62.Vehicular Technology

62• Vehicular Technology

62• Vehicular Technology Antenna Arrays for Mobile Communications Abstract | Full Text: PDF (206K) Cellular Radio Abstract | Full Text: PDF (152K) Ground Transportation Systems Abstract | Full Text: PDF (803K) Mobile Radio Channels Abstract | Full Text: PDF (603K) Mobile Satellite Communication Abstract | Full Text: PDF (445K) Mobile Telecommunications Standards Abstract | Full Text: PDF (246K) Multiple Access Mobile Communications Abstract | Full Text: PDF (217K) Paging Communication for Locating Mobile Users Abstract | Full Text: PDF (161K) Vehicle Navigation and Information Systems Abstract | Full Text: PDF (178K) Vehicular Electronics Abstract | Full Text: PDF (270K)

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELE...RONICS%20ENGINEERING/62.Vehicular%20Technology.htm15.06.2008 12:41:59

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELECTRICAL...NICS%20ENGINEERING/62.%20Vehicular%20Technology/W7709.htm

}{{}}

●

HOME ●

ABOUT US ●

CONTACT US ●

HELP

Home / Engineering / Electrical and Electronics Engineering

Wiley Encyclopedia of Electrical and Electronics Engineering Antenna Arrays for Mobile Communications Standard Article Arogyaswami Paulraj1, David Gesbert1, Constantinos Papadias2 1Stanford University, Stanford, CA 2Lucent Technologies, Holmdel, NJ Copyright © 1999 by John Wiley & Sons, Inc. All rights reserved. DOI: 10.1002/047134608X.W7709 Article Online Posting Date: December 27, 1999 Abstract | Full Text: HTML PDF (206K)

●

●

● ●

Recommend to Your Librarian Save title to My Profile Email this page Print this page

Browse this title ●

Search this title Enter words or phrases

Abstract The sections in this article are Early Forms of Spatial Processing Emerging Application of Space–Time Processing Channel Models Data Models SPace–Time Algorithms for the Reverse Link Multiuser Receiver Space–Time Algorithms for the Forward Link Applications of Space–Time Processing Summary About Wiley InterScience | About Wiley | Privacy | Terms & Conditions Copyright © 1999-2008John Wiley & Sons, Inc. All Rights Reserved.

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELE...ENGINEERING/62.%20Vehicular%20Technology/W7709.htm15.06.2008 12:43:34

❍

❍ ❍

Advanced Product Search Search All Content Acronym Finder

ANTENNA ARRAYS FOR MOBILE COMMUNICATIONS

551

Figure 1. The user signal experiences multipath propagation and impinges on a two-element array on a building rooftop.

ANTENNA ARRAYS FOR MOBILE COMMUNICATIONS Wireless cellular networks are growing rapidly around the world, and this trend is likely to continue for several years. The progress in radio technology enables new and improved services. Current wireless services include transmission of voice, fax, and low-speed data. More bandwidth-consuming interactive multimedia services such as video-on-demand and Internet access will be supported in the future. Wireless networks must provide these services in a wide range of environments, spanning dense urban, suburban, and rural areas. Varying mobility needs must also be addressed. Wireless local loop networks serve fixed subscribers. Microcellular networks serve pedestrians and other slow-moving users, and macrocellular networks serve high-speed vehicle-borne users. Several competing standards have been developed for terrestrial networks. AMPS (advanced mobile phone system) is an example of a first-generation frequency division multiple-access analog cellular system. Second-generation standards include GSM (global system for mobile) and IS-136, using time division multiple access (TDMA); and IS-95, using code division multiple access (CDMA). IMT-2000 is proposed to be the third-generation standard and will use mostly a wideband CDMA technology. Increased services and lower costs have resulted in an increased air time usage and number of subscribers. Since the

radio (spectral) resources are limited, system capacity is a primary challenge for current wireless network designers. Other major challenges include (1) an unfriendly transmission medium, with multipath transmission, noise, interference, and time variations, (2) the limited battery life of the user’s handheld terminal, and (3) efficient radio resource management to offer high quality of service. Current wireless modems use signal processing in the time dimension alone through advanced coding, modulation, and equalization techniques. The primary goal of smart antennas in wireless communications is to integrate and exploit efficiently the extra dimension offered by multiple antennas at the transceiver in order to enhance the overall performance of the network. Smart antenna systems use modems that combine the signals of multielement antennas in both space and time. Smart antennas can be used for both receive and transmit, both at the base station and at the user terminal. The use of smart antennas at the base alone is more typical, since practical constraints usually limit the use of multiple antennas at the terminal. See Fig. 1 for an illustration. Space–time processing offers various advantages. The first is array gain: multiple antennas capture more signal energy, which can be combined to improve the signal-to-noise ratio (SNR). Next, spatial diversity obtained from multiple antennas can be used to combat channel fading. Finally, space– time processing can help mitigate intersymbol interference (ISI) and cochannel interference (CCI). These leverages can be traded for improvements in: • • • •

Coverage: square miles per base station Quality: bit error rate (BER); outage probability Capacity: erlangs per hertz per base station Data rates: bits per second per hertz per base station

EARLY FORMS OF SPATIAL PROCESSING Adaptive Antennas The use of adaptive antennas dates back to the 1950s with their applications to radar and antijam problems. The pri-

J. Webster (ed.), Wiley Encyclopedia of Electrical and Electronics Engineering. Copyright # 1999 John Wiley & Sons, Inc.

552

ANTENNA ARRAYS FOR MOBILE COMMUNICATIONS

mary goal of adaptive antennas is the automatic generation of beams (beamforming) that track a desired signal and possibly reject (or null) interfering sources through linear combining of the signals captured by the different antennas. An early contribution in the field of beamforming was made in 1956 by Altman and Sichak, who proposed a combining device based on a phase-locked loop. This work was later refined in order to incorporate the adjustment of antenna signals in both phase and gain, allowing improved performance of the receiver in the presence of strong jammers. Howells proposed the sidelobe canceler for adaptive nulling. Optimal combining schemes were also introduced in order to minimize different criteria at the beamformer output. These include the minimum mean squared error (MMSE) criterion, as in the LMS algorithm proposed by Widrow; the signal-to-interferenceand-noise ratio (SINR) criterion proposed by Applebaum, and the minimum-variance beamformer distortionless response (MVDR) beamformer proposed by Capon. Further advances in the field were made by Frost, Griffiths, and Jim among several others. A list of references in beamforming can be found in Refs. 1 and 13. Besides beamforming, another application of antenna arrays is direction-of-arrival (DOA) estimation for source or target localization purposes. The leading DOA estimation methods are the MUSIC and ESPRIT algorithms (2). In many of the beamforming techniques (for instance in Capon’s method), the estimation of the source direction is an essential step. DOA estimation is still an area of active research. Antenna arrays for beamforming and source localization are of course of great interest in military applications. However, their use in civilian cellular communication networks is now gaining increasing attention. By enabling the transmission and reception of signal energy from selected directions, beamformers play an important role in improving the performance of both the base-to-mobile (forward) and mobile-tobase (reverse) links. Antenna Diversity Antenna diversity can alleviate the effects of channel fading, and is used extensively in wireless networks. The basic idea of space diversity is as follows: if several replicas of the same information-carrying signal are received over multiple branches with comparable strengths and exhibit independent fading, then there is a high probability that at least one branch will not be in a fade at any given instant of time. When a receiver is equipped with two or more antennas that are sufficiently separated (typically several wavelengths), they offer useful diversity branches. Diversity branches tend to fade independently; therefore, a proper selection or combining of the branches increases link reliability. Without diversity, protection against deep channel fades requires higher transmit power to ensure the link margins. Therefore, diversity at the base can be traded for reduced power consumption and longer battery life at the user terminal. Also, lower transmit power decreases the amount of co-channel-user interference and increases the system capacity. Independent fading across antennas is achievable when radio waves impinge on the antenna array with sufficient angle spread. Paths coming from different arriving directions will add differently (constructive or destructive manner) at each

antenna. This requires the presence of significant scatterers in the propagation medium, such as in urban or hilly terrain. Diversity also helps to combat large-scale fading effects caused by shadowing from large obstacles (e.g., buildings or terrain features). However, antennas located in the same base station experience the same shadowing. Instead, antennas from different base stations can be combined to offer a protection against such fading (macro diversity). Antenna diversity can be complemented by other forms of diversity. Polarization, time, frequency, and path diversity are some examples. These are particularly useful when physical constraints prevent the use of multiple antennas (for instance at the hand-held terminal). See Ref. 3 for more details. Combining the different diversity branches is an important issue. The main options used in current systems are briefly described below. In all cases, independent branch fading and equal mean branch powers are assumed. However, in nonideal situations, branch correlation and unequal powers will result in a loss of diversity gain. A correlation coefficient as high as 0.7 between instantaneous branch envelope levels is considered acceptable. Selection Diversity. Selection diversity is one of the simplest form of diversity combining. Given several branches with varying carrier-to-noise ratios (C/N), selection diversity consists in choosing the branch having the highest instantaneous C/N. The performance improvement from selection diversity is evaluated as follows: Let us suppose that M branches experience independent fading but have the same mean C/N, denoted by ⌫. Let us now denote by ⌫s the mean C/N of the selected branch. Then it can be shown that (4)

s = 

M
1 j=1

j

For instance, selection over two branches increases the mean C/N by a factor of 1.5. More importantly, the statistics of the instantaneous C/N is improved. Note that selection diversity requires a receiver behind each antenna. Switching diversity is a variant of selection diversity. In this method, a selected branch is held until it falls below a threshold T, at which point the receiver switches to another branch, regardless of its level. The threshold can be fixed or adaptive. This strategy performs almost as well as the selection method described above, and it reduces the system cost, since only one receiver is required. Maximum-Ratio Combining. Maximum-ratio combining (MRC) is an optimal combining approach to combat fading. The signals from M branches are first cophased to mutual coherence and then summed after weighting. The weights are chosen to be proportional to the signal level to maximize the combined C/N. It can be shown that the gain from MRC in mean C/N is directly proportional to the number of branches: s = M Equal-Gain Combining. Although optimal, MRC is expensive to implement. Also, MRC requires accurate tracking of the complex fading, which is difficult to achieve in practice. A simpler alternative is given by equal-gain combining, which consists in summing the cophased signals using unit weights.

ANTENNA ARRAYS FOR MOBILE COMMUNICATIONS

Fading

ISI Tx

Rx

Rx CCI Tx CCI

Co-channel Tx-user

Noise

553

Smart antennas can also be used at the transmitter to maximize array gain and/or diversity, and to mitigate ISI and CCI. In the transmit case, however, the efficiency of space– time processing schemes is usually limited by the lack of accurate channel information. The major effects induced by radio propagation in a cellular environment are pictured in Fig. 2. The advantages offered by space–time processing for receive and transmit are summarized in Table 1. In the following sections, we describe channel models and algorithms used in space–time processing. Both simple and advanced solutions are presented, and tradeoffs highlighted. Finally, we describe current applications of smart antennas.

Co-channel Tx-user

Figure 2. Smart antennas help mitigate the effects of cellular radio propagation.

The performance of equal-gain combining is found to be very close to that of MRC. The SNR of the combined signals using equal gain is only 1 dB below the SNR provided by MRC (4). EMERGING APPLICATION OF SPACE–TIME PROCESSING While the use of beamforming and space diversity proves useful in radio communication applications, an inherent limitation of these techniques lies in the fact that they exploit signal combining in the space dimension only. Directional beamforming, in particular, heavily relies on the exploitation of the spatial signatures of the incoming signals but does not consider their temporal structure. The techniques that combine the signals in both time and space can bring new advantages, and their importance in the area of mobile communications is now recognized (5). The main reason for using space–time processing is that it can exploit the rich temporal structure of digital communication signals. In addition, multipath propagation environments introduce signal delay spread, making techniques that exploit the complete space–time structure more natural. The typical structure of a space–time processing device consists of a bank of linear filters, each located behind a branch, followed by a summing network. The received space– time signals can also be processed using nonlinear schemes, for example, maximum-likelihood sequence detection. The space–time receivers can be optimized to maximize array and diversity gains, and to minimize: 1. Intersymbol interference (ISI), induced by the delay spread in the propagation channel. ISI can be suppressed by selecting a space–time filter that equalizes the channel or by using a maximum-likelihood sequence detector. 2. Co-channel-user interference (CCI), coming from neighboring cells operating at the same frequency. CCI is suppressed by using a space–time filter that is orthogonal to the interference’s channel. The key point is that CCI that cannot be rejected by space-only filtering may be handled more effectively using space–time filtering.

CHANNEL MODELS Channel models capture radio propagation effects and are useful for simulation studies and performance prediction. Channel models also help in motivating appropriate signalprocessing algorithms. The effects of radio propagation on the transmitted signal can be broadly categorized into two main classes: fading and spreading. Fading refers to the propagation losses experienced by the radio signal (on both the forward and reverse links). One type of fading, called selective fading, causes the received signal level to vary around the average level in some regions of space, frequency, or time. Channel spreading refers to the spreading of the information-carrying signal energy in space, and on the time or frequency axis. Selective fading and spreading are complementary phenomena. Channel Fading Mean Path Loss. The mean path loss describes the attenuation of a radio signal in free-space propagation, due to isotropic power spreading, and is given by the well-known inverse square law: Pr = Pt




λ 4πd

2 Gt Gr

where Pr and Pt are the received and transmitted powers, ␭ is the radio wavelength, d is the range, and Gt, Gr are the gains of the transmit and receive one-element antennas respectively. In cellular environments, the main path is often accompanied by a surface-reflected path which may interfere destructively with the primary path. Specific models have been developed that consider this effect. The path loss model becomes (4)

Table 1. Advantages of Space–Time Processing For transmit (Tx)

Reduces Tx CCI Maximizes Tx diversity Reduces ISI Increases Tx EIRP

For receive (Rx)

Reduces Rx CCI Maximizes Rx diversity Eliminates ISI Increases C/N

0

10 –30 Delay (µ s)

Power

Power

ANTENNA ARRAYS FOR MOBILE COMMUNICATIONS

Power

554

0 30 –fm 0 fm Doppler (Hz) Angle (deg)

Figure 3. The radio channel induces spreading in several dimensions. These spreads strongly affect the design of the space–time receiver.

Pr = Pt


h h 2 r r d2

Gt Gr

where ht, hr are the effective heights of the transmit and receive antennas, respectively. Note that this particular path loss model follows an inverse fourth-power law. In fact, depending on the environment, the path loss exponent may vary from 2.5 to 5. Slow Fading. Slow fading is caused by long-term shadowing effects of buildings or natural features in the terrain. It can also be described as the local mean of a fast fading signal (see below). The statistical distribution of the local mean has been studied experimentally and shown to be influenced by the antenna height, the operating frequency, and the type of environment. It is therefore difficult to predict. However, it has been observed that when all the above-mentioned parameters are fixed, then the received signal fluctuation approaches a normal distribution when plotted on a logarithmic scale (i.e., in decibels) (4). Such a distribution is called lognormal. A typical value for the standard deviation of shadowing distribution is 8 dB. Fast Fading. The multipath propagation of the radio signal causes path signals to add up with random phases, constructively or destructively, at the receiver. These phases are determined by the path length and the carrier frequency, and can vary extremely rapidly along with the receiver location. This gives rise to fast fading: large, rapid fluctuations of the received signal level in space. If we assume that a large number of scattered wavefronts with random amplitudes and angles of arrival arrive at the receiver with phases uniformly distributed in [0, 2앟), then the in-phase and quadrature phase components of the vertical electrical field Ez can be shown to be Gaussian processes (4). In turn, the envelope of the signal can be well approximated by a Rayleigh process. If there is a direct path present, then it will no longer be a Rayleigh distribution but becomes a Rician distributed instead.

Doppler Spread. When the mobile user is in motion, the radio signal at the receiver experiences a shift in the frequency domain (called the Doppler shift), the amplitude of which depends on the path direction of arrival. In the presence of surrounding scatterers with multiple directions, a pure tone is spread over a finite spectral bandwidth. In this case, the Doppler power spectrum is defined as the Fourier transform of the time autocorrelation of the received signal, and the Doppler spread is the support of the Doppler power spectrum. Assuming scatterers uniformly distributed in angle, the Doppler power spectrum is given by the so-called classical spectrum:


f − f 2 −1/2 3σ 2 c 1− S( f ) = , 2πf m fm

where f m ⫽ v/ ␭ is the maximum Doppler shift, v is the mobile velocity, f c is the carrier frequency, and ␴2 is the signal variance. When there is a dominant source of energy coming from a particular direction (as in line-of-sight situations), the expression for the spectrum needs to be corrected according to the Doppler shift of the dominant path f D: S( f ) + Bδ( f − f D ) where B denotes the ratio of direct to scattered path energy. The Doppler spread causes the channel characteristics to change rapidly in time, giving rise to the so-called time selectivity. The coherence time, during which the fading channel can be considered as constant, is inversely proportional to the Doppler spread. A typical value of the Doppler spread in a macrocell environment is about 200 Hz at 30 m/s (65 mi/h) in the 1900 MHz band. A large Doppler spread makes good channel tracking an essential feature of the receiver design. Delay Spread. Multipath propagation is often characterized by several versions of the transmitted signal arriving at the receiver with different attenuation factors and delays. The spreading in the time domain is called delay spread and is responsible for the selectivity of the channel in the frequency domain (different spectral components of the signal carry different powers). The coherence bandwidth, which is the maximum range of frequencies over which the channel response can be viewed as constant, is inversely proportional to the delay spread. Significant delay spread may cause strong intersymbol interference, which makes necessary the use of a channel equalizer. Angle Spread. Angle spread at the receiver refers to the spread of directions of arrival of the incoming paths. Like-

Table 2. Typical Delay, Angle, and Doppler Spreads in Cellular Radio Systems

Channel Spreading Propagation to or from a mobile user, in a multipath channel, causes the received signal energy to spread in the frequency, time, and space dimensions (see Fig. 3, and also Table 2 for typical values). The characteristics of the spreading [that is to say, the particular dimension(s) in which the signal is spread] affects the design of the receiver.

fc − fm < f < fc + fm

Environment Flat rural (macro) Urban (macro) Hilly (macro) Micro cell (mall) Pico cell (indoors)

Delay Spread (애s)

Angle Spread (deg)

Doppler Spread (Hz)

0.5 5 20 0.3 0.1

1 20 30 120 360

190 120 190 10 5

ANTENNA ARRAYS FOR MOBILE COMMUNICATIONS

Remote scatterers

Scatterers local to mobile user

Scatterers local to base Figure 4. Each type of scatterer introduces specific channel spreading characteristics.

wise, angle spread at the transmitter refers to the spread of departure angles of the paths. As mentioned earlier, a large angle spread will cause the paths to add up in a random manner at the receiver as the location of the receive antenna varies; hence it will be source of space-selective fading. The range of space for which the fading remains constant is called the coherence distance and is inversely related to the angle spread. As a result, two antennas spaced by more than the coherence distance tend to experience uncorrelated fading. When the angle spread is large, which is usually the case in dense urban environments, a significant gain can be obtained from space diversity. Note that this usually conflicts with the possibility of using directional beamforming, which typically requires well-defined and dominant signal directions, that is, a low angle spread.

555

Micro Cells. Micro cells are characterized by highly dense built-up areas, and by the user’s terminal and base being relatively close (a few hundred meters). The base antenna has a low elevation and is typically below the rooftops, causing significant scattering in the vicinity of the base. Micro-cell situations make the propagation difficult to analyze, and the macro-cell model described earlier no longer can be expected to hold. Very high angle spreads along with small delay spreads are likely to occur in this situation. The Doppler spread can be as high as in macro cells, although the mobility of the user is expected to be limited, due to the presence of mobile scatterers. Parametric Channel Model A complete and accurate understanding of propagation effects in the radio channel requires a detailed description of the physical environment. The specular model, to be presented below, only provides a simplified description of the physical reality. However, it is useful, as it describes the main channel effects and it provides the means for a simple and efficient mathematical treatment. In this model, the multiple elementary paths are grouped according to a (typically small) number L of main path clusters, each of which contains paths that have roughly the same mean angle and delay. Since the paths in these clusters originate from different scatterers, the clusters typically have near-independent fading. Based on this model, the continuous-time channel response from a single transmit antenna to the ith antenna of the receiver can be written as

f i (t) =

L


ai (θl )αlR (t)δ(t − τl )

(1)

l=1

Multipath Propagation Macro Cells. A macro cell is characterized by a large cell radius (up to a few tens of kilometers) and a base station located above the rooftops. In macro-cell environments, the signal energy received at the base station comes from three main scattering sources: scatterers local to the mobile, remote dominant scatterers, and scatterers local to the base (see Fig. 4 for an illustration). The following description refers to the reverse link but applies to the forward link as well. The scatterers local to the mobile user are those located a few tens of meters from the hand-held terminal. When the terminal is in motion, these scatterers give rise to a Doppler spread, which causes time-selective fading. Because of the small scattering radius, the paths that emerge from the vicinity of the mobile user and reach the base station show a small delay spread and a small angle spread. Of the paths emerging from the local-to-mobile scatterers, some reach remote dominant scatterers, such as hills or highrise buildings, before eventually traveling to the base station. These paths will typically reach the base with medium to large angle and delay spreads (depending of course on the number and locations of these remote scatterers). Once these multiple wavefronts reach the vicinity of the base station, they usually are further scattered by local structures such as buildings or other structures that are close to the base. These scatterers local to the base can cause large angle spread; therefore they can cause severe space-selective fading.

where 움Rl (t), ␪l, and ␶l are respectively the fading (including mean path loss and slow and fast fading), the angle, and the delay of the lth receive path cluster. Note that this model also includes the response of the ith antenna to a path from direction ␪l, denoted by ai(␪l). In the following we make use of the specular model to describe the structure of the signals in space and time. Note that in the situation where the path cluster assumption is not acceptable, other channel models, called diffuse channel models, are more appropriate (6). DATA MODELS This section focuses on developing signal models for space– time processing algorithms. The transmitted information signal is assumed to be linearly modulated. In the case of a nonlinear modulation scheme, such as the Gaussian minimum shift keying (GMSK) used in the GSM system, linear approximations are assumed to hold. The baseband equivalent of the transmitted signal can be written (7) u(t) =




g(t − kT )s(k) + n(t)

(2)

k

where s(k) is the symbol stream, with rate 1/T, g(t) is the pulse-shaping filter, and n(t) is an additive thermal noise. Four configurations for the received signal (two for the reverse link and two for the forward link) are described below.

556

ANTENNA ARRAYS FOR MOBILE COMMUNICATIONS

channel. As will be emphasized later, this is a challenging situation, as the transmitter typically lacks reliable information on the channel. For the sake of simplicity, we will assume here that a space-only beamforming weight vector w is used, as the extension to space–time beamforming is straightforward. The baseband signal received at the mobile station is scalar and is given by

F : Forward link R: Reverse link F R Hand-held terminals

Base station

Single-user case

x(t) = R R F

Hand-held terminals Multiuser case Figure 5. Several configurations are possible for antenna arrays, in transmit (T) and in receive (R).

These are also depicted in Fig. 5. In each case, one assumes M ⬎ 1 antennas at the base station and a single antenna at the mobile user. Reverse Link We consider the signal received at the base station. Since the receiver is equipped with M antennas, the received signal can be written as a vector x(t) with M entries. Single-User Case. Let us assume a single user transmitting towards the base (no CCI). Using the specular channel model in Eq. (1), the received signal can be written as follows:

x (t) =

L


w Ha (θl )αlF (t)u(t − τl ) + n(t)

(5)

l=1

F

Base station

L


a (θl )αlR (t)u(t − τl ) + n (t)

(3)

l=1

where 움Fl (t) is the fading coefficient of the lth transmit path in the forward link. Superscript H denotes the transpose-conjugation operator. Note that path angles and delays remain theoretically unchanged in the forward and reverse links. This is in contrast with the fading coefficients, which depend on the carrier frequency. Frequency division duplex (FDD) systems use different carriers for the forward and reverse links, which result in 움Fl (t) and 움Rl (t) being nearly uncorrelated. In contrast, time division duplex (TDD) systems will experience almost identical forward and reverse fading coefficients in the forward and reverse links. Assuming however that the transmitter knows the forward fading and delay parameters, transmit beamforming can offer array gain, ISI suppression, and CCI suppression. Multiuser Case. In the multiuser case, the base station wishes to communicate with Q users, simultaneously and in the same frequency band. This can be done by superposing, on each of the transmit antennas, the signals given by Q beamformers w1, . . ., wQ. At the mth user, the received signal waveform contains the signal sent to that user, plus an interference from signals intended for all other users. This gives

xm (t) =

Lq Q



F wH q a (θlm )αlm (t)uq (t − τlm ) + nm (t)

(6)

q=1 l=1

where a(␪l) ⫽ (a1(␪l), . . ., aM(␪l))T is the vector array response to a path of direction ␪l, and where T refers to the transposition operator.

Note that each information signal uq(t) couples into the Lm paths of the mth user through the corresponding weight vector wq, for all q.

Multiuser Case. We now have Q users transmitting towards the base. The received signal is the following sum of contributions from the Q users, each of them carries a different set of fading, delays, and angles:

A Nonparametric Model

x (t) =

Lq Q



R a (θlq )αlq (t)uq (t − τlq ) + n (t)

(4)

q=1 l=1

where the subscript q refers to the user index. Forward Link Single-User Case. In this case, the base station uses a transmitter equipped with M antennas to send an information signal to a unique user. Therefore, space–time processing must be performed before the signal is launched into the

The data models above build on the parametric channel model developed earlier. However, there is also interest in considering the end-to-end channel impulse response of the system to a transmitted symbol rather than the physical path parameters. The channel impulse response includes the pulse-shaping filter response, the propagation phenomena, and the antenna response as well. One advantage of looking at the impulse response is that the effects of ISI and CCI can be described in a better and more compact way. A second advantage is that the nonparametric channel only relies on the channel linearity assumption. We look at the reverse-link and single-user case only. Since a single scalar signal is transmitted and received over several branches, this corresponds to a single-input multiple-output

ANTENNA ARRAYS FOR MOBILE COMMUNICATIONS

s(n)

Oversampling only increases the number of scalar observations per transmitted symbol, which can be regarded mathematically as increasing the number of channel components, in a way similar to increasing the number of antennas. Hence the model above also holds true when sampling at T/2, T/3, . . .. However, though mathematically equivalent, spatial oversampling and temporal oversampling lead to different signal properties.

s(n) Modulator

Space–time combiner Propagation channel

s(n)

s(n) Global channel

Structure of the Linear Space–Time Beamformer

Space–time combiner

Figure 6. The source signal s(n) can be seen as driving a singleinput multiple-output filter with M outputs, where M is the number of receive antennas.

(SIMO) system, depicted in Fig. 6. The model below is also easily extended to multiuser channels. Let h(t) denote the M ⫻ 1 global channel impulse response. The received vector signal is given by the result of a (noisy) convolution operation: x (t) =




h (t − kT )s(k) + n (t)

Space combining is now considered at the receive antenna array. Let w be a M ⫻ 1 space-only weight vector (a single complex weight is assigned to each antenna). The output of the combiner, denoted by y(k) is as follows: y(k) = w Hx (k) The resulting beamforming operation is depicted in Fig. 7. The generalization to space–time combining is straightforward: Let the combiner have m time taps. Each tap, denoted by w(i), i ⫽ 0, . . ., m ⫺ 1, is an M ⫻ 1 space weight vector defined as above. The output of the space–time beamformer is now written as

(7)

k

y(k) =

From Eqs. (2) and (3), the channel response may also be expressed in terms of the specular model parameters through

h (t) =

L


a (θl )αlR (t)g(t − τl )

Signal Sampling. Consider sampling the received signal at the baud (symbol) rate, that is, at tk ⫽ t0 ⫹ kT, where t0 is an arbitrary phase. Let N be the maximum length of the channel response in symbol periods. Assuming that the channel is invariant for some finite period of time [i.e., 움Rl (t) ⫽ 움Rl ], the received vector sample at time tk can be written as (9)

where H is the sampled channel matrix, with size M ⫻ N, whose (i, j) term is given by

[H ]ij =

L


m−1


(11)

which can be reformulated as y(k) = W HX (k)

(12)

where W ⫽ (w(0)H, . . ., w(m ⫺ 1)H)H and X(k) is the data vector compactly defined as X(k) ⫽ (x(k)H, . . ., x(k ⫺ m ⫹ 1)H)H. ISI and CCI Suppression The formulation above gives insight into the algebraic structure of the space–time received data vector. Also it allows us to identify the conditions under which the suppression of ISI and/or CCI is possible. Recalling the signal model in Eq. (9), the space–time vector X(k) can be in turn written as X (k) = H S (k) + N (k)

ai (θl )αlR g(t0 + jT − τl )

Channel

l=1

(13)

Equalizer w*1

and where s(k) is the vector of N ISI symbols at the time of the measurement:

s1

h1

x1 (k)

...

w*2

s (k) = (s(k), s(k − 1), . . ., s(k − N + 1))T

x2 (k)

To allow for the presence of CCI, Eq. (9) can be generalized to

x (k) =

w (i)Hx (k − i)

i=0

(8)

l=1

x (k) = H s (k) + n (k)

557

Q


H qs q (k) + n (k)

. . . sQ

(10)

. . . hQ

...

. . .

Σ

w*m xm (k)

q=1

where Q denotes the number of users and q the user index. Most digital modems use sampling of the signal at a rate higher than the symbol rate (typically up to four times).

Figure 7. Structure of the spatial beamformer. The space–time beamformer is a direct generalization that combines in time the outputs of several spatial beamformers.

558

ANTENNA ARRAYS FOR MOBILE COMMUNICATIONS

where S(k) ⫽ (s(k), s(k ⫺ 1), . . ., s(k ⫺ m ⫹ N ⫹ 2))T and where

H

0 BB BB =B BB @

H

0

0 .. .

H .. .

0

···

··· .. . ..

0 .. .

.

0

0

H

1 CC CC CC CA

WqHH f = 0

(14)

is a mM ⫻ (m ⫹ N ⫺ 1) channel matrix. The block Toeplitz structure in H stems from the linear time-invariant convolution operation with the symbol sequence. Let us temporarily assume a noise-free scenario. Then the output of a linear space–time combiner can be described by the following equation: y(k) = W HH S (k)

(15)

In the presence of Q users transmitting towards the base station, the output of the space–time receiver is generalized to

y(k) =

Q


W HH q S q (k)

(16)

q=1

ISI Suppression. The purpose of equalization is to compensate for the effects of ISI induced by the user’s channel in the absence of CCI. Tutorial information on equalization can be found in Refs. 7, and 8. In general, a linear filter Wq is an equalizer for the channel of the qth user if the convolution product between Wq and the channel responses yields a Dirac function, that is, if Wq satisfies the following so-called zeroforcing condition: WqHH q = (0, . . ., 0, 1, 0, . . ., 0)

signal of user q using Wq, the following conditions must be satisfied (possibly approximately):

(17)

Here, the location of the 1 element represents the delay of the combined channel–equalizer impulse response. Note that from an algebraic point of view, the channel matrix H q should have more rows than columns for such solutions to exist: mM ⱖ m ⫹ N ⫺ 1. Therefore, it is essential to have enough degrees of freedom (number of taps in the filter) to allow for ISI suppression. Note that zero-forcing solutions can be obtained using temporal oversampling at the receive antenna only, since oversampling by a factor of M provides us theoretically with M baud-rate branches. However, having multiple antennas at the receiver plays a important role in improving the conditioning of the matrix H , which in turn will improve the robustness of the resulting equalizer in the presence of noise. It can be shown that the condition number of the matrixe H q is related to some measure of the correlation between the entries of h(t). Hence, a significant antenna spacing is required to provide the receiver with sufficiently decorrelated branches. CCI Suppression. The purpose of CCI suppression in a multiple access network is to isolate the contribution of one desired user by rejecting that of others. One way to achieve this goal is to enforce orthogonality between the response of the space–time beamformer and the response of the channel of the users to be rejected. In other words, in order to isolate the

for all

f = q ∈ [1, . . ., Q]

(18)

If we assume that all the channels have the same maximum order N, Eq. (18) provides as many as (Q ⫺ 1)(m ⫹ N ⫺ 1) scalar equations. The number of unknowns is again given by mM (the size of Wq). Hence a receiver equipped with multiple antennas is able to provide the number of degrees of freedom necessary for signal separation. This requires mM ⱖ (Q ⫺ 1)(m ⫹ N ⫺ 1). At the same time, it is desirable that the receiver capture a significant amount of energy from the desired user; hence an extra condition on Wq should be W qH H q ⬆ 0. From an algebraic perspective, this last condition H f 其f⬆q should not have the same column requires that H q and 兵H subspaces. The required subspace misalignment between the desired user and the interferers in space–time processing is a generalization of the condition that signal and interference should not have the same direction, needed for interference nulling using beamforming. Joint ISI and CCI Suppression. The complete recovery of the signal transmitted by one desired user in the presence of ISI and CCI requires both channel equalization and separation. A space–time beamformer is an exact solution to this problem if it satisfies both Eq. (17) and Eq. (18), which can be further written as H 1 , . . ., H q−1 , H q , H q+1 , . . ., H Q ) = (0, . . ., 0, 1, 0, . . ., 0) WH q (H (19) where the location of the 1 element designates both the index of the user of reference and the reconstruction delay. The existence of solutions to this problem requires the multiuser def H 1, . . ., H Q) to have more rows than channel matrix H * ⫽ (H columns: mM ⱕ Q(m ⫹ N ⫺ 1). Here again, smart antennas play a critical role in offering a sufficient number of degrees of freedom. If, in addition, the global channel matrix H * has full column rank, then we are able to recover any particular user using space–time beamforming. In practice, though, the performance of an ISI–CCI reduction scheme is limited by the SNR and the condition number of H . * SPACE–TIME ALGORITHMS FOR THE REVERSE LINK General Principles of Receive Space–Time Processing Space processing offers several important opportunities to enhance the performance of the radio link. First, smart antennas offer more resistance to channel fast fading through maximization of space diversity. Then, space combining increases the received SNR through array gain, and allows for the suppression of interference when the user of reference and the co-channel-users have different DOAs. Time processing addresses two important goals. First, it exploits the gain offered by path diversity in delay-spread channels. As the channel time taps generally carry independent fading, the receiver can resolve channel taps and combine them to maximize the signal level. Second, time processing can combat the effects of ISI through equalization. Linear zero-forcing equalizers address the ISI problem but do

ANTENNA ARRAYS FOR MOBILE COMMUNICATIONS

not fully exploit path diversity. Hence, for these equalizers, ISI suppression and diversity maximization may be conflicting goals. This not the case, however, for maximum likelihood sequence detectors. Space-time processing allows us to exploit the advantage of both the time and space dimensions. Space-time (linear) filters allow us to maximize space and path diversity. Also, space-time filters can be used for better ISI and CCI reduction. However, as was mentioned above, these goals may still conflict. In contrast, space–time maximum likelihood sequence detectors (see below) can handle harmoniously both diversity maximization and interference minimization. Channel Estimation Channel estimation forms an essential part of most wireless digital modems. Channel estimation in the reverse link extracts the information that is necessary for a proper design of the receiver, including linear (space–time beamformer) and nonlinear (decision feedback or maximum likelihood detection) receivers. For this task, most existing systems rely on the periodic transmission of training sequences, which are known both to the transmitter and receiver and used to identify the channel characteristics. The estimation of H is usually performed using the nonparametric FIR model in Eq. (9), in a least-squares manner or by correlating the observed signals against decorrelated training sequences (as in GSM). A different strategy consists in addressing the estimation of the physical parameters (path angle and delays) of the channel using the model developed in Eq. (8). This strategy proves useful when the number of significant paths is much smaller than the number of channel coefficients. Channel tracking is also an important issue, necessary whenever the propagation characteristics vary (significantly) within the user slot. Several approaches can be used to update the channel estimate. The decision-directed method uses symbol decisions as training symbols to update the channel response estimate. Joint data–channel techniques constitute another alternative, in which symbol estimates and channel information are recursively updated according to the minimization of a likelihood metric: 储X ⫺ H S储2. Signal Estimation Maximum Likelihood Sequence Detection (Single User). Maximum likelihood sequence detection (MLSD) is a popular nonlinear detection scheme that, given the received signal, seeks the sequence of symbols of one particular user that is most likely to have been transmitted. Assuming temporally and spatially white Gaussian noise, maximizing the likelihood reduces to finding the vector S of symbols in a given alphabet that minimizes the following metric: X − H S 2 min X S

(20)

where the channel matrix H has been previously estimated. Here, 储 ⭈ 储 denotes the conventional euclidean norm. Since X contains measurements in time and space, the criterion above can be considered as a direct extension of the conventional ML sequence detector, which is implemented recursively using the well-known Viterbi algorithm (7). MLSD offers the lowest BER in a Gaussian noise environment, but is no longer optimal in the presence of co-channel-

559

users. In the presence of CCI, a solution to the MLSD problem consists in incorporating in the likelihood metric the information on the statistics of the interferers. This however assumes that the interferers do not undergo significant delay spread. In general though, the optimal solution is given by a multiuser MLSD detection scheme (see below). Maximum Likelihood Sequence Detection (Multiuser). The multiuser MLSD scheme has been proposed for symbol detection in CCI-dominated channels. The idea consists in treating CCI as other desired users and detecting all signals simultaneously. This time, the Q symbol sequences S1, S2, . . ., SQ are found as the solutions to the following problem:

 2 Q
   X − min  H S q q   S q} {S

(21)

q=1

where again all symbols should belong to the modulation alphabet. The resolution of this problem can be carried out theoretically by a multiuser Viterbi algorithm. However, the complexity of such a scheme grows exponentially with the numbers of users and the channel length, which limits its applicability. Also, the channels of all the users are assumed to be accurately known. In current systems, such information is very difficult to obtain. In addition, the complexity of the multiuser MLSD detector falls beyond current implementation limits. Suboptimal solutions are therefore necessary. One possible strategy, known as onion peeling, consists in first decoding the user having the largest power and then subtracting it out from the received data. The procedure is repeated on the residual signal, until all users are decoded. Linear receivers, described below, constitute another form of suboptimal but simple approach to signal detection. Minimum mean square error detection is described below. Minimum Mean Squared Error Detection. The space–time minimum mean squared error (STMMSE) beamformer is a space–time linear filter whose weights are chosen to minimize the error between the transmitted symbols of a user of reference and the output of the beamformer defined as y(k) ⫽ W qHX(k). Consider a situation with Q users. Let q be the index of the user of reference. Wq is found by: min E| y(k) − sq (k − d)|2 Wq

(22)

where d is the chosen reconstruction delay. E here denotes the expectation operator. The solution to this problem follows from the classical normal equations: X (k)X X (k)H ]−1 E[X X (k)sq (k − d)∗ ] W q = E[X

(23)

The solution to this equation can be tracked in various manners, for instance using pilot symbols. Also, it can be shown that the intercorrelation term in the right-hand side of Eq. (23) corresponds to the vector of channel coefficients of the decoded user, when the symbols are uncorrelated. Hence Eq. (23) can also be solved using a channel estimate. STMMSE combines the strengths of time-only and spaceonly combining, hence is able to suppress both ISI and CCI. In the noise-free case, when the number of branches is large enough, Wq is found to be a solution of Eq. (19). In the pres-

560

ANTENNA ARRAYS FOR MOBILE COMMUNICATIONS

ence of additive noise, the MMSE solution provides a useful tradeoff between the so-called zero-forcing solution of Eq. (19) and the maximum-SNR solution. Finally the computational load of the MMSE is well below than that of the MLSD. However MLSD outperforms the MMSE solution when ISI is the dominant source of interference. Combined MMSE–MLSD. The purpose of the combined MMSE–MLSD space–time receiver is to be able to deal with both ISI and CCI using a reasonable amount of computation. The idea is to use a STMMSE in a first stage to combat CCI. This leaves us with a signal that is dominated by ISI. After channel estimation, a single-user MLSD algorithm is applied to detect the symbols of the user of interest. Note that the channel seen by the MLSD receiver corresponds to the convolution of the original SIMO channel with the equalizer response. Space–Time Decision Feedback Equalization. The decision feedback equalizer is a nonlinear structure that consists of a space–time linear feedforward filter (FFF) followed by a nonlinear feedback filter. The FFF is used for precursor ISI and CCI suppression. The nonlinear part contains a decision device which produces symbol estimates. An approximation of the postcursor ISI is formed using these estimates and is subtracted from the FFF output to produce new symbol estimates. This technique avoids the noise enhancement problem of the pure linear receiver and has a much lower computational cost than MLSD techniques. Blind Space–Time Processing Methods The goal of blind space–time processing methods is to recover the signal transmitted by one or more users, given only the observation of the channel output and minimal information on the symbol statistics and/or the channel structure. Basic available information may include the type of modulation alphabet used by the system. Also, the fact that channel is quasiinvariant in time (during a given data frame) is an essential assumption. Blind methods do not, by definition, resort to the transmission of training sequences. This advantage can be directly traded for an increased information bit rate. It also helps to cope with the situations where the length of the training sequence is not sufficient to acquire an accurate channel estimate. Tutorial information on blind estimation can be found in Ref. 9. Blind methods in digital communications have been the subject of active research over the last twenty years. It was only recently recognized, however, that blind techniques can benefit from the utilization of the spatial dimension. The main reason is that oversampling the signal in space using multiple antennas, together with the exploitation of the signal–channel structure, allows for efficient channel and beamformer estimation techniques.

HOS methods look at third- and fourth-order moments of the received data and exploit simple relationships between those moments and the channel coefficients (assuming the knowledge of the input moments) in order to identify the channel. In contrast, the SOS of the output of a scalar (single input-single output) channel do not convey sufficient information for channel estimation, since the second-order moments are phase-blind. In SIMO systems, SOS does provide the necessary phase information. Hence, one important advantage of multiantenna systems lies in the fact that they can be identified using second-order moments of the observations only. From an algebraic point of view, the use of antenna arrays creates a lowrank model for the vector signal given by the channel output. Specifically, the channel matrix H in Eq. (14) can be made tall and full-column-rank under mild assumption on the channels. The low-rank property allows one to identify the column span of H from the observed data. Along with the Toeplitz structure of H , this information can be exploited to identify the channel. Direct Estimation. Direct methods bypass the channel estimation stage and concentrate on the estimation of the spacetime filter. The use of antenna arrays (or oversampling in time or space in general) offers important advantages in this context too. The most important one is perhaps the fact that, as was shown in Eq. (17), the SIMO system can be inverted exactly using a space–time filter with finite time taps, in contrast with the single-output case. HOS methods for direct receiver estimation are typically designed to optimize a nonlinear cost function of the receiver output. Possible cost functions include Bussgang cost functions [Sato, decision-directed, and constant modulus (CM) algorithms] and kurtosis-based cost functions. The most popular criterion is perhaps the CM criterion, in which the coefficients of the beamformer W are updated according to the minimization (though gradient-descent algorithms) of W ) = E[|y(k)|2 − 1]2 J(W where y(k) is the beamformer output. SOS techniques (sometimes also referred to as ‘‘algebraic techniques’’) look at the problem of factorizing, at least implicitly, the received data matrix X into the product of a blockToeplitz channel matrix H and a Hankel symbol matrix S

X ≈H S

(24)

A possible strategy is as follows: Based on the fact that H is a tall matrix, the row span of S coincides with the row span of X. Along with the Hankel structure of S, the row span of S can be exploited to uniquely identify S. MULTIUSER RECEIVER

Blind Channel Estimation. A significant amount of research work has been focused lately on identifying blindly the impulse response of the transmission channel. The resulting techniques can be broadly categorized into three main classes: higher-order statistics (HOS) methods, second-order statistics (SOS) methods, and maximum-likelihood (ML) methods.

The extension of blind estimation methods to a multiuser scenario poses important theoretical and practical challenges. These challenges include an increased number of unknown parameters, more ambiguities caused by the problem of user mixing, and a higher complexity. Furthermore, situations where the users are not fully synchronized may result in an

Direct-sequence CDMA (DSCDMA) systems are expected to gain a significant share of the cellular market. In CDMA, the symbol stream is spread by a unique spreading code before transmission. The codes are designed to be orthogonal or quasiorthogonal to each other, making it possible for the users to be separated at the receiver. See Ref. 11 for details. As in TDMA, the use of smart antennas in CDMA system improves the network performance. We first introduce the DSCDMA model; then we briefly describe space–time CDMA signal processing. Signal Model. Assume M ⬎ 1 antennas. The received signal is a vector with M components and can be written as

x (t) =

pq (t − kT ) + n (t) sq (k)p

where sq(k) is the information bit stream for user q, and pq(t) is the composite channel for user q that embeds both the physical channel hq(t) (defined as in the TDMA case) and the spreading code cq(p) of length P: P−1
p=0

Conventional rake

Figure 8. The space–time rake receiver for CDMA uses a beamformer to spatially separate the signals, followed by a conventional rake.

SPACE–TIME ALGORITHMS FOR THE FORWARD LINK General Principles of Transmit Space–Time Processing In transmit space–time processing, the signal to be transmitted is combined in time and space before it is radiated by the antennas to encounter the channel. The goal of this operation is to enhance the signal received by the desired user, while minimizing the energy sent towards co-channel-users. Space– time processing makes use of the spatial and temporal signature of the users to differentiate them. It may also be used to preequalize the channel, that is, to reduce ISI in the received signal. Multiple antennas can also be used to offer transmit diversity against channel fading. Channel Estimation

q=1 k=−∞

pq (t) =

561

biner (see Fig. 8). The beamformer reduces the CCI at the rake input and thus improves the system capacity.

Space–Time Processing for Direct-Sequence Code Division Multiple Access

Q
∞


Beamformer

abruptly time-varying environment which makes the tracking of the channel or receiver coefficients difficult. As in the nonblind context, multiuser reception can be regarded as a two-stage signal equalization plus separation. Blind equalization of multiuser signals can be addressed using extensions of the aforementioned single-user techniques (HOS, CM, SOS, or subspace techniques). Blind separation of the multiuser signals needs new approaches, since subspace methods alone are not sufficient to solve the separation problem. In CDMA systems, the use of different user spreading codes makes this possible. In TDMA systems, a possible approach to signal separation consists in exploiting side information such as the finite-alphabet property of the modulated signals. The factorization Eq. (24) can then be carried out using alternate projections (see Ref. 10 for a survey). Other schemes include adjusting a space–time filter in order to restore the CM property of the signals.

x x x x x x

ANTENNA ARRAYS FOR MOBILE COMMUNICATIONS




hq t − cq (p)h

pT P



Space–Time Receiver Design. A popular single-user CDMA receiver is the RAKE combiner. The rake receiver exploits the (quasi)orthogonal codes to resolve and coherently combine the paths. It uses one correlator for each path and then combines the outputs to maximize the SNR. The weights of the combiner are selected using diversity combining principles. The rake receiver is a matched filter to the spreading code plus multipath channel. The space–time rake is an extension of the above. It consists of a beamformer for each path followed by a rake com-

The major challenge in transmit space–time processing is the estimation of the forward link channel. Further, for CCI suppression, we need to estimate the channels of all co-channelusers. Time-Division Duplex Systems. TDD systems use the same frequency for the forward link and the reverse link. Given the reciprocity principle, the forward and reverse link channels should be identical. However, transmit and receive take place in different time slots; hence the channels may differ, depending on the ping–pong period (time duration between receive and transmit phases) and the coherence time of the channel. Frequency-Division Duplex Systems. In frequency division duplex (FDD) systems, reverse and forward links operate on different frequencies. In multipath environment, this can cause the reverse and forward link channels to differ significantly. Essentially, in a specular channel, the forward and reverse DOAs and times of arrival (TOAs) are the same, but not the path complex amplitudes. A typical strategy consists in identifying the DOAs of the dominant incoming path, then using spatial beamforming in transmit in order to focus energy in

562

ANTENNA ARRAYS FOR MOBILE COMMUNICATIONS

these directions while reducing the radiated power in other directions. Adaptive nulls may also be formed in the directions of interfering users. However, this requires the DOAs for the co-channel-users to be known. A direct approach for transmit channel estimation is based on feedback. This approach involves the user estimating the channel from the downlink signal and sending this information back to the transmitter. In the sequel, we assume that the forward channel information is available at the transmitter. Single-User Minimum Mean Squared Error The goal of space–time processing in transmit is to maximize the signal level received by the desired user from the base station, while minimizing the ISI and CCI to other users. The space–time beamformer W is chosen so as to minimize the following MMSE expression:




 2 min E W HH Fq S q (k) − sq (k − d)2 + α W

Q


!

W HH Fk H

FH k W

where 움 is a parameter that balances the ISI reduction at the reference mobile and the CCI reduction at other mobiles. d is the chosen reconstruction delay. H Fq is the block-Toeplitz matrix [defined as in Eq. (14)] containing the coefficients of the forward link channel for the desired user. H Fk , k ⬆ q, denotes the forward channel matrix for the other users. Multiuser Minimum Mean Squared Error Assume that Q co-channel-users, operating within a given cell, communicate with the same base station. The multiuser MMSE problem involves adjusting Q space–time beamformers so as to maximize the signal level and minimize the ISI and CCI at each mobile. Note that CCI that originates from other cells is ignored here. The base communicates with user q through a beamformer Wq. All beamformers Wq, q ⫽ 1, . . ., Q, are jointly estimated by the optimization of the following cost function:

min

W q ,q=1,...,Q




 2 F  E W H q H q S q (k) − sq (k − d) 2

q=1

+α

Q


APPLICATIONS OF SPACE–TIME PROCESSING We now briefly review existing and emerging applications of space–time processing that are currently deployed in base stations of cellular networks.

k=1,k = q

(25)

Q


The basic approach in space–time coding is to split the encoded data into multiple data streams, each of which is modulated and simultaneously transmitted from a different antenna. Different choices of data-to-antenna mapping can be used. All antennas can use the same modulation and carrier frequency. Alternatively different modulation (symbol waveforms) or symbol delays can be used. Other approaches include use of different carriers (multicarrier techniques) or spreading codes. The received signal is a superposition of the multiple transmitted signals. Channel decoding can be used to recover the data sequence. Since the encoded data arrive over uncorrelated faded branches, diversity gain can be realized.

F FH WH k H qH q W k

! (26)

k=1,k = q

It turns out that the problem above decouples into Q independent quadratic problem, each having the form shown in Eq. (25). The multiuser MMSE problem can therefore be solved without difficulty. Space–Time Coding When the forward channel is unknown or only partially known (in FDD systems), transmit diversity cannot be implemented directly as in TDD systems, even if we have multiple transmit antennas that exhibit low fade correlation. There is an emerging class of techniques that offer transmit diversity in FDD systems by using space–time channel coding. The diversity gain can then be translated into significant improvements in data rates or BER performance.

Switched-Beam Systems Switched beam systems (SBSs) are nonadaptive beamforming systems that involve the use of four to eight antennas per sector at the base station. Here the system is presented for receive beamforming, but a similar concept can be used for transmit. The cell usually consists of three sectors that cover a 120⬚ angle each. In each sector, the outputs of the antennas are combined to form a number of beams with predesigned patterns. These fixed beams are obtained through the use of a Butler matrix. In most current cellular standards (including analog FDMA and digital FDMA–TDMA), a sector and a channel–time-slot pair are assigned to one user only. In order to enhance the communication with this user, the base station examines, through an electronic sniffer, the best beam output and switches to it. In some systems, two beams may be picked up and their outputs forwarded to a selection diversity device. Since the base also receives signals from mobile users in surrounding cells, the sniffer should be able to detect the desired signal in the presence of interferers. To minimize the probability of incorrect beam selection, the beam output is validated by a color code that identifies the user. In digital systems, beam selection is performed at baseband, after channel equalization and synchronization. SBSs provide array gain, which can be traded for an extended cell coverage. The gain brought by SBS is given by 10 log m, where m is the number of antennas. SBSs also help combat CCI. However, since the beams have a fixed width, interference suppression can occur only when the desired signal and the interferer fall into different beams. As a result, the performance of such a system is highly dependent on propagation environments and cell loading conditions. The SBS also experiences several losses, such as cusping losses (since there is a 2 to 3 dB cusp between beams), beam selection loss, mismatch loss in the presence of nonplanar wavefronts, and loss of path diversity. Reuse within Cell Since cellular communication systems are (increasingly) interference-limited, the gain in CCI reduction brought by the use of smart antennas can be traded for an increase in the

ANTENNAS

number of users supported by the network for a given quality of service. In current TDMA standards, this capacity improvement can be obtained through the use of a smaller frequency reuse factor. Hence, the available frequency band is reused more often, and consequently a larger number of carriers are available in each cell. Assuming a more drastic modification of the system design, the network will support several users in a given frequency channel in the same cell. This is called reuse within cell (RWC). RWC assumes these users have sufficiently different space–time signatures so that the receiver can achieve sufficient signal separation. When the users become too closely aligned in their signatures, space–time processing can no longer achieve signal recovery, and the users should be handed off to different frequencies or time slots. As another limitation of RWC, the space–time signatures (channel coefficients) of each user needs to be acquired with good accuracy. This can be a difficult task when the powers of the different users are not well balanced. Also, the propagation environment plays a major role in determining the complexity of the channel structure. Finally, angle spread, delay spread, and Doppler spread strongly affect the quality of channel estimation. As an additional difficulty, the channel estimation required in forward link space-time processing is made difficult in FDD systems. SUMMARY Smart antennas constitute a promising but still emerging technology. Space–time processing algorithms provide powerful tools to enhance the overall performance of wireless cellular networks. Improvements, typically by a factor of two in cell coverage or capacity, are shown to be possible according to results from field deployments using simple beamforming. Greater improvements can be obtained from some of the more advanced space–time processing solutions described in this paper. The successful integration of space–time processing techniques will however also require a substantial evolution of the current air interfaces. Also, the design of space–time algorithms must also be application- and environment-specific. BIBLIOGRAPHY 1. R. A. Monzingo and T. W. Miller, Introduction to Adaptive Arrays, New York: Wiley, 1980. 2. D. Johnson and D. Dudgeon, Array Signal Processing, Englewood Cliffs, NJ: Prentice-Hall, 1993. 3. J. D. Gibson (ed.), The Mobile Communications Handbook, Boca Raton, FL: CRC Press, 1996. 4. W. C. Jakes, Microwave Mobile Communications, New York: Wiley, 1974. 5. Proc. 4th Workshop on Smart Antennas in Wireless Mobile Communications, Sanford, CA: Center for Telecommunications and Information Systems Laboratory, Stanford University, 1997. 6. R. S. Kennedy, Fading Dispersive Communication Channels, New York: Wiley, 1969. 7. J. G. Proakis, Digital Communications, New York: McGrawHill, 1989. 8. S. U. H. Qureshi, Adaptive equalization, Proc. IEEE, 53: 1349– 1387, 1985.

563

9. S. Haykin (ed.), Blind Deconvolution, Englewood Cliffs, NJ: Prentice-Hall, 1994. 10. Special issue on blind identification and estimation, Proc. IEEE, to be published 1998. 11. M. K. Simon et al., Spread Spectrum Communications Handbook, New York: McGraw-Hill, 1994. 12. A. Paulray and C. B. Papadias, Space-time processing for wireless communications, IEEE Signal Processing Magazine, 14 (6): 49–83, Nov. 1997. 13. S. U. Pillai, Array Signal Processing, New York: Springer-Verlag, 1989.

AROGYASWAMI PAULRAJ DAVID GESBERT Stanford University

CONSTANTINOS PAPADIAS Lucent Technologies

ANTENNA NOISE. See RADIO NOISE.

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELECTRICAL...NICS%20ENGINEERING/62.%20Vehicular%20Technology/W7705.htm

}{{}}

●

HOME ●

ABOUT US ●

CONTACT US ●

HELP

Home / Engineering / Electrical and Electronics Engineering

Wiley Encyclopedia of Electrical and Electronics Engineering Cellular Radio Standard Article William C. Y. Lee1 1AirTouch Communication, Walnut Creek, CA Copyright © 1999 by John Wiley & Sons, Inc. All rights reserved. DOI: 10.1002/047134608X.W7705 Article Online Posting Date: December 27, 1999 Abstract | Full Text: HTML PDF (152K)

●

●

● ●

Recommend to Your Librarian Save title to My Profile Email this page Print this page

Browse this title ●

Search this title Enter words or phrases

Abstract The sections in this article are History of the Cellular Radio Systems Mobile Radio Environment: A Difficult Environment For Cellular Radio Systems Reasons for Digital Cellular Requirement For Cellular and Pcs Digital Modulations and Multiple Access The Specifications Of Different Cellular/PCS Systems About Wiley InterScience | About Wiley | Privacy | Terms & Conditions Copyright © 1999-2008John Wiley & Sons, Inc. All Rights Reserved.

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELE...ENGINEERING/62.%20Vehicular%20Technology/W7705.htm15.06.2008 12:43:53

❍

❍ ❍

Advanced Product Search Search All Content Acronym Finder

126

CELLULAR RADIO

nels. The channel bandwidth is 30 kHz. Mobile cellular telecommunications systems (2) leave two unique features: 1. First, they invoke the concept of frequency reuse for increasing spectrum efficiency. The same set of frequency channels can be assigned to many cells. These cells are called co-channel cells. The separation between two cochannel cells is engineered by the D/R ratio (see Fig. 1), where D is the co-channel cell separation and R is the cell radius. A 4-mile cell implies R ⫽ 4 miles. The D/R ratio is characteristic of a cellular system. If the D/R ratio is high, the voice quality is improved by reducing the system’s user capacity. 2. A second feature, handing off communications from one frequency to another occurs when a mobile unit enters a new cell. The scheme is called a handoff in North America and a handover in Europe. The system handles this operation automatically, and the users do not need to intervene. A good handoff algorithm can reduce both the call drop rates and interference. In general, there are two kinds of handoffs: (1) Soft handoffs, which implies making a new connection before breaking the old one; and (2) Hard handoffs, which is breaking the old connection before making the new one.

CELLULAR RADIO The cellular radio system is sometimes called a mobile phone system or a car phone system. Due to the daily needs of subscribers, cellular systems have expanded considerably all over the world. This article discusses the history of cellular systems and the difficulty of deploying them in the mobile radio environment, elaborating on employing digital cellular systems, Personal Communication Services (PCS) mobile satellite systems, and the future IMT-2000 system.

The first installation of a cellular system occurred in Tokyo in 1979, using a minor modification of AMPS. The first AMPS cellular system installed in the United States took place in Chicago in 1983. Analog cellular systems are in use over most of the world, employing different versions of AMPS: In Japan the Nippon Telephone and Telegraph (NTT) AMPS system, in the UK, the Total Access Communications System (TACS); and in northern Europe, the Nordic Mobile Telephone (NMT).

D = 4.6R

R 1 1

HISTORY OF THE CELLULAR RADIO SYSTEMS

Start-Up Period (From 1964 to 1987). In 1964, AT&T Bell Labs actively developed a high-capacity mobile radio phone system called Advanced Mobile Phone Service (AMPS) (1), which is an analog frequency modulation (FM) system. The system consists of many so-called cells. Each cell has one or multiple transceivers. Because of the cell formation, the system is referred to as a cellular system. In the analog AMPS system, mobile units are compatible with all the cellular systems operating in the United States, Canada, and Mexico. A spectrum of 50 MHz (limited to 825 MHz to 849 MHz for mobile transmissions and 869 MHz to 896 MHz for base station transmissions) is shared by two cellular system providers in each market (city). Each one provider operates over a bandwidth of 25 MHz in a duplex fashion (using 12.5 MHz in each direction between cell sites and mobile units). There are 416 channels, comprising 21 setup channels and 395 voice chan-

1

D

Analog System

1 1

3

2 4

6

5 7

1

1

Figure 1. Hexagonal coils in an AMPS system. R ⫽ radius of cells, D ⫽ minimum separation of co-channel cells, q ⫽ D/R ⫽ 4.6, K ⫽ number of cells in a cluster ⫽ 7. Clusters are indicated, and the six cells that effectively interfere with cell 1 are numbered 2 through 7. The shaded cells are co-channel cells.

J. Webster (ed.), Wiley Encyclopedia of Electrical and Electronics Engineering. Copyright # 1999 John Wiley & Sons, Inc.

CELLULAR RADIO

The major difference is their reduced channel bandwidths of 25 kHz instead of 30 kHz as in AMPS. Mature Period (From 1987 to 1992). From 1987 to 1992, the 90 MSA (metropolitan statistical area) markets, as well as most of the 417 RSA (rural service area) markets, had cellular operations in the United States. The number of subscribers reached 1 million. The cell-split (reducing the size of cells) technique and dynamic frequency assignment were applied to increase the user capacity. When the cell radius R is less than half a kilometer, the cell is called a microcell. In such small cells it is harder to reduce the so-called co-channel interference in order to increase capacity, requiring special technological approaches called microcell technology. The world was also becoming more aware of the potential future markets. Suddenly, finding the means to increase capacity became urgent. Digital System Introduction Period. In 1987, the capacity of the AMPS cellular system started to show its limitations. The growth rate of cellular subscribers far exceeded expectations. In 1987, the Cellular Telecommunication Industry Association (CTIA) formed a subcommittee for Advanced Radio Technology to study the use of a digital cellular system (3) to increase capacity. At that time, the Federal Communications Commission (FCC) had clearly stated that no additional spectrum would be allocated to cellular telecommunications in the foreseeable future. Therefore, the existing analog and forthcoming digital systems would have to share the same frequency band. In December 1989, a group formed by the Telecommunication Industry Association (TIA) completed a draft of a digital cellular standard. The digital AMPS, which must share the existing spectrum with the analog AMPS, is a duplex time-division multiple-access (TDMA) system. The channel bandwidth is 30 kHz. There are 50 TDMA frames per second in each channel. Three or six time slots per frame can serve three calls or six calls at the same time in one channel. The speech coding rate is 8 kb per second. An equalizer is needed in the receiver to reduce the intersymbol interference that is due to the spread in time delay caused by the dispersed time arrival of multipath waves. The North American TDMA system was first called IS-54 by the TIA. Later, the system was modified and renamed as IS-136. During this period, not all mobile telephone systems in Europe were compatible. A mobile phone unit working in one country could not operate in another country. In 1983, in response to the need for compatibility, a special task force called the Special Mobile Group (4) was formed among European countries to develop a digital cellular system called GSM (group of special mobile systems) in 1994, then changed to stand for global system for mobile communications. The operating principles of the GSM system resemble those of the AMPS in radio operation, but the system parameters are different; this will be described later. In the United States, in addition to the TDMA being considered above, another particularly promising technology is code-division multiple access (CDMA) (3). It is a spread spectrum technique with a bandwidth of 1.25 MHz. The maximum

127

number of traffic channels is 55. This CDMA system is called IS-95 or cdmaOne. There are three mobile data systems in the United States: namely Ardis, operated by IBM/Motorola; Ram, operated by Ericsson; and CDPD (Cellular Digital Packet Data) system. The transmission rates for all data systems are around 8 kbps. Only CDPD operates in the cellular spectrum band. The Future. Starting in 1996, the so-called PCS systems were deployed. They were cellular-like systems, but operated in the 1.8 GHz band in Europe and the 1.9 GHz range in North America. In Europe, the so-called DCS-1800 PCS systems were endorsed, which are based on the GSM system. In the United States, the PCS had three versions: DCS-1800 (a GSM version), TDMA-1900 (IS-136 version), and CDMA-1900 (IS-95 version). The PCS could have six operational licenses (A, B, C, D, E, F) in each city. Therefore, more competitors would be in the mobile phone services business. In addition, the mobile satellite systems that use the LEO concept (low earth orbit) were deployed. Iridium (66 satellites) and Globalstar (48 satellites) were launched at 900 km and 400 km altitudes, respectively. These systems can integrate with cellular systems and enhance cellular coverage domestically and roam internationally as a global system. Other LEO systems are also in the development stage. There is a special LEO system called Teledesic that will be operating at 26 GHz with 840 satellites in orbit. This system is used for wideband data and video channels to serve subscribers in a high capacity network. A future cellular system, called the International Mobile Telephone (IMT-2000) system, is now in the planning stage. A universal cellular standard (or PCS) system with high capacity and high transmission rate may be realized by the year 2002. MOBILE RADIO ENVIRONMENT: A DIFFICULT ENVIRONMENT FOR CELLULAR RADIO SYSTEMS Understanding the Mobile Radio Environment The Limitations of Nature. In the mobile radio environment, there are many attributes that limit the system performance for wireless communication. In the past, there were attempts to adapt digital equipment such as data modems and FAX machines used for wireline to celluar systems. The data engineers at that time only realized the blanking and burst interruption in the voice channel as a unique feature of handoffs and power control. They modified data signaling by overcoming the impairments caused by blanking and burst signaling interruption. This modified data modem did not work as expected in the cellular system. Actually, the blanking and burst interruption scheme was not the sole cause of the inadequate data transmission and would have been relatively easy to handle. But without entirely understanding the impairments, the unexpected poor performance could not be offset by merely overcoming the blanking and burst signaling impairment. Choosing the Right Technologies. In designing radio communication systems, there are many different technologies, and among them no single technology is superior to the others. Choosing a technology depends on real conditions in the envi-

128

CELLULAR RADIO

ronment of a particular communication. In satellite communication or microwave link transmission, the radio environments are different from that of the mobile radio environment. There are many good technologies that work in satellite communication and microwave link transmission, but they may not be suitable for the terrestrial mobile radio environment. Therefore, choosing the right technology must depend on the transmission environment.

h1

;;;; ;;;;

he h eb a h ec

A

C

B

Figure 3. Effective antenna heights at base station based on different locations of mobile stations.

Description of the Mobile Radio Environment The mobile radio environment is one of the most complex ones among the various communication environments. Nature Terrain Configuration. Because the antenna height of a mobile unit or a portable unit is very close to the ground, the ground-reflected wave affects the reception of the signal from the transmitting site via the direct path. The free space loss is 20 dB/dec (dec stands for decade, a period of ten) or, in other words, it is inversely proportional to the distance d⫺2. However, in the mobile radio environment, due to the existing ground-reflected wave and the small incident angle ␪, as shown in Fig. 2, the total energy of the ground-reflected wave is reflected back to space. Due to the nature of electromagnetic waves, when the wave hits the ground, the phase of the wave changes by 180⬚. Therefore, at the mobile, the direct wave and reflected wave cancel each other instead of adding constructively. As a result, the signal that is received becomes very weak. A simple explanation is as follows: If the path length of the direct wave is d, and the path length of the reflected wave is d ⫹ ⌬d. Then the received power of the two combined waves is proportional to d⫺4 as demonstrated below Pr ∝


1 d

−

1 d +
d

2

=





d d(d +
d)

2

=

(
d)2 d4

(1)

where ⌬d is assumed to be much less than d and ⌬d is a function of the antenna height h1 at the base station. From Eq. (1), the mobile radio path loss follows the inverse fourth power rule or 40 dB/dec, and the antenna height gain follows the second power rule or 6 dB/oct. In the mobile radio environment, the average signal strength at the mobile unit varies due to the effective antenna height he at the base station measured from the mobile unit location. Since the mobile unit is traveling, the effective antenna height is always changing as a function of terrain undulations and so is the average signal strength. This phenomenon is shown in Fig. 3. This two-wave (direct wave and ground-reflected wave) model is only used to explain the propagation loss of 40 dB/ dec in the mobile radio environment, not the multipath fading.

Base-station antenna

;;;;; Gr

h1

ou

Dir

nd

-re

fle

ect

cte

w av

dw

e

av

e

Mobile station antenna

h2

Figure 2. Two-wave propagation model.

Man-Made Effect Man-Made Communities. These can be classified as metropolitan areas, urban areas, suburban areas, and open areas, and so on. The distribution of buildings and homes depends on the population size. The reception of the signal is affected by the differences in man-made communities and results in different propagation path loss. Man-Made Structures. Different geographical areas use different construction materials, different types of construction frames, and different sizes of buildings. Cities such as Los Angeles, San Francisco, and Tokyo are in earthquake zones and follow earthquake construction codes. The signal reception in those cities is different from that in others. Man-made structures will affect the propagation path loss and multipath fading due to reflection and the signal penetration through the buildings. Man-Made Noise. This can be classified into two categories: Industrial noise or the automotive ignition noise. The high spikes in automotive ignition or in machines are like impulses in the time domain; their power spectrum density will cover a wide spectrum in the industrial frequency domain. At 800 MHz, automotive ignition noise is determined by the number of vehicles. For a traffic volume from 100 cars/h to 1000 cars/ h, the noise figure increases 7 dB. As the application of Ultra High Frequency (UHF) devices and microwave systems increase, so does the noise pollution for cellular systems. As we will mention later, a communication system is designed to maintain the minimum required carrier-to-interference ratio (C/Is). The interference I may, under certain circumstances, be included in noise the N. If the interference level is higher, the level of the carrier, C should be also higher in order to meet the (C/I)s requirement. This means that when the manmade noise level is high, either the transmission power at the base station should be increased or the cell size must be reduced. Moving Medium. If the mobile unit is in motion, the resulting signal from multipath waves at one location is not the same at another, thus the mobile receiver observes an instantaneous fluctuation in amplitude and phase. The amplitude change is called Rayleigh phase, and the phase change is a uniformly distributed process, or random FM in FM systems. The signal fading can be fast or slow depending on the speed of the vehicle. When the vehicle speed is slow, the average duration of fading is long. This average fading duration can be, for example, 7 ms at ⫺10 dB below the average level when the vehicle speed is 24 km/h at a propagation frequency of 850 MHz. In an analog system, a fade duration of 7 ms does not affect the analog voice; the ear cannot detect these short

CELLULAR RADIO

fades. However, the fade duration of 7 ms is long enough to corrupt the digital (voice and data) transmissions. At a transmission rate of 20 kbps, 140 bits will sink in the fade. Furthermore, the vehicle speed of all the users is not constant, and the use of interleaving and channel coding to protect the information bits is very difficult. Furthermore, voice communication is operating in real time unlike data transmission which can be in any time-delay fashion. Many schemes used by data communication cannot be used for digital voice communication. Dispersive Medium. Because of human-made structures, the medium becomes dispersive. In a dispersive medium, two phenomena occur. One is time delay spread and the other is selective fading. The time delay spread is caused by a signal transmission from the base station reflected from different scatterers and arriving at the mobile unit at different times. In urban areas, the mean time delay spread, ⌬, is typically 3 애s; in suburban areas, ⌬ is typically 0.5 애s. In an open area, ⌬ is typically 0.2 애s, and in an in-building floor, ⌬ is around 0.1 애s or less. These time delay spreads do not affect the analog signal because the ear cannot detect the short delay spread. However, in a digital system when a symbol (bit) is sent, many echoes arrive at the receiver at different times. If the next symbol is sent out before the first one dies down, intersymbol interference occurs. The dispersive medium also causes frequency selective fading (Fig. 4). The selective fading will not hurt the moving receiver because when the mobile unit is moving, only the average power is considered. Then, in order to make a mobile phone call when the mobile unit is at a standstill, it usually requires that all the signal strengths from four frequencies have two strong setup channels and two strong voice channels. A pair of frequencies is formed by a channel carrying a call on both a forward link and a reverse link. When the mobile unit is moving, the average power of the four frequencies is the same. Then we base our quality estimates on one (C/I)s value. But when the mobile unit is still, the signals of four frequencies at one location are different due to frequency selective fading. Unless all four frequencies are above the acceptable threshold level, the call cannot be connected. Concept of C/I In designing high-capacity wireless systems, the most important parameter is the carrier-to-interference ratio (C/I). The C/I ratio and the D/R ratio are directly halted. The D/R ratio

Mean ∆ =

∆

0.1 0.2 0.5 3.0

µs µs µs µs

In-building Open area Surburban Urban

s(t) in dB

10 9 8 7 6 q 5 4 3 2 1 0

0

5

10 15 20 (C/ I)s in dB

25

129

30

Figure 5. Relationship between q and (C/I)s [Eq. (3)].

is determined by the C/I ratio. Usually, with a given received signal level C, the lower the interference level, the higher the C/I ratio and hence the quality improves. There is a specific C/I level, namely (C/I)s, that the system design criterion is based on. We may derive the relationship between C/I and D/R as follows: Assume that the first tier of six co-channel interference cells is the major cause for the interference I. Based on the 40 dB/dec propagation rule, we obtain

C C = 6 I


≈ Ii

C R−4 (D/R)4 = = −4 6 · Ii 6·D 6

(2)

i=1

A general equation of the co-channel interference reduction factor q can be expressed, from Eq. (2), as q = (D/R)s = (6(C/I)s )1/4

(3)

where (C/I)s is obtained from a subjective test corresponding to the required voice (or data) quality level, as mentioned previously. Equation (3) is plotted in Fig. 5. The (C/I)s ratio is chosen according to either the required voice or data channel quality. The Predicted Signal-Strength Models Since the (C/I)s is a system design parameter, system planning engineers would like to use an effective model to predict both C and I in a given area. There are two different prediction models. One predicts the average signal strength along the radio path based on the path loss slope. The Okumura and Hata’s model (5,6) represents these types of models. The other predicts the local-mean signal strength along a particular mobile path (street or road) based on the particular terrain contour. Lee’s model (7) represents these types of models. REASONS FOR DIGITAL CELLULAR Compatibility in Europe

t1

Time (at receiving end)

Figure 4. Time delay spread ⌬ at the receiving end when transmitting one bit in a dispersive medium.

Again, due to the lack of a standard mobile radio system in Europe during the early 1980s, the mobile phone unit used in each country could not be used in other countries. Starting in 1982, ETSI (European Telecomms Standard Institute) formed a group called the Group of Special Mobile to construct an international mobile radio system called GSM for Europe.

130

CELLULAR RADIO

The system chosen was to be a digital system using TDMA for the access scheme. The GSM advanced intelligent network (AIN) was adopted from the wireline telephone network. GSM was the first digital mobile phone system in the world.

needed one standard for the entire North American cellular industry. The Advantages of a Digital System Digital systems offer the following advantages:

Capacity in North America Frequency spectrum is a very limited resource commonly shared by all wireless communications. Among the wireless communications systems, cellular is the most spectral efficient system where the so-called spectral efficiency is related to the number of traffic channels per cell. From this number, we can derive the Erlang/cell ratio, which translates to Erlang/km2, or the number of traffic channels/km2 based on the traffic model and the size of the cells. However, a spectrum of only 50 MHz has been allocated to cellular operators in the United States. Furthermore, since two operators are licensed in each market, the spectrum of 50 MHz must be split in two. Therefore, system trunking efficiency is reduced and interference caused by an operator in one market often contaminates the other operator’s allocated spectrum. Furthermore, manufacturing companies were always considering lowering the cost of cellular units and increasing sales volume. As a result, the specification of cellular units could not be kept tight, and thus more interference prevailed. Once interference increased and could not be controlled by cellular operators, both voice quality and system capacity decreased. In 1987, the top 10 US markets were already feeling the constraints of channel capacity; they would not be able to meet the market demand in the future. The solution for this increasing need for high capacity was to go digital (1–3). Going digital was the best solution because of the nature of the digital waveform. If system compatibility is not an issue, the top 10 US markets might and could go digital by themselves. However, for the sake of compatability the United States

1. The digital waveform is discrete in nature. Therefore, the digital waveform can be regrouped easily for transmission needs. 2. Digital transmission is less susceptible to noise and interference. 3. Digital modulation can confine the transmitted energy within the channel bandwidth. 4. Digital equipment may consume less battery power, and hence may reduce equipment weight. 5. Digital systems can provide reliable authentication and privacy (encryption). REQUIREMENT FOR CELLULAR AND PCS In 1996, the Telecommunications Act Bill was passed by the U.S. Congress and stated, in simple terms, that everyone could get into everyone’s business. Cellular service is moving toward digital and is trying to compete with PCS. The PCS spectrum was auctioned in early 1996. There are wideband PCSs and narrowband PCSs (see Fig. 6). The spectrum of wideband PCS is allocated at 1900 MHz in order to operate the same technology as the cellular system. The spectrum of narrowband PCS is allocated at 900 MHz and is used for twoway paging. The joint requirements of both cellular and PCS are as follows: From the end users perspective: The PCS and cellular units should be light in weight, small in size, and have

Wideband PCS – for cellular- like systems Base Rv

Base Tx

15 5 15 5 5 15 5 10 5 15 5 15 5 5 15 A D B EF C UD A D B EF C 1800 UV unlicensed voice UD unlicensed data

1850 1870

1900 UV 1930 1950 1970 1990

Narrowband PCS – for two-way paging systems Five 50 kHz channels paired with 50 kHz channels 901.00 .05

.1

.15

.20 901.25

940.00 .05

.1

Three 50 kHz channels paired with 12.5 kHz channels 901.75 901.7875 .7625 .7750

930.40 .45

Three 50 kHz unpaired channels Figure 6. Spectrum allocated for wideband PCS and narrowband PCS.

940.75

.80

.85

940.90 MHz

.50

930.55 MHz

.15

.20 940.25 MHz

CELLULAR RADIO

long talk-time capabilities without battery recharging and good quality in voice and data. The unit should be employable for initiating and receiving calls anywhere using any telephone feature. The important requirement of PCS and cellular is to please the vast majority of subscribers who always prefer to carry a single unit, not many units. This unit can be classified according to the different grades of service. From the system providers perspective: The PCS should provide full coverage and large system capacity to serve end users. An end user unit ideally should be serviced by one system with different grades of service and unless there are natural limitations by the various personal communication environments (such as mobile vehicle, pedestrian, and indoor public communication). Then one end user unit should be capable of accessing more than one system.

131

system. It is a low-risk system to develop but was voted down by the industry in 1987. FDMA is not suitable for high-speed data transmission. TDMA was first developed in Europe and is called GSM. TDMA has been developed in North America. For the ADC (American Digital Cellular system), CDMA needs more advanced technology and is relatively harder to implement than the other two multiple access schemes, especially in the mobile radio environment. However, the improved user capacity of CDMA has given the cellular industry the incentive to develop this system. Therefore, digital transmission in the mobile radio environment has only two competing multiple accesses. The North America selected TDMA based on the influence from the European GSM.

THE SPECIFICATIONS OF DIFFERENT CELLULAR/PCS SYSTEMS Analog Systems

DIGITAL MODULATIONS AND MULTIPLE ACCESS Digital Modulation Schemes Digital modulation schemes can be selected to confine the transmitted energy of a digital voice signal in a given frequency bandwidth while transmitting in a mobile radio environment. The information may have to be modulated by signal phases or frequencies, rather than amplitudes, because the multipath fading impairs the signal amplitude.

Each traffic channel in an analog system uses two frequencies, one receiving and one transmitting frequency. In general, we often refer to ‘‘a 30 kHz channel’’ when we really mean a bandwidth of 30 kHz on one of two frequencies. Therefore, the total occupied spectrum for each traffic channel is 60 kHz. There are three analog systems: The AMPS from North America, the NTT system from Japan, and the TACS system from the UK. Their specifications are listed in Table 1.

Multiple Access

TDMA Systems

Digital transmission can use time-division multiple access (TDMA), frequency-division multiple access (FDMA), or codedivision multiple access (CDMA), but in analog transmission only FDMA can be used. FDMA provides many different frequency channels, where each is assigned to support a call. TDMA means chopping a relatively broadband channel over time into many time slots. Each time slot is assigned to support a call. CDMA means generating many different code signatures over a long code-bit stream channel, where each code signature is assigned to convey a call. FDMA is a narrowband

The following TDMA systems can be grouped into two different duplexing techniques, FDD and TDD: FDD (frequency division duplexing). Each traffic channel consists of two operational frequencies. The analog system can only use a FDD system, whereas the digital system has a choice. GSM. The term GSM often implies DCS-1800 and DCS1900 services. They are in the same family, only the

Table 1. Large-Capacity Analog Cellular Telephones Used in the World Japan System transmission frequency (MHz): Base station Mobile station Spacing between transmission and receiving frequencies (MHz) Spacing between channels (kHz) Number of channels Coverage radius (km) Audio signal: Type of modulation Frequency deviation (kHz) Data transmission rate (kb/s) Message protection

North America

England

870–885 925–940 55

869–894 824–849 45

917–950 872–905 45

25, 12.5 600 5 (urban area) 10 (suburbs)

30 832 (control channel 21 ⫻ 2) 2–20

25 1320 (control channel 21 ⫻ 2) 2–20

FM ⫾5 0.3 Transmitted signal is checked when it is sent back to the sender by the receiver.

FM ⫾12 10 Principle of majority decision is employed.

FM ⫾9.5 8 Principle of majority decision is employed.

Source: Report from International radio Consultative Committee (CCIR) 1987.

132

CELLULAR RADIO

Table 2. Physical Layer Parameters of GSM Parameter Radio carrier bandwidth TDMA structure Time slot Frame interval Radio carrier number Modulation scheme Frequency hopping Equalizer Frequency hop rate Handover a

Specifications 200 kHz 8 time slots per radio carrier 0.577 ms 8 time slots ⫽ 4.615 ms 124 radio carriers (935–960 MHz downlink, 890–915 MHz uplink) Gaussian minimum shift keying with BTa ⫽ 0.3 Slow frequency hopping (217 hops/s) Equalization up to 16 애s time dispersion 217 hops/s Hard handover

BT ⫽ Bandwidth ⫻ Time.

7. No equalizer implemented 8. Handoff 9. Transmission rate—6.5 kbps/slot 10. Forward error correction—3 kbps 11. Dispatch capability TDD (time division duplexing). Transmission and reception are shared by one frequency. Certain time slots are for transmission and certain time slots are for reception. CT-2 (Cordless Phone Two). CT-2 was developed by GPT Ltd. in the UK for so-called Telepoint Applications. Phone calls can be dialed out but cannot be received. The transmission parameters for CT-2 are as follows: 1. Full duplex system

carrier frequencies are different. We list the physical layer parameters in Table 2. NA-TDMA (North American-TDMA). NA-TDMA, sometimes called ADC, is North America’s standard system. It incorporates both 800 MHz and 1900 MHz system versions. The network follows philosophy of the GSM intelligent network. The physical layer is shown in Table 3. The PDC (personal digital cellular) system. This system was developed in Japan and is very similar to the NATDMA system, but its radio carrier bandwidth is 25 kHz. IDEN (Integrated Digital Enhanced Network). This system was developed by Motorola. It was called MIRS (Mobile Integrated Radio System); then Motorola modified the system and renamed it IDEN. This system uses the SMR (Special Mobile Radio) band, which is specified by Part 90 of FCC CFR code of Federal regulations in the private sector. The system now can be used for cellularlike commercial services. The physical parameter system is as follows: 1. Full-duplex communication system 2. Frequency—806 to 824 MHz (mobile transmitter), 851 MHz 3. Channel bandwidth—25 kHz 4. Multiple access—TDMA 5. Number of time slots—6 6. Rate of speech coder—VSELP (Vector sum excitation linear predicted)

2. Voice coder—32 kbps adaptive differential pulsecode modulation (ADPCM). 3. Duplexing—TDD. The portable and base units transmit and receive on the same frequency but different time slots. 4. Multiple access—TDMA-TDD, up to four multiplexed circuits 5. Modulation—앟/4 DQPSK differential QPSK, roll-off rate ⫽ 0.5 6. Data rate—192 ksps (192 kilo symbols per second or 384 kbps) 7. Spectrum allocation—1895 MHz to 1918.1 MHz. This spectrum has been allocated for private and public use. 8. Carrier frequency spacing—300 kHz PHS (Personal Handy-Phone System). It was developed in Japan. Now there are three operators: NTT, STEL, and DDI. The system serves for the low tier subscribers, like teenagers. There are around seven million customers. The specifications for transmission parameters are as follows: 1. Full duplex system 2. Voice coder—32 kpbs adaptive differential pulsecode modulation (ADPCM). 3. Duplexing—TDD. The portable and base units transmit and receive on the same frequency but different time slots. 4. Multiple access—TDMA-TDD, up to four multiplexed circuits. 5. Modulation—앟/4 DQPSK, roll-off rate ⫽ 0.5

Table 3. Physical Layer of NA-TDMA Parameter Radio carrier bandwidth TDMA structure Time slot Frame interval Radio carrier number Modulation scheme Equalizer

6. Data rate—192 ksps (or 384 kbps). Specifications

30 kHz 3 time slots per radio carrier 6.66 ms 20 ms 2 ⫻ 416 (824–849 MHz reverse link, 869–894 MHz forward link) 앟 ⫺ DQPSK 4 Equalization up to 60 애s time dispersion

7. Spectrum allocation—1895 MHz to 1918.1 MHz. This spectrum has been allocated for private and public use. 8. Carrier frequency spacing—300 kHz. Another system called PACS (Personal Access Communication Systems) (3) is in the same system family as PHS. DECT (Digital European Cordless Telephone) (3). DECT is a European standard system for slow motion or in-

CELLULAR RADIO

133

Table 4. Comparative Low-Earth-Orbiting Mobile Satellite Service Applications System Characteristics Number of satellites Constellation altitude (NM) Unique feature

Loral/QUALCOMM 48 750

Motorola IRIDIUM 66 421

TRW ODYSSEY 12 5600

Constellation ARIES (b) 48 550

Ellipsat ELLIPSO 24 1767 ⫻ 230

Transponder

Transponder

Transponder

Transponder

Circuit capacity (US) Signal modulation Gateways in US Gateway spectrum band Coverage

6500 CDMA 6 C-band existing Global

Onboard processing 3835 TDMA 2 New Ka band Global

4600 CDMA 2 New Ka band Global

100 FDMA/CDMA 5 Unknown Global

1210 CDMA 6 Unknown Northern hemisphere

building communications. Its system structure is as follows: 1. Duplex method—TDD 2. Access method—TDMA 3. RF (radio frequency) power of handset—10 mW 4. Channel bandwidth—1.728 MHz/channel 5. Number of carriers—five (a multiple-carrier system) 6. Frequency—1800 to 1900 MHz DECT’s characteristics are as follows: 1. Frame—10 ms 2. Time slots—12 3. Bit rate—38.8 kb/slot 4. Modulation—GFSK (Gaussian Filtered FSK) 5. Handoff—Yes CDMA Systems CDMA is another multiple-access scheme using different orthogonal code sequences to provide different call connections. It is a broadband system and can be classified by two approaches: (1) Frequency Hopping System approach (3), and (2) direct sequence system approach (3). The commercial CDMA system applies the direct sequence approach. Developed in the United States, it is called the IS-95 Standard System. The first CDMA system was deployed in Hong Kong and then in Los Angeles in 1995. CDMA is a high-capacity system. It has been proven, theoretically, that CDMA system capacity can be 20 times higher than analog capacity. In a CDMA system, all the cells share the same radio carrier in an operating system. The handoff from cell to cell is soft (i.e., not only is the frequency kept unchanged, but the cell is connected in both the old cell and the new cell in the handoff region). The IS-95 CDMA is now called cdmaOne. The CDMA radio specifications are as follows: 1. 2. 3. 4. 5. 6. 7. 8.

CDMA shares the spectrum band with AMPS Total number of CDMA radio carriers is 18. Radio carrier bandwidth is 1.2288 MHz. Pseudo noise (PN) chip rate is 1.2288 Mcps. Pilot channel is one per radio carrier. Power control step is 1 dB in 1 ms. Soft handoffs are used. Traffic channels are 55 per each radio carrier.

9. Vocoder is Qualcomm Code Excited Linear Prediction (QCELP) at a variable rate. 10. Modulation is Quaternary Phase Shift Keying (QPSK). 11. Data frame size is 20 ms. 12. Orthogonal spreading is 64 Walsh functions. 13. Long PN code length is 242 ⫺ 1 chips 14. Short PN code length is 215 ⫺ 1 chips. Mobile Satellite Systems Mobile satellite systems (MSS) are used to enhance terrestrial radio communication, either in rural areas or in terms of global coverage. Therefore, MSS becomes, in a broad sense, a PCS system. By taking advantage of reduced transmitting power and short time delays, the low earth orbit (LEO) systems are being developed. However, there is a drawback. Each LEO system needs many satellites to cover the earth. There are many LEO systems, as shown in Table 4. There is also another LEO referred to as the Teledesic system, which will operate at 24 GHz with a spectrum band of 500 MHz. This LEO system is not just for enhancing cellular or PCS coverage, but also can replace the terrestrial long-distance telephone network in the future. IMT-2000 Since the CDMA One system has been successfully deployed in Korea and the United States, in mid-1997 the European countries under the auspices of the so-called (ETSI) European Telecom. Standard Institute, Japan (ARIB) Association of Radio Industrial and Business, and the United States (TIA) Telecom Industrial Asso. began planning a universal singlestandard system for the so-called IMT-2000 (International Mobile Telephone—Year 2000). There are three general proposals. The proposals disagree on many issues, but they do agree on the following general guideline principles: 1. 2. 3. 4. 5. 6.

Use wideband CDMA (WCDMA). Use direct sequence as spread spectrum modulation. There should be a multiband, single mobile unit. The standard band should be 5 MHz. There is a need for international roaming. There should be given up IPR (intellectual property right) issues in developing the new global system among all the international vendors.

134

CEMENT INDUSTRY

The IMT-2000 system will require a great deal of compromise in selecting technologies due to the political differences in the international standards bodies. The formal IMT-2000 system will be adapted by the ITU (International Telecommunication Union). It remains uncertain if there will be a single universal IMT-2000 by the year 2000. BIBLIOGRAPHY 1. S. H. Blecher, Advanced mobile phone services, IEEE Trans. Veh. Technol., VT-29: 238–244, 1980. 2. W. C. Y. Lee, Mobile Communications Design Fundamentals, 2nd ed., New York: Wiley, 1993. 3. W. C. Y. Lee, Mobile Cellular Telecommunications, Analog and Digital Systems, New York: McGraw-Hill, 1995. 4. B. J. T. Malliner, An overview of the GSM system, Conf. Proc., Digital Cellular Radio Conference, Hagen FRG, October 1988. 5. Y. Okumura et al., Field strength and its variability in VHF and UHF land-mobile radio service, Rev. Electr. Commun. Lab., 16: 825–873, 1968. 6. M. Hata, Emperical formula for propagation loss in land mobile radio services, IEEE Trans. Veh. Technol., VT-29: 317–325, 1980. 7. W. C. Y. Lee, Spectrum efficiency in cellular, IEEE Trans. Veh. Technol., 38: 69–75, 1989. Reading List J. Gabion, The Mobile Comms Handbook, IEEE Press, New York, 1985. M. Morly et al., The GSM System.

WILLIAM C. Y. LEE AirTouch Communication

CELLULAR STANDARDS. See MOBILE TELECOMMUNICATIONS STANDARDS.

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELECTRICAL...NICS%20ENGINEERING/62.%20Vehicular%20Technology/W7714.htm

}{{}}

●

HOME ●

ABOUT US ●

CONTACT US ●

HELP

Home / Engineering / Electrical and Electronics Engineering

Wiley Encyclopedia of Electrical and Electronics Engineering Ground Transportation Systems Standard Article Adam Szelag1 and Leszek Mierzejewski1 1Warsaw University of Technology Copyright © 1999 by John Wiley & Sons, Inc. All rights reserved. DOI: 10.1002/047134608X.W7714 Article Online Posting Date: December 27, 1999 Abstract | Full Text: HTML PDF (803K)

●

●

● ●

Recommend to Your Librarian Save title to My Profile Email this page Print this page

Browse this title ●

Search this title Enter words or phrases

Abstract The sections in this article are Railroads High-Speed Railroads Maglev Conventional Railroads Urban Transport Systems Electric Transport Power Supply Systems Moving Load on Power Supply Due to Electric Vehicles Electrical Parameters of the Traction Supply Network Overhead Catenaries Systems Approach to Electrified Ground Transportation Choice of Electric Traction System

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20E...GINEERING/62.%20Vehicular%20Technology/W7714.htm (1 of 2)15.06.2008 12:44:13

❍

❍ ❍

Advanced Product Search Search All Content Acronym Finder

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELECTRICAL...NICS%20ENGINEERING/62.%20Vehicular%20Technology/W7714.htm

Sizing of Power Supply Equipment and Installation Modeling and Simulation Methods as a Tool for Analysis of an Electrified Ground Transport System Control and Signaling in Transportation Systems Impact of an Electrified Ground Transport System on the Environment and the Technical Infrastructure Impact of Traction Substations on Power Utility Systems Electromagnetic Compatibility of Electrified Ground Transport Systems About Wiley InterScience | About Wiley | Privacy | Terms & Conditions Copyright © 1999-2008John Wiley & Sons, Inc. All Rights Reserved.

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20E...GINEERING/62.%20Vehicular%20Technology/W7714.htm (2 of 2)15.06.2008 12:44:13

J. Webster (ed.), Wiley Encyclopedia of Electrical and Electronics Engineering c 1999 John Wiley & Sons, Inc. Copyright


GROUND TRANSPORTATION SYSTEMS Ground transportation systems (GTSs) move people and goods over special ground or underground routes. A GTS includes not only the means of transport, but also the entire infrastructure necessary for providing it. The great variety of GTSs in use may be classified as follows: • • •

Road systems (buses, guided buses, trolleybuses, automobiles) Rail systems (railroads, street cars, subways, light rail) nonconventional systems (maglev, people movers, automatic guided transport, monorails, vehicles with gyroscopes, industrial transport).

This paper is concerned with electrical means of ground transport, so the main focus is on the operation of transport systems with electrically propelled vehicles—electric vehicles (EVs). When power to vehicles is delivered via a special power supply network, the system is called an electrified (network) transport system; when the source of energy is installed in the vehicle, the system is called autonomous. Electrified transport systems can be divided according to the type of power supply: ac or dc, as shown in Tables 1 and 2. EVs supplied by an electric transport power supply system (ETPSS) collect electrical energy from an overhead catenary (railroads, light rail, streetcars [one pole contact wire; see Fig. 1], trolleybuses [two separated negative and positive overhead wires; see Fig. 2(a)]) or from a third rail (railroads in the UK and Berlin, subways), mounted on isolators, through a shoe-type current collector [Fig. 2(b)] placed beside or below the EV. When energy is delivered to a guideway rather than to the EV (as in maglev), no catenary is required (Fig. 3). Nonconventional systems are usually electrified with a special power supply network. Autonomous vehicles are powered by their own source of electrical energy: electrochemical batteries, diesel generators, gas turbine generators, or fuel cells installed onboard. Drive systems for conventional electrical transport are equipped with dc series (usually) or compound motors, ac three-phase asynchronous or synchronous motors, ac single-phase commutator motors, or ripple current motors; some nonconventional systems apply linear induction (asynchronous) or synchronous motors (maglev or subway). According to the type of service, transport systems can be distinguished as follows: mainline (long-distance, intercity: high-speed railroad, upgraded conventional railroad, maglev), regional, suburban, urban (streetcars, light rail, subways, trolleybuses), and local (people movers, monorail).

Railroads The most widespread transport systems in the world are railroads using double-rail track: narrow-gauge (1000 mm to 1067 mm, as in Japan, Taiwan, Central America, and South Africa), standard (1435 mm) or wide-gauge (1524 mm in the former Soviet Union, Finland, and Mongolia; 1668 mm in Spain, Portugal, and South America; 1600 mm in Ireland). The lower motion resistance on rails than on roads allows hauling heavy freight trains 1

2

GROUND TRANSPORTATION SYSTEMS

Fig. 1. Basic scheme of electrified ground transport power supply system with rail electric vehicles supplied by overhead catenary.

(with masses of 2000 tonnes to 6000 tonnes and in some cases over 10,000 tonnes), but requires significant expense for the technical infrastructure (track, fixed installation of power supply system, signaling and control systems).

High-Speed Railroads Because roads are overcrowded and railroads exceed other means of transport in energy efficiency, environmental friendliness, and comfort of service, there has been new interest in electrified rail transport. In order to compete with air transport over distances on the order of hundreds of kilometers, high-speed railroads (HSRs) are being developed (Table 3). HSRs trains in Europe are usually composed of a set of cars, with powered cars at both ends. International interoperability of trains in Europe requires operation under different power supply systems. In addition, TGV-type trains, the French equivalent of HSRs, are in service in Spain (AVE), and Trans-European Euro-star multisystem trains use TGV technology. HSR projects are ongoing as well in Russia, the United States, South Korea, Canada, and Taiwan. Difficulties with power delivery from overhead contact lines to vehicles limit the maximum speed of such rail systems to 350 km/h. Thus, maglev systems are being constructed as the next generation of high-speed ground transport systems.

Maglev Maglev (magnetic levitation) is technologically the most advanced transportation concept. It uses magnetic forces to suspend the vehicle over the guideway and for lateral guiding. (Fig. 3), and a linear induction motor to drive the vehicle. The suspension may be electromagnetic (as in the TRANSRAPID test line in Germany) or

GROUND TRANSPORTATION SYSTEMS

3

Fig. 2. (a) Trolleybus two-pole catenary; (b) third-rail power supply.

Fig. 3. A scheme for a maglev car.

electrodynamic (as in the Yamanashi and Miyazaki test lines in Japan), and may make use of superconducting materials. This system has following advantages over the conventional ones: • • • • • •

No adhesion needed for drive or for braking, the resistance to motion being solely aerodynamic, so that higher gradients and sharper curves can be used Maintenance of horizontal (lateral) position by magnetic forces No contact vibration and less noise, as there are no rolling wheels or rails Minimum maintenance High safety and reliability at speeds up to 500 km/h No catenary, as power is delivered to the guideway.

4

GROUND TRANSPORTATION SYSTEMS

The primary energy consumption per passenger kilometer (with service speed over 400 km/h) in slightly higher than for automobiles (which have a quarter the speed), but only half that for airplanes over the same distance (1). The practical maximum speed for service is 500 km/h, on account of increasing aerodynamic resistance. The first commercial maglev train has been successfully operating at low speed for 11 years on a 600 m long track (a 90 s journey) at Birmingham Airport.

Conventional Railroads Some of the high technology developed for HSRs is being transferred to conventional railroads, which, though they do not reach very high speed, can apply ac motors with power electronic converters and braking energy regeneration to improve the quality of service, lower the cost of maintenance, energy consumption, and environmental impact, and increase the speed. The flexibility of multisystem power supply for trains (composed of locomotives with cars or of multiple traction units) allows one to use the same electrical vehicles on railroad lines and in cities on light-rail lines (as in Karlsruhe, Germany) or even on nonelectrified sections of railroads (when vehicles are additionally equipped with diesel generators or storage batteries).

Urban Transport Systems Electrified transport started at the end of the nineteenth century in towns, and is now enjoying a significant comeback in cities overcrowded with cars and buses causing environmental pollution. Subways, light-rail systems (on dedicated track on bridges and in excavations avoiding level crossings with roads), and modern

GROUND TRANSPORTATION SYSTEMS

5

streetcars (separated from road traffic on streets, with traffic-light priority) yield the highest traffic capacity (Table 4). Low-floor, easy-access streetcars allow one to use traditional routes and drive smoothly into city centers. Nonconventional solutions include people-movers, mini-trams, and fully automated guided transport (Lille, France, 1983; Docklands, London, 1987; SkyTrain, Vancouver, 1986; Automated Skyway Express, Jacksonville, FL, 1992) on special guideways or monorails, used over short distances in areas where efficient passenger transport is in heavy demand (as in airports and at expositions) and needs to be integrated effectively with other transport modes (such as the Chiba Urban Monorail System in Japan). In order to lower the cost of the infrastructure in less-populated areas, trolleybuses are used, and tests are being performed on optically guided hybrids such as diesel–electric buses or minitrams equipped with batteries or gyroscopes, fed from low-voltage power points at stops. The use of linear motors in rail subways in Tokyo and Osaka, Japan, has allowed decreased tunnel diameter and lower capital costs. The choice of an urban transport system depends on the required transport capacity T C (Table 4), which may be calculated using the following formulae:

where T e = transport capacity in one direction [passengers/h] Pv = average capacity of a car [passengers] n = number of cars in a train

6

GROUND TRANSPORTATION SYSTEMS

v = average service speed [km/h] h = headway of trains [h] d = distance between trains [km]

Electric Transport Power Supply Systems An ETPSS is composed of electric power transmission circuits and fixed installations, which are to deliver energy to EVs from dedicated for GTS power stations or public utility systems (PUSs). An ETPSS distributes energy as required by the traction system. Specifications include dc or ac, the voltage, and, if ac, the frequency and whether one- or three-phase. The electric circuits of an ETPSS include power stations or transformer stations, power lines, traction substations (ac or dc), supply feeders, and a traction power network (supply and return network). A typical scheme of an ETPSS is presented in Fig. 1, and designs of different systems are shown in Fig. 4 (dc system), Fig. 5(a–d) (ac industrial-frequency system) and Fig. 6 (low-frequency 16 23 Hz or 25 Hz systems). Characteristic parameters of different traction systems are given in Table 5. Schemes of Traction Power Supply. Power supply schemes of traction systems are divided into feeder sections, which are separated by isolators in the catenary and supplied from the traction substation (TS). When one feeder supplies one section one refers to unilateral supply; when two feeders from different sides supply a section of catenary, to bilateral supply. When sections of catenary for parallel tracks are connected

GROUND TRANSPORTATION SYSTEMS

7

Fig. 4. Dc electrified transport system power supply.

together, this connection is called track-sectioning equipment (or parallel connection points); when equipped with high-speed breakers (HSBs), the combination is called a traction sectioning cabin [Fig. 7(a,b,c)]. Unilateral supply is used on 25 kV 50 Hz ac, 15 kV 16 23 Hz, and 0.6 kV dc traction systems. Bilateral supply is common within 0.8 kV, 1.5 kV, and 3 kV dc systems, while main railway lines electrified with 3 kV dc utilize bilateral supply with traction sectioning cabins. The main purposes of bilateral supply with traction sectioning cabins are lowering voltage drops, improving short-circuit protection, sharing current between different feeders, improving flexibility of the power supply, and, in case of a feeder or TS failure, the capability of switching equipment to make rearrangements to maintain power supply through an adjoining feeder or TS. DC Traction Systems. In a dc traction system (Fig. 4) the ETPSS includes ac power supply lines (high or medium voltage) transmitting power from the power utility system to ac–dc traction substations. Ac

8

GROUND TRANSPORTATION SYSTEMS

Fig. 5. Ac industrial-frequency electrified transport system: (a) general scheme, (b) with booster transformer, (c) with special three-phase–two-phase transformers, (d) with auto transformer (AT).

three-phase power delivered to the TS is then converted to dc power at the required voltage. Mostly 6- or 12-pulse diode rectifiers are used. If voltage control is required (for instance to increase the voltage level under high load or reduce it under no-load conditions), thyristor rectifiers are used, which can also operate as HSBs. When EVs are equipped with regenerative braking equipment, the installation of inverting converters in the TS allows transferring excess energy from EV to ETPSSPUS. Dc power is supplied to the traction supply network from dc positive busbars through feeders (cable or overhead) with HSBs (short-circuit protective devices) and

GROUND TRANSPORTATION SYSTEMS

9

Fig. 6. Ac low-frequency electrified transport system.

switching apparatus, while negative busbars, under normal conditions isolated from the ground (by a spark gap or special grounding equipment) is connected to the rails via return feeders. However, in some systems the polarity is reversed, and the catenary is connected to negative busbars. Return Network of DC Systems. The return network (RN) of a dc power supply is composed of running rails and return feeders. In order to lower the electrical resistance, the rails of the track (for double-track lines, of both tracks) are connected in parallel, with the installation of special impedance bonds on sections with track circuits [Fig. 8(a,b)]. In a fourth-rail system (as in the London Underground, or when wheels with rubber tires are used to lessen vibration and noise), the running rails are separated from the traction current flow and retu feeders are connected to a fourth rail, which collects return current from the EV. This last approach eliminates stray current outflow from the rails to ground. Protection of DC Traction Systems. The ETPSS must maintain the power supply with high reliability and must detect and quickly eliminate fault conditions. This accomplished by installing dc HSBs to protect every feeder from short circuit (SC). An HSB clears the fault in 15 ms to 30 ms, depending on the inductance and resistance of the circuit and the internal tripping time of the HSB. Usually an HSB trips when the

10

GROUND TRANSPORTATION SYSTEMS

Fig. 7. Schemes for sectioning the catenary of a two-track railroad: (a) unilateral supply from one TS, (b) bilateral supply from two TSs, (c) bilateral supply with traction sectioning cabin.

instantaneous value of the current exceeds the preset value Is (instantaneous series trip relay). When the SC happens close to the TS, the steady-state current may reach tens of thousands of amperes, which has to be within the switching-off capacity of the HSB. For a distant SC the minimum current ISCmin (which depends on the resistance of the circuit and the TS and the system voltage) may be lower than the maximum load current IL . In that case it is impossible to set the level of protection I8 to distinguish between the load and the SC current, as when Is > IL (in order not to switch off the load current) one has Is > ISCmin and the HSB does not clear the distant SC. In the other case, when Is > ISCmin , the SC will be cleared, but load currents IL > Is will be switched off, which will cause a traffic disturbance. In case it is impossible to overcome this problem by changes in the supply arrangements (such as shortening the feeder section, switching from unilateral to bilateral supply with interlocking between the adjacent feeder HSBs, or decreasing resistance of the supply circuit), a special detector must be used. This differential protection device distinguishes between the instantaneous current load IL and the SC current Isc . The instantaneous series trip relay responds to di/dt and its duration, or to the change (rise) of current,
i (Fig. 9), over a certain time; it is possible to set Is >IL >ISCmin . As most SCs in a traction catenary are temporary, after the HSB switches off the supply from the section, an impedance relay is used to measure the impedance of the faulty section. If the impedance is high enough, the HSB automatically makes a few tries at switching on voltage to the section. Protection against ground fault is obtained by installing special devices, such as spark gaps or grounding switches, between negative busbars and the ground electrode of the TS. Under normal operating conditions this equipment separates the negative electrode from the ground. In case the voltage between the ground electrode and the negative electrode of the TS exceeds the preset value, the equipment makes a connection between

GROUND TRANSPORTATION SYSTEMS

11

Fig. 8. Parallel connection of rails of a two-track railroad line: (a) using centers of impedance bonds, (b) using additional return wires.

them. If the grounding current flowing through the switch is high enough, then the HSB of the appropriate feeder may clear the fault; otherwise, after a time delay, the special relay trips the ac power switch (time of operation up to 200 ms) of the TS. In order to make the SC from the catenary to the support structures a full SC, they are bonded to the rails [Fig. 20(a)], either individually or in a section group (by a ground wire). Double isolation with the ground wire and individual grounding of support structures may also be used [Fig. 19(b)], but that requires isolation of the foundations from the earth. In case there may be stray currents, special separating devices such as spark gaps or overvoltage contactors (switches) are included in the bonding wire, to connect the ground wire and structures to the rails when the voltage rises above a specified level. Additional protection against SCs of lower level and longer duration is undervoltage protection with a specified voltage– time characteristic. The overhead catenary may additionally be provided with thermal overload protection and lightning protection. The type and the setting of the traction power supply network protection depend on the fixed installation arrangements, local codes of practice, and conditions as well as the characteristics of the operating EV (type of drive system, with or without regenerative braking). AC Industrial-Frequency Systems. A scheme of a simple single-phase ETPSS with rail return, shown in Fig. 5(a), includes three-phase HV power lines connecting the PUS with the traction TSs. In practice more sophisticated solutions are in common use:

12

GROUND TRANSPORTATION SYSTEMS

Fig. 9. Time curve of instantaneous current during distant short circuit and during starting of a train (with change of motor connections from series to parallel).

• • •

Booster-transformer systems more efficient if provided with a return wire—Fig. 5(b), which allow one to reduce induction disturbance effects in telephone networks Ac systems with special three-phase–two-phase transformers (Scott, Leblanc, or Woodridge type) installed in the TS [Fig. 5(c)], to reduce asymmetry in the PUS Autotransformer systems [Fig. 5(d)], most expensive but most efficient, yielding lower voltage drops because the rail–catenary voltage U/2 is half the substation voltage U, so longer distances between TSs are possible and the system is suitable for HSRs or for heavy freight traffic (2 × 25 kV at 50 Hz).

A typical scheme for support of catenary and return-wire connections is shown in Fig. 20(c) below. AC Low-Frequency Systems. The ETPSS of ac low-frequency systems is composed of high-voltage (HV) or medium-voltage (MV) lines connecting the PUS with TSs (Fig. 6) (equipped with static or electromechanical converters of three-phase industrial-frequency power to the one-phase 16 22 . Hz or 25 Hz power required for Europe and the United States, respectively. In some ETPSSs, dedicated power stations or systems with one-phase low-frequency output are used. Protection of AC Traction Supply Networks. Protective equipment for ac traction supply networks differs from that typically used in power utility systems due to the following peculiarities: • • • • • •

EVs are moving along the railroad line, changing the load values and positions and thus the parameters and configuration of the circuit. There are big differences between short-term and long-term loads. Even if unilateral supply is applied in case of fault, the SC current may be energized by more than one feeder or even by a regenerating EV. Two-phase supply of the same section is prohibited. The load current may exceed the minimum value for a distant SC. Quick identification and elimination of the faulty section is required.

The TS is equipped with typical devices for transformers such as overcurrent, ground SC, Bucholtz, overtemperature, and differential protection, while the traction network requires the following protection equipment:

GROUND TRANSPORTATION SYSTEMS • • • •

13

Fast distance-directional impedance protection with current-step relay Thermal overload protection of the catenary and differential protection of the supply-cable Voltage protection against connecting different phases to the same section Lightning and overvoltage protection.

Moving Load on Power Supply Due to Electric Vehicles The dynamic requirement for traction characteristic of an EV is to develop maximum available traction force F (torque M) during starting and then maintaining maximum power Pmax within a range of speeds v up to maximum service speed. Electric Vehicles with DC Motors. Dc series motors have characteristics very similar to those required, though with technical limitations on their traction characteristic [Fig. 10(a)] due to (1) adhesion and sliding, (2) rated power (heating), (3) commutation, and (4) the maximum speed vmax . The maximum rated power of dc traction motors on locomotives is in the range 800 kW to 900 kW with weight/power ratios of 7 kg/kW to 12 kg/kW; for such motors in automobiles the corresponding ranges are 150 kW to 250 kW and 15 kg/kW to 20 kg/kW. Control of a dc series motor depends on its electrical and electromechanical characteristics:

where ω = rotational speed U M = voltage supplying motor Ir = motor rotor current Rd = additional series resistance (starting resistance up to speed v1 ) Rs = motor’s internal resistance CE = electrical constant CM = mechanical constant φ = magnetic flux in motor M e = electromagnetic torque k = coefficient of flux weakening A, B = constants in approximation of magnetizing curve of a motor According to Eqs. (2) 3, 4 it is possible to control the speed of the motor by: • • •

Controlling the supply voltage U M Changing the additional resistance Rd Controlling the flux φ by means of its weakening (coefficient k)

14

GROUND TRANSPORTATION SYSTEMS

Fig. 10. Traction characteristics of EVs with dc motors: (a) rheostatic control and step-flux weakening, (b) switching the connection of motors from series to parallel during starting with rheostatic control and step-flux weakening, (c) chopper control during starting and flux weakening.

GROUND TRANSPORTATION SYSTEMS

15

The traction characteristic of a dc-driven EV (traction fce F versus speed v) is divided into three zones [Fig. 10(b)]: (1) Constant-torque mode for starting: maintaining Ir = const and k = const, so M e = const, F = const, rheostatic control (variable Ra ↓) with change from series to parallel connection of motors [Fig. 10(b)], or voltage control [U M , Fig. 10(c)] with a chopper or a controlled rectifier, (2) Constant-power mode: P = const, Ir = const, U M = const,  k ↓ (k from maximum to minimum (magnetic flux weakening) (3) Natural traction characteristic with reduced current Ir at speed v > v2 , P ↓,  k ↓ (k = const = min), U M = const = max. Rheostatic control of U M is not efficient (dissipates energy) and creates traction characteristics like those shown in Fig. 10(b) for different values of Rd and levels flux weakening, whereas continuous control of U M by static converters and continuous flux weakening k ↓ limits operation to the area below the curve of F (v) in Fig. 10(c). The main disadvantages of dc commutator motors in an EV drive are high maintenance costs due to mechanical brush–commutator contact between stator and rotor, low reliability, and large weight and size per unit rated power.

Electric Vehicles with AC Motors. The development of power electronics and static power converters makes three-phase ac motors (induction or asynchronous) more convenient for use on EVs, due to lower maintenance requirements, lower size/power and weight/power ratios (up to 1.5 MW rated power per motor), and more convenient regenerative braking operation. Variable-voltage, variable-frequency three-phase control inverters allow one to obtain the required traction characteristic of EVs with ac motors. The electromagnetic torque of an asynchronous motor is defined as

where M emax (M er ) = maximum (breakdown) torque U s = supply voltage p = number of poles f s = frequency of voltage U s , ωs = 2πf s synchronous speed s = slip (ωs − ωr ) = ωs ωr = rotor speed From Eq. (5) for M e (ω s ) it is possible to formulate control algorithms (involving variation of the supply voltage U s , frequency f s , and slip s) for an asynchronous motor to get the required traction characteristic of the EV in the area limited by the curves for maximum torque, constant power, maximum slip, and maximum speed V max . Similarly to the traction characteristic of a dc-driven EV, we may distinguish three zones of operation of an EV with an ac induction motor, within which the operation point may be chosen (Fig. 11):

16

GROUND TRANSPORTATION SYSTEMS

Fig. 11. Traction characteristics of EVs with induction motors and inverter control.

(1) Constant torque M e (constant traction force F): M e supply voltage U s increases with speed v up to its nominal value U sn , and f s is increased to achieve U s / F s = const, slip s = const. (2) Constant power: U s = const;  ↓ stator frequency f s increases above the nominal value f sn : slip s increases up to the maximum value Smax ; torque M s is inversely proportional to speed ωs (3) Constant slips at its maximum value: U s = const; power Ps motor current Is and φ decrease; stator frequency f s is still increasing proportionally to the speed; torque M and traction force F are inversely proportional to the square of the speed; zone is limited by the maximum speed V max The traction characteristic F(v) of the EV, according to its dynamic motion requirements must develop a sufficient starting force and enough acceleration force to reach the maximum speed, taking into account the specifications for the traction drive (specified service speed, restrictions of speed on some sections, need to compensate deviations from the time table, changing grade, and curves). An example of the traction force characteristic F(v) and resistance to motion W(v) for the same locomotive with dc motor drive but with different types of trains (curves W 1 to W 5 ) are presented in Fig. 12(a). The traction characteristic F(v) of EVs with dc motors and step flux weakening makes a set of curves with a number of steady-state operating points (accelera- tion = 0) A1 , B1 , C1 for train W 1 , A2 , B2 , C2 , D2 for train W 2 . . . . For an inverter-supplied ac-motor-driven locomotive, the area of operation is limited only by these curves [Fig. 12(b)], so the possible steady-state operating points lie on the curves W 1 to W 5 below the limiting characteristic of the maximum F(v). Equation for an Electric Vehicle Motion. Power delivered to an EV is spent to produce the traction force F increasing the speed (acceleration) and overcoming the resistance W to motion of the train, which is given by

Here W is the total force of resistance to motion [N], v is the speed of the EV [km/h], and ao , ao , a2 are approximate coefficients (dependent on the type of EV—usually obtained empirically) for the frictional and

GROUND TRANSPORTATION SYSTEMS

17

Fig. 12. Steady-state operation points (traction force F equal to resistance to motion) for different EVs and route profiles W 1 to W 5 . (a) With dc motors and step flux weakening, the set of labeled points is accessible. (b) With ac motors and nonstep frequency control, all points within the area limited by the maximum available force F are accessible.

aerodynamic resistance. Typical dependences of these coefficients are as follows:

where A, B, C = empirical constants, K = coefficient dependent on the type of bearings N = number of cars,

18

GROUND TRANSPORTATION SYSTEMS n = number of axles, m = mass of the EV [tonnes,]

Also in Eq. (7), W c is the local curve resistance [N] given approximately for 1435 mm gauge and classical trains by

where R = radius of curvature of the track [m] g = acceleration of gravity [m/s2 ], Modern vehicles with special self-steering wheel axles allow fitting to the rail at curves, which reduces curve resistance in comparison with a typical stiff wheel axle. Finally in Eq. (7), W g is the local grade resistance (positive when the track upgrades and negative when it downgrades), given by W g = mgw [N]

where w is the grade of a section of track [‰]o .]

in which h = equivalent height of grade [m] s = length of the section with constant grade [km] The force of resistance to motion may be expressed in thousandths (per mill, ‰ of the force of gravity on the EV. Usually for low-speed urban transport vehicles the average resistance to motion is assumed to be 4‰ 12‰ for trams and 12‰ 18‰ for trolleybuses. Additional resistance to motion that must be taken into account is caused by wind and by air resistance in tunnels. During coasting, resistance to motion is higher than during motoring (because the kinetic energy covers mechanical losses in motors and transmission). The increase in resistance to motion may be calculated using the increase coefficient kc :

where m4 = driven mass of the EV (the part of the total mass on driven wheels) The kinetic energy Ek of a train at speed v is given by

where m = mass of train k = coefficient that takes into account the energy of rotational parts We have 1.03 for foreight wagons 1.05 for passenger coaches

GROUND TRANSPORTATION SYSTEMS

19

k = 1.1–1.15 for multiple traction units 1.2 for streecars 1.3 for trolley buses The equation of the EV dynamics (equation of motion) for rotational motion of the EV motors is

and for the translational motion of the EV is

where ω = angular velocity of a motor J = moment of inertia of the train (with respect to an axle of a motor) t = time M = drive torque of a motor M r = resistance to motion torque at an axle of a motor F = traction force W = Total force of resistance to motion s = distance The useful mechanical power Pm delivered by the EV is equal

while the power P demanded from the source of energy is

where η = efficiency of the source of power, supply network, main circuit of the EV with traction motors, and gear.

Unitary Energy Consumption. The specific energy consumption j is an average parameter given for a certaint type of EV of given mass, speed, route, and starting and stopping distances and is defined as the energy required to transport 1 tonne of mass a distance of 1 km:

where A = energy consumption [Wh] m = mass [tonnes] s = distance [km]

20

GROUND TRANSPORTATION SYSTEMS

The unitary energy consumption depends significantly on the mode of operation of the EV—maximum and average speed, acceleration, and so on. Special optimization techniques are applied (including on board computer systems) to operate the EV with the minimum energy consumption for a given timetable. Note that quite apart from traction purposes, energy is required for auxiliary needs (air-conditioning, light, etc.) which increase the power demand by 10% to 20%. In order to reduce energy consumption, regenerative braking of an EV can be used, whereby kinetic energy is transformed during braking (for stopping or on downgrades) to electrical energy (the EV operates as a generator), which may be • • •

Used onboard the EV or in the TS (after storage in batteries or elsewhere) Delivered via the power supply network to other EVs consuming current Returned to the TS through inverting rectifiers, which transfer energy to the ac power utility network

Electrical Parameters of the Traction Supply Network The geometrical layout of wires and rails in the traction supply network (TSN) constitutes a complicated structure of mutually connected or coupled electrical components. These elements variously influence the electromagnetic and electromechanical processes occurring in traction systems (dc and ac). The main characteristic parameters of a TSN are R (resistance), L (inductance), M (mutual inductance), and C (capacitance); they are found by calculation or measurement and are usually expressed per kilometer of length. An equivalent simplified scheme of a catenary–rails–ground circuit is shown in Fig. 13(a). Usually during calculations of voltages and currents the capacitance C is neglected, but for transients, overvoltages, and resonances all parameters must be taken into account due to their significant influence on short-time phenomena in circuits. Some elements (rails, power rails, and messenger wires) are made of ferromagnetic materials, which causes their parameters to depend on current and frequency. The impedance of catenaries is dependent on their configuration and location with respect to ground, since the ground is a part of the return network, and may be described as

where ra Li Le M

= resistance per unit length [/km] = internal inductance per unit length of the circuit [H/km] = external inductance per unit length of the circuit with ground return [H/km] = mutual inductance per unit length of the circuit [H/km]

which may be rewritten in the form

where the x’s are reactances per unit length. For nonferromagnetic magnetic materials the resistance ra is equal to that for dc circuits. In case ferromagnetic materials are used in dc traction circuits during transients or with harmonics or in ac circuits, the skin effect must be considered, as the resistance increases for ac components:

where k = coefficient dependent on frequency, cross-section, perimeter, and permeability (2–4) r0 = resistance per unit length for dc component [/km],

GROUND TRANSPORTATION SYSTEMS

21

Fig. 13. A scheme of an elementary section of an electric traction power supply network with self- and mutual impedances: (a) simplified one-track, (b) dual-track.

22

GROUND TRANSPORTATION SYSTEMS The self-impedance zr of a rail may be described by the following simplified equation (5):

where µo = 4 π × 10 − 7 H/m, µr = relative permeability [for a ferromagnetic conductor with significant current change, see Table 6(a) (5)] ρr = resistivity of a rail [·m], ω = 2πf with f = frequency [Hz] Re = radius of an equivalent circular cross section for a rail =L/2π [m] where L = length of the perimeter of a cross section of a rail [m] The internal impedance per unit length of a conductor 5 is

where µr = 1 for Al and Cu. The external impedance of a conductor (Carson–Clem–Pollaczek formulae) (5) is

where De , the equivalent distance (depth) of the earth current from the surface of the ground, is given by

and ρ is the ground resistivity [·m]. The mutual impedance of conductors a and b (5) is equal to

where Dah is the distance between the conductors a and b [m] For multiconductor networks there are formulas allowing one to make the calculations for the equivalent one-conductor circuit. The equivalent impedance ze of a messenger-wire-contact-wire circuit is equal to

where Zw = impedance of contact wire Zm = impedance of messenger wire Zwm = mutual impedance between contact wire and messenger wire

GROUND TRANSPORTATION SYSTEMS

23

For a pair of identical conductors connected in parallel ails of one track, rails or catenary of two tracks) the following equation is used:

where Zeq = equivalent impedance of the pair of conductors Za = self-impedance of the conductor, Zab = mutual impedance between the two identical conductors a and b When the rails are taken into account as a return network, the equivalent impedance z1 of a catenary of one track is equal to

24

GROUND TRANSPORTATION SYSTEMS

where Zc = self-impedance of catenary [/km] Zr = self-impedance of catenary [/km] Zer = mutual impedance from catenary to rails [/km] Zrc = mutual impedance from rails to catenary [/km] Ic = current flowing in catenary [A] Ir = current flowing in rails [A] For the equivalent circuit shown in Fig. 14(a) the voltage matrix equation is

with

so the voltage drops between catenary and rails are

For directions of currents as in Fig. 14(b) (two-track railroad) the impedance matrix Z is equal to

and for directions as in Fig. 14(c),

After solving the equations when the currents are known in relation to the source (substation) voltage U, it is possible to calculate the lumped impedance of the circuit as the ratio of U to a sum of currents. The representation of electric traction power supply circuits depends on the kind and purpose of the performed analysis, which may include: • •

Balance of power and energy, with losses and energy consumption (parameters R, L, C for ac and R for dc systems) Voltage drops (R, L for ac and R for dc systems)

GROUND TRANSPORTATION SYSTEMS

25

Fig. 14. An equivalent scheme for a traction power supply network with the ground as a return wire: (a) one-track, (b,c) dual-track.

• • • •

SC time-constant calculations (R, L of circuits for both ac and dc systems) Overvoltages and response to switching on and off (R, L, C for ac and dc systems) Frequency-dependent characteristics and high harmonics (R, L, C for ac and dc systems) Rail–ground voltage calculations; stray currents (R, L for ac and R for dc systems) Typical parameters of a traction network are presented in Table 6(b) and 6 (6).

26

GROUND TRANSPORTATION SYSTEMS

Fig. 15. Plain trolley wire: L, span distance; f cmax , maximum sag; F c , tensile force in the contact wire.

Fig. 16. Sagged simple overhead catenary.

Fig. 17. Sagged multiple overhead catenary.

Overhead Catenaries An overhead catenary (OC) consists of one or more wires elastically suspended above the EV, which receive current via a current collector (pantograph type or trolley type) installed on the roof. The design of an OC, which with the pantograph constitutes a vibrating system, depends basically on the requirement of maintaining proper contact with the pantograph at the required current and maximum speed. The main types of OC are as follow: • • •

Plain trolley wire (composed to contact wire only; used for low-speed EVsuch as a trolleybus or a streetcar) (Fig. 15) Sagged simple (Fig. 16) or multiple (Fig. 17) OC, composed of one or more messenger wires and suspended on one or more dropper contact wires Compound single or multiple OC (in Fig. 18 is shown a compound single OC with dampers to reduce mechanical oscillations) Basic technical data of an OC are as follows (Fig. 19):

• • •

System height hk Height of contact wire suspension, h Span distance (between the adjacent support structures), L

GROUND TRANSPORTATION SYSTEMS

27

Fig. 18. Compound single overhead catenary with dampers.

Fig. 19. Effect of a moving pantograph’s position on an overhead catenary.

The suspension of an OC is made using special support structures (Fig. 20) with stagger (snaking) of the contact wire (Fig. 21), which allows use of the whole contact part of the pantograph pan and reduces local wear. Wear results from both mechanical and electrical effects. The contact force creates friction, which causes mechanical wear (larger force means more wear). Larger contact force, however, leads to lower equivalent contact resistance Rz and less electrical wear, which usually has major importance. The heat Q emitted at the point of contact the pantograph pan with the contact wire is

where F = contact force between pantograph and contact wire µ = friction coefficient between pantograph and contact wire i(t) = collected current Rz = equivalent resistance between pantograph and contact wire In order to maintain constant conditions of mechanical operation of an OC with a pantograph it is required to keep constant the forces tensioning the contact wire, F c (Fig. 15), and the messenger wire, F m (Fig. 19), which influence the sag f c of the contact wire. The maximum sag can be calculated as follows (for the plain OC,

28

GROUND TRANSPORTATION SYSTEMS

Fig. 20. Schemes of support structures of an overhead catenary and connections of wires: (a) with a parallel feeder and a ground wire connecting the supports, ground wire bonded to rail (dc system); (b) with parallel feeder and double isolation of catenary and ground wire connecting neutral parts between two isolators (dc system); (c) used in ac traction systems (bonding of a grounded support to the ground rail and return wire).

Fig. 15):

where f emax = maximum sag, midspan [m]

GROUND TRANSPORTATION SYSTEMS

29

Fig. 21. Plan view of catenary, showing staggering of catenary and position of pantograph along the track axis.

m = mass of contact wire per unit length [kg/m] L = span distance [m] F c = tensile force of contact wire [N] The temperature-dependent length change
I of catenary between the two supports (span length) is given by

where
t = temperature change [◦ C], c1 = coefficient thermal expansion of the contact wire lsqb;◦ C − 1 ]. To reduce this effect, mechanical tensioning equipment (TE) is used (a balance weight as in Fig. 22(b)– 22(d), or a pneumatic or hydraulic system) on both ends of a section of catenary (up to 1600 m) including a number of spans. An typical simplified plan view of a span with two tensioning sections (1 and 2), with the indicated section of current collection frothe contact wires of both sections, is shown in Fig. 23. An OC that is first tensioned without the TE, is called a noncompensated OC (because the tensile forces change with the temperature); see Fig. 21(a). An OC with the messenger wire first tensioned without the TE and the contact wire tensioned with the use of the TE is called half-compensated [Fig. 21(b)]. If both the messenger wire and the contact wire are tensioned with the TE [Fig. 21(c,d)], the OC is called fully compensated. The contact wire is usually made from Cu, possibly alloyed with Ag or Si; typical cross sections are 100 mm2 to 170 mm2 . A messenger wire is typically made of Cu, bronze, steel, or combined steel and aluminium, with cross section 50 mm2 to 180 mm2 . Parallel feeders, usually suspended on the other side of the support structures (Fig. 20) to increase the equivalent cross section of the catenary, are made of Cu or Al (with Al–Fe wire), as are the protective ground wires used for connecting bonded [Fig. 20(a)] or grounded [Fig. 20(b)] support structures. Typical parameters of overhead catenaries are given in Table 7. Conductor rails in third-rail [Fig. 2(b)] or fourth-rail systems are made of steel, aluminum (low wear resistance), or aluminum with a steel surface (good conductivity and wear resistance). There are different types of current collectors (CCs) used: trolley collector [Fig. 2(a)], shoe-type [Fig. 2(b)], bow collector, and pantograph (Fig. 19). The pantograph’s pan head is made of Cu, graphite, or metal-plated carbon with addition of lubricant. During running of the EV its pantograph, pushed by a force F, raises the contact wire (CW) to a height
h that depends on F, the CW’s elasticity, the speed of the vehicle, the vibration of the vehicle, and the pantograph’s mass. (In Fig. 19 two positions of the pantograph are shown: at the point A close to the support, and at B, the midspan of the OC, with heights denoted respectively as
hA and
hB ) The total force F is the sum of the following forces:

30

GROUND TRANSPORTATION SYSTEMS

Fig. 22. Types of overhead catenary with compensation equipment: (a) noncompensated, (b) half-compensated, (c, d) fully compensated.

where F µ = friction in the pantograph joints, (− for movement up; + for down) F i = inertia force (− for movement up; + for down) F a = aerodynamic interaction (vertical component of aerodynamic force) By sinusoidal approximation of the contact point (pantograph–contact-wire trajectory) or its first harmonic (Fig. 24) it is possible to describe its vertical movement (ordinate y) along the span (position x) with the

GROUND TRANSPORTATION SYSTEMS

31

Fig. 23. Plan view of two tensioning sections of an OC.

following assumptions: infinite catenary length, constant span length L, constant speed v of the EV, and constant parameters of catenary and pantograph. We have

32

GROUND TRANSPORTATION SYSTEMS

Fig. 24. Trajectory of vertical position y the contact point between pantograph and contact wire along the span x.

where ω0 = natural frequency of pantograph m0 = equivalent mass of pantograph referred to the contact point ki = coefficient of vertical aerodynamic force on pantograph Then the simplified equation of the contact point x is

where ms (x) = equivalent mass of catenary (referred to the contact point) y = y(t)= ordinate of contact-point trajectory rs = coefficient of friction e(x) = elasticity of catenary In practice the characteristics of the catenary are assessed by calculation and measurement, from which the following parameters are obtained: •

The frequency of free vibrations, ωs , which may be approximated by

where F m = tensile force of messenger wire [N] F = tensile force of contact wire [N] mc = mass per unit length of contact wire [kg/m], mm = mass per unit length of messenger wire [kg/m] Le = span distance [m] In case the pantograph is moving at the critical speed ver equal to

mechanical resonance occurs and ωs = ωo .

GROUND TRANSPORTATION SYSTEMS

33

The coefficient of nonuniformity of static elasticity,

where emax = midspan elasticity emin = elasticity at support structure The propagation velocity vw of the transverse waves,

The reflection coefficient for transverse waves,

Doppler’s coefficient

where v is the speed of the EV. The amplification coefficient

The coefficient of contact continuity,

where to = time that wire and pantograph are out of electrical contact tp = time of EV run on the catenary section For high-speed railroad lines with speeds of 200 km/h to 250 km/h, the catenary parameters are given in Table 8.

Systems Approach to Electrified Ground Transportation In order to represent the technical complexity of electrified ground transportation (EGT) treated as a multivariable system, the rail system may be divided into a finite number of subsystems (Fig. 25), where

34

GROUND TRANSPORTATION SYSTEMS

power utility system traction substation (ac or dc) = traction power supply system = electric vehicles = traction signaling, command, and control system = railroad = demand for transport service (freight and passenger) = influence of EGT system (EGTS) on the environment and surrounding technical infrastructure = performed transport service (in passenger-kilometers, passengers per hour, or tonne-kilometers of freight) The systems approach allows us to define internal and external interdependences between subsystems, which are described as follows: U 1 = system voltage; power capacity of PUS (external input) Y 1 = voltage; short-circuit power of PUS at point of connection of TS to PUS Z1 = influence of traction substation load on PUS (changes of load, harmonics, asymmetry) U 2 = voltage; short-circuit power of PUS at inlet busbars of TS Y 2 = currents of feeders; voltage at output busbars of TS U 3 = voltage at output busbars of TS; currents taken by trains Y 3 = voltage at pantographs of EVs; currents in the contact line Z3 = locations of EVs in violation of the timetable Z4 = voltage in the contact line below the specified limit, causing abnormal operation of TPSS U 4 = trains’ positions and speeds as required by timetable; data and commands from control and signaling system delivered to trains; voltage in the contact line Y 4 = actual mode of operation of trains; power; current; actual location of trains Z5 = actual traffic situation on the line: traffic disturbances, delays, compatibility disturbances (harmonic distortion caused by power traction circuits in CS installations) Y 5 = trains’ positions and speeds as required by timetable; data and commands from control and signaling system delivered to trains U 5 = traffic technology; specified timetable (type of trains, masses, speeds, etc.) according to expected and real transport service demand U 6 = return current of EVs; mechanical load on track due to EV movement

GROUND TRANSPORTATION SYSTEMS

35

Fig. 25. Electrified ground transport system divided into a number of subsystems.

Y 6 = current distribution in rails; stray currents; rail–ground voltage

Choice of Electric Traction System The choice of the electric traction system depends on many different conditions, such as:

• • • • • •

Type of the transport system (urban, suburban, long distance) Traffic capacity and expected power demand Environmental impact and safety Availability of technical infrastructure (e.g., public power supply system and its power capacity) Compatibility with existing transport systems Cost–benefit analysis of different options

The choice of the system voltage determines the required parameters of the power supply structure, the rolling stock, and the surrounding technical infrastructure.

36

GROUND TRANSPORTATION SYSTEMS

Sizing of Power Supply Equipment and Installation Calculation of the required power to be delivered to an EV is based on the assumed traffic volume and the resulting timetable. From traction calculations—solving the equations in Eq. (13) of EV motion—the power demand of vehicles for peak traffic hours (maximum power demand) is obtained and used for sizing the power equipment. Then the power flow in the supply system (dc or ac) and the loads on the catenary, feeders, and traction substation are determined. The following parameters are calculated to evaluate the required parameters of the installations: • • • • • •

Values of rms current (average over 15 min for catenary; over 30 min and 1 h for feeders), Peak values of feeder currents (to set the level of short-circuit protection), Power demand from traction substation and its time profile Minimum allowed voltage at pantograph of EV and on busbars of traction substation Voltage drops in rails (leading to safety hazards and stray currents) Power delivery efficiency

The calculated values must be (within a certain margin) below the capacity of the power supply installations, and the power required from locomotives must be within their capacity. As the above usually requires simultaneous solution of a number of electromechanical equations of EV motion and electrical equations of power flow, modeling and simulation methods are widely used. A simplified method of sizing is based on knowledge of j, the specific energy consumption of an EV, from which an average power demand per unit length, P, on a section of the power supply is obtained:

where n = frequency of traffic [trains/h] j = specific energy consumption [Wh/tonne·km] v = average speed of EV [km/h] m = mass of a train [tonne] l = length of a power supply section [km] Assuming an average voltage in the contact line, the current load per unit length of a section may be calculated, and thence the load on the traction substation (using dc or ac power flow calculations). Sizing of ac power supply installations from a public utility network is based on the obtained TS load profile.

Modeling and Simulation Methods as a Tool for Analysis of an Electrified Ground Transport System Modeling and simulation of such a complex system as an EGTS involve many interacting problems concerning subsystems (Fig. 25). Many assumptions and simplifications must be made in formulating the model and calculations. The preparation of an EGTS simulator requires not only analytical, but also logical techniques to obtain either a theoretical or a methodological solution. The process of modeling and simulation consists of: • •

Splitting the EGTS into functional subsystems (Fig. 25) Constructing a mathematical model of the each subsystem with sufficient accuracy

GROUND TRANSPORTATION SYSTEMS • •

37

Reviewing the algorithms and other tools that might be applied to these models, taking into consideration the available computing facilities Solving the identified problems using suitable mathematical methods and iterative procedures In view of the complexity of EGTSs, models in practice are oriented towards the following main aspects:

• • • •

Electromechanical problems (motion of EV, power demand of EV, influence of voltage conditions on traction characteristics of EV). Main model: equation of EV motion. Power flow problems (power flow analysis in traction system and PUS, sizing of equipment and installations). Main model: power flow model of traction system (dc or ac) and PUS. Electromagnetic problems concerned with harmonics and transients: disturbances, SCs, dynamic states of the system. Specialized circuit and electromagnetic-field-oriented models are required to analyze shortterm interdependences between subsystems and elements. Traffic technology (timetable, traffic flow, management).

Simulation methods are applied for the preliminary study, design work, and verification of the real system’s operation, as some experiments are difficult or even impossible to carry out on a functioning transport network (7,8,9,10,11,12,13,14,15,15).

Control and Signaling in Transportation Systems A transportation system has to ensure appropriate (as high as economically possible) reliability, availability, maintainability, and safety (RAMS). All RAMS figures are increasing with the introduction of new systems and new technologies, although they never reach 100%. New systems also allow increased speed and line capacity and decreased traveling times for high-speed and heavy-traffic lines, as well as improved economy of operation, especially for secondary and low- traffic lines. Conventional Signaling Systems (with Colored Light Signals). Widely used conventional signaling systems (with colored light signals) are based on track occupancy checking devices. The track is divided into sections. Occupancy of a section of track may be checked by track circuits or by axle counters. Some railroads have also tried special infrared train-end devices, but they are not used in practice because of operational problems. On the basis of the track occupancy information, special installations—block systems on lines and interlocks at stations—prepare routes for trains. A route for a train must be set using only unoccupied tracks. Then the route has to be proven and locked. When the route is locked for a train, a color is displayed on the light signal that gives permission to the driver to enter the route. The colors displayed on the signal depend on the railroad administration and the traffic situation. There are two basic principles for signaling: speed signaling (signal shows maximum allowed speed for passing it and the next signal) and distance signaling (signal shows distance to the point that should not be passed—usually as a number of unoccupied block sections after the signal for which the route is set, proven, and locked). As a result, information is passed to the driver, who is responsible for keeping the train running within the limits. Track Circuits. The basic method of track occupancy detection is the separation of the track into blocks equipped with track circuits. Widely used track circuits are low-frequency track circuits with insulation joints and jointless track circuits. They are described below. Other means of detection of trains on tracks include impulse track circuits (used on nonelectrified railroads), binary-coded track circuits, and axle counters. Low-Frequency Track Circuits. These are used mainly on dc electrified systems. Both rails are used to transmit an ac signal from a low-voltage transmitter T1 to a receiver R1 (where T1 is situated at one end of the circuit and R1 at the other) or receivers R3A and R3B (where T3 is situated at the center, and R3A and

38

GROUND TRANSPORTATION SYSTEMS

Fig. 26. Track circuits: (a) low-frequency track circuits with insulated joints (IJ), both single-rail (TC2) and double-rail (TC1, TC3); (b) jointless track circuit; (c) track-to-train communication using jointless track circuit.

R3B at the ends) [Fig. 26(a)]. Block rails are separated by insulation joints (IJ) at both ends of the block. As rails are used for the return path of the traction current IT (both rails in double-rail, TC1, but one rail only in single-rail, TC2), inductive bonds L are used, which present high impedance at the signaling frequency, but low resistance to the dc current IT . When there is no train on the block (TC1), relay R1 is energized with signaling current Is1 from T1. In case a train occupies the track (TC3), the signaling current Is3 is shunted by a wheel set of the train and relays R3A and R3B are deenergized, signaling the occupancy of the track, which is shown by changing the color of a semaphore light. Typical operating frequencies are: 25 Hz, 33 13 Hz, 50 Hz, 60 Hz, 75 Hz, 83 13 Hz, 100 Hz and 125 Hz. The length of the circuit depends on the frequency, voltage, and power of the transmitter and on the electrical parameters of the rails, ties, and ballast. The track circuit is described using transmission-line equations. Single-rail dc track circuits used on ac electrified railway lines are based on the same principle. Jointless Track Circuits. Here two rails are used both for the traction return current and for the signaling current, whose frequency f is in e audio range (typically 1.5 kHz to 3 kHz) [Fig. 26(b)]. The signaling current

GROUND TRANSPORTATION SYSTEMS

39

is generated by a transmitter T and received by a tuned receiver R. No insulated joints are used, and the track circuit is electrically terminated at its ends by terminating bonds (TBs) composed of L and C elements. A TB presents low impedance at the signaling current frequency f , but high impedance at other signal frequencies. In order to avoid interference, neighboring track circuits operate at different frequencies. Other means of detection of a train on the block include (1) comparison of the phase and amplitude of signals detected by receivers at both ends of the circuit from a transmitter at the center, and (2) frequency modulation. Vigilance; Automatic Train Protection, Control, and Operation. As long as track-to-train communication takes place via the driver’s eyes, onboard systems are not able to monitor driver behavior. The only possibility is a passive vigilance device periodically (every n seconds) checking only the presence of the driver (e.g., the dead man used by British Railways). Track-to-train transmission must be available to introduce active vigilance devices. As an example we mention SHP, used by Polish State Railways, which checks driver vigilance 600 m before each signal. If the driver does not operate the vigilance button within 5 s, an emergency brake is applied. An automatic train protection (ATP) system is an intelligent overlay on the conventional signaling system. The track–train transmission includes the maximum speed and (possibly) the permitted distance. The onboard ATP equipment does not allow the driver to override the limits. As examples we mention many systems used by European railways: ZUB, EBICAB, KHP, INDUSI, ASEC, SELCAB, BACC, EVM, KVB, and others. An automatic train control (ATC) system combines ATP functions with some functions of conventional signaling systems. Examples are LZB and TVM. In practice the boundary between ATP and ATC systems is not strictly defined, so we speak of ATP–ATC systems. ATP–ATC systems may be based on spot transmission, on transmission, sections, or on continuous transmission. Spot transmission means sending data at certain points where the train passes a trackside device: a transponder (balise), short track circuit, or cable loop. Sent information is then valid at the moment of transmission. If an in-fill section (leaky cable, medium loop, or local radio) is added before the spot device, a stop signal that has changed to a proceed signal can be seen by the system without the need to stop. Continuous transmission (via long loops, coded track circuit, or radio) meets the requirement that the information be up to date. As an example we consider jointless track circuits, which are used for both track-occupancy detection and track-to-train communication [Fig. 26(c)]. The terminating impedae bonds (TB) are to be tuned both to the track signal frequency f s and to the track-to-train signal frequency f TT . The signal Is flowing in the rails is detected by an onboard receiver (OBR) mounted ahead of the first axle under the locomotive moving from the receiver R to the transmitter T (the track current Is is shunted by the axles and does not flow in the rails behind the first axle). An automatic train operation (ATO) system is an autodriver (like an autopilot in an airplane). Such a system requires more information, as for example the start and length of the platform and the side on which the platform is situated. It may take into account timetable information and energy-saving criteria. Such systems are used in subway trains, where there is no mixed traffic (programming of ATO is much more complex for mixed traffic such as occurs on ordinary railroad lines) and working conditions for drivers are bad (running mostly underground).

European Train Control System—A New Unified Automatic Train Protection and Control System. The European railroads, together with the European signaling industries, are preparing a new standard for track–train transmission and supervision of drivers. It is the European Train Control System (ETCS), functionally similar to the Advanced Train Control System (ATCS) specified by the Association of American Railroads. ETCS is designed for passenger and freight traffic command and control in order to achieve interoperability of trains between different railroad networks in Europe (no need to change locomotives or drivers, or even to stop trains at borders). Data received from the track, together with data available onboard, are used for calculation of static and dynamic speed profiles, which are continuously compared with actual train speed and traveled distance. Most of the functions can be performed by either the trackside or the onboard equipment.

40

GROUND TRANSPORTATION SYSTEMS

ETCS guarantees safe operation of trains on the set route and within the speed limits, especially on high-speed lines. ETCS is divided into the following three levels of application: •

•

•

Level 1. Train is equipped with onboard safety computer, maintenance computer with man–machine interface, recorder, odometer, and antenna to receive data from track installations. Track is equipped with switchable beacons connected via line-side electronic units (LEUs) to the signals or directly to the interlocking or blocking system. Optionally level 1 may be equipped with in-fill channels for actualization of data received from trackside. This may be done by Euro-loops, or by a specific transmission module (STM). Level 2.Train is additionally equipped with the ETCS EIRENE radio. Track is equipped with nonswitchable beacons (used for location reference and for transmission of permanent data) and a radio block center (RBC), which communicates on one side with interlocks and blocking systems and on the other side with trains, giving all switchable information. Colored light signals are no longer required (if all entering trains are equipped with ETCS level 2 onboard equipment), as their function is taken over by radio transmission. Level 3.Train is additionally equipped with a train integrity unit, which forms a basis for the supervision of track occupancy. Track circuits and axle counters are no longer required. Level 3 allows trains to operate under moving-block schemes (which increases the line capacity), where the headway between trains is continuously regulated using information about the position and the speed of the preceding train.

Traffic Management Systems. On top of the control and signaling systems there is a dispatching system whose aim is traffic management. Different railroads use different dispatching systems. Introduction of a unified ATP–ATC system—ETCS—will allow unification of traffic management. The European railroads are planning a new standard for traffic management: the European Railway Traffic Management System (ERTMS). They have already started to define its scope. It includes at least the ETCS control command, and traffic management functions. It may also include management of locomotives and cars, power supply management, ATO, timetable planning, track maintenance planning, and other functions.

Impact of an Electrified Ground Transport System on the Environment and the Technical Infrastructure One of the main advantages of electrified over nonelectrified transport is its lower impact on the environment. Namely, an EGTS makes less noise; it uses electric energy, which may be produced far away from the area of its consumption, so there is no emission of fumes along the line; and its transport capacity and energy efficiency are larger, so that it is more reliable and cost-effective. However, an EGTS has some disadvantages: the large capital investment required, landscape problems due to overhead catenares and the use of land for TSs and power lines, and technical problems such as the effect of TSs on the power utility system, electromagnetic compatibility, and stray currents.

Impact of Traction Substations on Power Utility Systems The effect of traction substations on the public power utility system (PUS) supplying them depends on the type of electric traction system: dc or ac. Dc traction substations present a load that is nonlinear (current harmonics are created by rectifier or inverter operation) and nonsteady (flickers are created by changes of power demand). Ac (50 Hz) traction substations—apart from harmonics (generated by power electronic converters installed in

GROUND TRANSPORTATION SYSTEMS

41

electric vehicles), reactive power loads, and fluctuations of taken power—create non-three-phase loads, which cause asymmetry in the PUS. Harmonic Distortion in a Public Utility System. International and country standards impose limits on the individual or total harmonic distortion (THD) harmonics caused by nonlinear loads and converters on a PUS. These limits depend on the voltage, and use of a high-voltage supply with high power capacity and fault level significantly decreases the distortions. When it is impossible to decrease harmonic distortion below the limits, the traction substation should be equipped with filters or be supplied individually via a separate line or transformer. Asymmetry. Ac single-phase traction substations cause unbalance in three-phase PUSs, which then causes negative phase sequence current (nps). The level of nps can be assessed for one unbalanced load as follows (15):

where ST = traction line-to-line load Ssc3 = three-phase short-circuit level at point of TS connection to PUS For ac TSs, when asymmetry and harmonics exert a combined influence on the PUS, both disturbances must be analyzed together. In case the limits on these disturbances are exceeded, special measures must be undertaken, such as changing the power supply arrangements or installing symmetrizing equipment. Voltage Flickers. Due to step changes of traction loads, voltage flickers (VFs) may be observed at the point of common coupling (PCC) of a TS, which influence other energy consumers. The limits of allowed VF are usually expressed as the maximum permitted repeated voltage change
U [%] as a function of frequency f [min1 ], which may be specified by the PUS company as in Ref. 9 (for 38 kV lines):

or as suggested by the PUS company in Poland: •

For HV lines,

•

For MV lines,

42

GROUND TRANSPORTATION SYSTEMS

It is possible to recalculate the above values of
U for the corresponding permitted step changes in current load
I using the following formulas (for DC traction systems):

where U = supply voltage, U DC = Dc voltage of the rectifier TS, η = efficiency of the TS, X, R = reactance and resistance of PCC (corresponding to a fault level), cos φ = load power factor (assumed lagging) and then compare the results with the load profile of TS obtained from traction power supply system calculations. Stray Currents. Rails that are used as a part of return current network are not perfectly isolated from the ties, ballast, and ground, and a significant part of the return current I flows to the TS through the ground and conducting elements buried in it [stray currents, Ih Fig. 27(a)]. In areas when rails are at higher potential than the surrounding ground (anode zone of rails), current flows out from the rails to the ground, while in areas where the ground is at higher potential than the rails (cathode zone of rails) currents return to the rails. If the return rails are connected to the negative busbar of the traction rectifier substation, the anode zone moves with the EV taking current, while the cathode zone is situated around connections of return feeders to rails [Fig. 27(a)]. Stray currents cause electrochemical corrosion of metallic structures buried in the ground, because they operate as galvanic cells, the ground being an electrolyte, and the ions (charged atoms) of metal move away from the structures (mainly in cathode zones of rails). The mass of eroded metal is governed by Faraday’s law. The return-path traction current can be described using the resistance per unit length of the rails, rR [/km], and the resistance from the rails to the ground, rR−G [ km] or conductance gR−G [S/km], gR−G = 1/rR−G [Fig. 27(b)]. The flow of current dIR (x) from an elementary section dx of rails is shown in Fig. 27(c), where Ib (x) = stray current at distance x from the EV Ir (x) = current flowing in rails at distance x from the EV U R−G (x) = voltage between rails and remote ground at distance x dU RG (x) = change of voltage between rails and remote ground at an elementary section dx of rails The equations are as follows:

where the minus sign means that for positive values of U R (x) and dx the current IR in the rails decreases [Fig. 27(c)]. Solving the above equations, we get

GROUND TRANSPORTATION SYSTEMS

43

Fig. 27. Stray currents in dc traction: (a) scheme of stray current flow with one EV receiving current I from one rectifier substation, (b) equivalent dc circuit, (c) current flow away from elementary section dx of rails, (d) rail current IR and stray Is current along the section, (e) rail voltage U R along the section.

44

GROUND TRANSPORTATION SYSTEMS

where C1 and C2 are integration constants, and

Graphs of U R , IR , and Is for unilateral supply of one EV taking current I are shown in Fig. 27(d) and (e). Usually the supply schemes, especially for a streetcar return network, are more complicated, utilizing a distributed return network. One then applies superposition method, using the above equations for each load, or else matrix methods. In order to eliminate stray current migration from rails, the following remedies should be considered: • • • • •

Lowering the voltage between rails and ground (by lowering of the resistance of rails or shortening the distances between traction substations) Proper configuration of return feeder cables, higher system voltage, open bonding of the catenary support structures [Fig. 20(a)], or grounding with isolation of the foundation from the earth [Fig. 20(b)] Increasing the resistance between rails and ground (depending on the type of ties and ballast)—in some circumstances this may cause high step or contact voltage between rails and ground), or providing a special path for return current separated from the running rails Isolation and sectioning of installations buried in the ground If an area (e.g., in a tunnel or near the sea) is especially vulnerable to stray currents (which may assessed by measurements or calculations), special measures are to be undertaken as follows: cathodic protection with an imposed current (which decreases the voltage between the rails and the protected installation in the anode zone of the rails, decreasing the stray current flow), or electrical drainage (equipotentialization between the rails and the protected installation in the cathode zone of the rails).

Electromagnetic Compatibility of Electrified Ground Transport Systems As may be seen in Fig. 25(f), an EGTS is a complex system with subsystems that, in order to operate safely and reliably, have to fulfil electromagnetic compatibility requirements. The difficulties in achieving compatibility result from the different categories of subsystems: the high-power supply system and EVs are in the megawatt and kilovolt range, while low-power signaling and control systems operate the range of milliamperes and volts. Another problem is the operating frequency bands of installations. Even if different subsystems operate within separate frequency ranges, there may appear harmonics, which can cause disturbances that are transferred from one subsystem to another by capacitive, conductive (galvanic), and inductive coupling and wave phenomena. For safety reasons the voltages transferred from power circuits to conducting structures of an electrified railway have to be kept below certain limits [the maximum allowed longitudinal continuous induced voltage

GROUND TRANSPORTATION SYSTEMS

45

Fig. 28. Disturbances in electrified transport system and neighboring technical infrastructure: TS, traction substation; R, return network; TC, traction circuits; OC, overhead catenary; CC, current collector; IF, input filter; CH, chopper; 4QS, four-quadrant converter; IN, inverter.

to ground (at the fundamental system frequency) 60 V; in short-term fault conditions it is 430 V (2, 7)]. A special ground wire is installed, to connect support structures and protect against SC or isolation breakage, by bonding to the rails [Fig. 20(a,c)], or double isolation is used [Fig. 20(b)]. Voltage drops and harmonics flowing in rails (conductive coupling) may cause the following malfunctions of track circuits: • •

Right-side failure.Safe, but causing disturbances in traffic when unoccupied track is signaled as occupied Wrong-side failure.Dangerous when a block of track with a train on it is signalled as unoccupied

According to the type of traction circuit (TC) used, limits on harmonic currents in rails are imposed. In case disturbances exceed the limits, special measures should be undertaken to reduce disturbances at their source (using filters, or changing the equipment or its mode of operation) or to make the TC immune from them. The main source of disturbances in EGTSs are (Fig. 28): • • • •

TS. Dc traction rectifier substations (dc side voltage harmonics) Power Electronic Converters of the EV. Dc-dc converters [choppers (CH) at fixed frequency] or ac-dc converters (4QS); inverters (IN) supplying ac motors (with frequency of operation variable from 2 Hz to 120 Hz) Current Collectors (CC). Radio-frequency harmonics Transients.starting of EV, faults in catenary

The current I delivered from TS to EV is composed not only of characteristic harmonics (typical due to type of equipment used) but also of noncharacteristic harmonics (due to faulty operation of converters) or asymmetry and sideband harmonics (due to intermodulation of harmonics imposed by different sources). The catenary (OC) and return railway current network (R) in Fig. 28 show frequency dependence of their parameters, which, together with the dc side filter of the rectifier substation (in dc power supply systems) and the input filter (IF) of the EV (which by moving along the line changes the parameters of the energy source),

46

GROUND TRANSPORTATION SYSTEMS

create various reso nant circuits. In ac systems resonance may cause significant catenary overvoltages (more than twice the nominal voltage) (16). The level of disturbances in telecommunication lines, which must be kept below the specified limits, may be assessed using the psophometric current IP (or voltage U P ), defined as (2)

where In (U n ) = nth current (voltage) harmonic ωn = weighting factor for the nth harmonic According to Refs. 2 and 7, the limits on psophometric voltage are: in public phone cables 0.5 mV to 1 mV; in railroad phones 1 mV to 2 mV (cabled) and 5 mV (open-wire). On railroads limits may be imposed on the psophometric value of the harmonic component of dc side voltage, as high as 0.5% of the nominal voltage in Poland or 10 V in Italy (both for 3 kV dc systems). Trackside equipment (TSE) (Fig. 28) is also vulnerable to inductively coupled disturbances caused by current flow in rails and on-board power circuits of an EV. Modeling and simulation methods are widely applied for prediction of disturbing interference effects and their elimination at the design state of an ETGS (6,7,8,9, 11), while for installations compatibility tests should be undertaken according to local standards and codes (2, 9,10,11, 17, 18).

BIBLIOGRAPHY 1. IEEE Trans. Vehicular Technol., Whole Journal (a special issue on maglev), VT-29 (1): 1980. 2. Anon., CCITT Directives Concerning the Protection of Telecommunication Lines Against Harmful Effects from Electric Power and Electrified Railway Lines. Induced Currents and Voltages in Electrified Railway Systems, Geneva: 1989. 3. K. G. Markvardt. Power Supply of Electrified Railways (in Russian), Moscow: Transport Press, 1982. 4. G. E. Littler, The performance of an ac electrified railway system at harmonic frequencies, Trans. Inst. Eng. Austral. Electrical Eng. Trans., EE16 (4): 184–198, 1980. 5. M. Krakowski, Earth-Return Circuits (in Polish), Warsaw: WNT Press 1979. 6. A. Szelag. W. Zajac, P. M. Martinez, Harmonic distortion caused by electric traction vehicles with a.c. motors fed by d.c. supply system—CAD analysis presented at the Eur. Power Electronic Conf. EPE’95, Seville, 1995. 7. B. Mellitt, et al. Computer-based methods for induced-voltage calculations in ac railways, IEE Proc., 137, Pt. B (1): 1990. 8. R. J. Hill, Electric railway traction, Power Eng. Journal, Part 1, pp. 47–56, Feb. 1994; Part 2, pp. 143–152, June 1994; Part 3, pp. 275–286, Dec. 1994; Part 4, pp. 201–206, Aug. 1995; Part 5, pp. 87–95, Apr. 1996; Part 6, pp. 31–39, Feb. 1997; Part 7, pp. 259–266, Dec. 1997. 9. C. D. Waters, et al. Dublina area rapid transit, IEE Proc., 135, Pt. B (3): 134–150, 1988. 10. V. D. Nene, Advanced Propulsion Systems for Urban Rail Vehicles, Englewood Cliffs, NJ: Prentice-Hall, 1985. 11. A. Mariscotti P. Pozzobon Harmonic distortion and electromagnetic interference in electric rail transport systems, Third Int. Sci. Conf. “Drives and Supply Systems for Modern Electric Traction,” Warsaw, 1997, pp. 131–136. 12. Papers presented at IEE Third Vacation School on Electric Traction Systems, University of Birmingham, 1995. 13. A. Szelag, Algorithms and tools for analysis of electrified railway lines, 28th Universities Power Eng. Conf., Staffordshire University, UK, 1993, pp. 354–357. 14. N. B. Rambukwella, B. Mellitt, W. S. Chan, Traction equipment modelling for AC and DC supplied railway systems using DC drives, Eur. Power Electronics Conf., Aachen, 1989, pp. 379–384. 15. D. C. Howroyd, Public-supply system distortion and unbalance from single-phase a.c. traction, Proc. IEE, 124 (10): 853–858, 1977.

GROUND TRANSPORTATION SYSTEMS

47

16. R. E. Morisson M. J. Barlow, Continuous overvoltages on a.c. traction system, IEEE Trans. Power Appar. Syst., PAS-102: 1211–1217, 1983. 17. Proposed 1996 Manual Revisions to Chap. 33, Electr. Energy Utilization, Part 6, Power Supply and Distribution Requirements for Railroad Electrification Syst., Amer. Railway Eng. Assoc., 97, 754, January 1996, pp. 255–272. 18. J. A. Taufiq, C. J. Goodman, B. Mellitt, Railway signalling compatibility of inverter fed induction motor drives for rapid transit, Proc. IEE, 133B (2): 71–84, 1986.

READING LIST Anon., Mass transit, Signal Technol. Today, Sept./Oct. 1996, pp. 36–49, 58–61. Anon., Functional Requirements Specification for ETCS v. 3.0, EEIG ERTMS User Group, Jan. 1996. J. F. Gieras, Linear Induction Drives, Oxford: Clarendon Press, 1994. W. Harprecht, F. Kiessling, R. Seifert, 406.9 km/word speed record-energy transmission during the record run of the ICE train of the DB, Elekir, Bahnen. 86 (9): 268-289, 1988. T. A. Kneschke, Simple method for determination of substation spacing for AC and DC electrification systems. IEEE Trans, Ind. Appl., IA-22: 763-780, 1986. L. Mierzejewski, A. Szelag, M. Galuszewski. Power Supply of DC Electric Traction Systems (in Polish), Warsaw: Warsaw University of Technology Press, 1989. M. Taplin, The history of tramways and evolution of light rail. Online: http://www/lrta.org/mrthistory.htm

ADAM SZELAG LESZEK MIERZEJEWSKI Warsaw University of Technology

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELECTRICAL...NICS%20ENGINEERING/62.%20Vehicular%20Technology/W7702.htm

}{{}}

●

HOME ●

ABOUT US ●

CONTACT US ●

HELP

Home / Engineering / Electrical and Electronics Engineering

Wiley Encyclopedia of Electrical and Electronics Engineering Mobile Radio Channels Standard Article Rodney G. Vaughan1 1Industrial Research Limited, New Zealand Copyright © 1999 by John Wiley & Sons, Inc. All rights reserved. DOI: 10.1002/047134608X.W7702 Article Online Posting Date: December 27, 1999 Abstract | Full Text: HTML PDF (603K)

●

●

● ●

Recommend to Your Librarian Save title to My Profile Email this page Print this page

Browse this title ●

Search this title Enter words or phrases

Abstract The sections in this article are The Mobile Channel Multipath Propagation Effects Channel Model using Discrete Effective Scatterers Statistical Basis of a Mobile Channel The Two-Path Model Statistical Approach Using Two-Path Model MAny-Path Model Path Loss and the Mobile Channel About Wiley InterScience | About Wiley | Privacy | Terms & Conditions Copyright © 1999-2008John Wiley & Sons, Inc. All Rights Reserved.

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELE...ENGINEERING/62.%20Vehicular%20Technology/W7702.htm15.06.2008 12:44:34

❍

❍ ❍

Advanced Product Search Search All Content Acronym Finder

J. Webster (ed.), Wiley Encyclopedia of Electrical and Electronics Engineering c 1999 John Wiley & Sons, Inc. Copyright


MOBILE RADIO CHANNELS The term mobile channel refers to the transfer function of adio link when one or both of the terminals are moving. The moving terminal is typically in a vehicle such as a car, or a personal communications terminal such as a cellphone. Normally one end of the radio link is fixed, and this is referred to as the base station. In the link, there is usually multipath radiowave propagation, which is changing with time, or more specifically, as a function of position of the moving terminal. The effects of this multipath propagation dominate the behavior and characterization of the mobile channel. The radio frequency of the link ranges from hundreds of kilohertz, as in broadcast AM radio, to microwave frequencies, as in cellphone communications. Indeed, even optical frequencies are used, as in an infrared link used for indoor computer communications. The kind of channel most often referred to as “mobile,” however, is that using microwave frequencies, and this article concentrates on the characteristics of a mobile microwave radio link. Much of the channel behavior can be scaled by the carrier frequency and by the speed of the mobile terminal. Current spectral usage is a result of many different historical developments, so the bands used by mobile radio channels have evolved to be at many frequencies. For example, current vehicular and personal communications terminals mostly use frequencies around 400 MHz, 900 MHz, and 1.8 GHz. In the future, higher frequencies will be used. The frequency has a definitive bearing on the rate at which the channel changes. Some examples of mobile channels include: domestic cordless telephones; cellular telephones and radiotelephones; pagers; satellite communication terminals, including navigational services such as Global Positioning System (GPS) reception; and radio networks for local data communications. Finally, the reception by portable receivers of broadcast radio, at frequencies of a few hundred kilohertz (AM radio) are common forms of the mobile radio channel. The use of mobile channels has grown very quickly in the last decade. This growth will continue. It is driven by a combination of consumer demand for mobile voice and data services and advances in electronic technology. A limiting factor to the growth is that many users must share the radio spectrum, which is a finite resource. The spectral sharing is not only local, it is also international, and so spectral regulatory issues have also become formidable. The increasing pressure to use the spectrum more efficiently is also a driving force in regulatory and technical developments. To a user, a mobile or personal communications system is simple: it is a terminal, such as a telephone, that uses a radio link instead of a wire link. The conspicuous result is that the terminal is compact for portability, and it has an antenna, although for personal communications the antenna is often no longer visible. To the communication engineer, however, the mobile terminal is just a component in a vast, complex circuit. The mobile channel is one link in the circuit, but this link is the most complex, owing to its use of radio waves in complicated propagation environments and of radio signal-processing technology needed to facilitate wireless transmission among multiple users. In mobile channels, efficient spectral utilization is a function of the basic limitations on controlling radiowave behaor in complicated physical environments, including the launching and gathering of the waves. Thus antennas and propagation are key topics, and their roles characterize the channel behavior. 1

2

MOBILE RADIO CHANNELS

Fig. 1. Various channels in a mobile communications link. The term “mobile channel” refers to the analog aspects of the channel, excluding modulation and coding.

The Mobile Channel The mobile channel covers many different transfer functions that have different properties. Figure 1 illustrates individual channels (1). The figure shows half of a link, with the other half essentially an inverse process. The multipath propagation environment represents the physical environment of the radio waves in the mobile channel. The flow of information is described here for transmission, but the description adapts readily to reception. A raw information channel refers to the transfer function that separates the transmitted and received raw information. For speech, for example, degradations of the immediate acoustical environment from reverberation and acoustic noise form part of the channel “seen” by the user at the receiving end. The quality of the information channel may be subjective, although standard metrics of distortion and signal-to-noise ratio can be applied for characterization. The electrical signal is often digitized for efficient transmission, and the digital channel is nonlinear, but its channel quality can be measured directly as a bit error ratio (BER). This digital form is sometimes rearranged by encoding techniques for more robust transmission of the information. The digital information is coded into analog waveforms and then mixed, or heterodyned, to the radio carrier frequency and transmitted via the antenna. The distinguishing feature of the mobile channel is the changing multipath propagation between transmitter and receiver. The receiving antenna gathers the many incident electromagnetic waves from the multipath environment. These multipath contributions mutually interfere in a random, time-varying manner, and so

MOBILE RADIO CHANNELS

3

statistical techniques are needed to characterize the channel. In the physical transmission media, the waves that bear the information are the signals of the electromagnetic propagation channel. The antenna reduces the signals from a vector form of orthogonal polarizations to a scalar voltage. The signal at the open circuited receiving antenna terminal is the output of the electromagnetic signal channel. The antenna needs to be terminated in order to maximize the power received by the front end. The signal-to-noise ratio (SNR) is established at this point in the link, and the resulting signal is the output of the radio channel. The antenna is a critical part of the mobile channel, and it can control much of the channel behavior. The baseband equivalent form of the radio channel, which is the radio channel shifted in frequency to a low-pass spectral position, is the signal that engineers use for mathematical characterization and most electronic (including digital) signal processing. The analog form of the radio channel is what will be referred to from here on as the mobile channel. Multiple Access for Mobile Channels. Most mobile communications systems are for multiple users, and a multiple access technique is required to allow the spectrum to be shared. In cellular systems, for example, the frequencies are reused at geographically spaced locations. For indoor systems, the frequency reuse spacing may be between floors. In a system design, the multiple access technique interacts with the choice of channel modulation and signal coding. The three basic techniques are frequency division multiple access (FDMA), which has channels occupying different narrow bandwidths simultaneously; code division multiple access (CDMA), in which multiple users share wider bandwidths simultaneously by using differently coded waveforms; and time division multiple access (TDMA), in which users share a bandwidth by occupying it at multiplexed times. Some systems employ a combination of these techniques. Multiple access is not a part of the mobile channel as such. However, the reader should remain aware that multiple access is part of the communations system and the choice of technique has an influence on the mobile channel bandwidth, its usage, and the type of signaling employed. Multiple access also brings in coand adjacent-channel interference, in which the unwanted signals at a receiver may not be noiselike, but in fact be signals with very similar characteristics to the wanted signal. In systems with densely packed users, the system capacity is interference-limited.

Multipath Propagation Effects Multipath radiowave propagation is the dominant feature of the mobile channel. More often than not, the transmitted signal has no line-of-sight path to the receiver, so that only indirect radiowave paths reach the receiving antenna. For microwave frequencies, the propagation mechanisms are a mixture of specular (i.e., mirrorlike) reflection from electrically smooth surfaces such as the ground, walls of buildings, and sides of vehicles; diffraction from edges of buildings, hills, etc.; scattering from posts, cables, furniture, etc.; and diffuse scattering from electrically rough surfaces such as some walls and grounds. Some multipath propagation occurs in nearly all communications links. The basic phenomenon is that several replicas of the signal are received, instead of one clean version. The result can be seen as television ghosts, for example. On transmission lines, reflections from mismatches on the line give the same effect, for example, echoes on telephone lines. On a long distance point-to-point radio link, a direct line-of-sight wave, a single ground bounce, and atmospherically refracted waves can all contribute to the received signal. When signal replicas are too close together to be discriminated and processed as discrete contributions, the received signal becomes distorted. This distortion limits the capacity of the channel. The phenomena is akin to the severe acoustic distortion known as the railway station effect, where increasing power output (volume) does not increase the intelligibility of the message. In digital communications, the distortion caused by multipath propagation creates an analogous effect: an increase in transmitted power does not decrease the BER as Shannon’s theorem might suggest. The amount and nature of the multipath propagation sets the level of power at which the BER becomes essentially independent of the SNR. The effect has often been referred to as the “irreducible BER,” but the use of signal processing, in particular equalization, can in fact reduce the BER

4

MOBILE RADIO CHANNELS

further. Experimental examples of the irreducible BER in the digital channel are given below, but this article otherwise concerns the analog mechanisms and the statistical nature of the mobile channel.

Fading in the Mobile Channel. Fast Fading. The interference, or phase mixing, of the multipath contributions causes time- and frequency-dependent fading in the gain of the channel. The time dependence is normally from the changing position of the mobile terminal, and so is also referred to as space dependence. At a given frequency, the power of the received signal, and thus the gain of the mobile channel, changes with time. This changing SNR is called signal fading and is often experienced as audible “swooshing” or “picket fencing” when an FM station (with a radio frequency of about 100 MHz) is received by the antenna on a moving car. If the mobile terminal is stationary, the signal may continue to experience some fading, and this is caused by changes in the multipath environment, which may include moving vehicles, etc. In nearly all situations the changing mobile position dominates the time variation of the mobile channel. Usually, the multipath environment is taken, or at least modeled, as unchanging. This is called the static multipath assumption. In this case, a static mobile experiences an unchanging channel. If now the radio frequency is swept, then the gain of the transfer function experiences fading similar to that due to changes in position, because the electrical path distances of the multipath components are frequency-dependent. For a continuous wave (CW) signal, the time- and frequency-dependent fades can be some 40 dB below the mean power level, and up to 10 dB above the mean. This indicates the large dynamic range required of the receiver just to handle the multipath interference. This fading is variously called the fast fading, or short term fading, or Rayleigh fading after the Rayleigh distribution of the signal magnitude. The maximum density of fading is a fade about every half wavelength on average, and this occurs typically in urban outdoor environments. The fast fading dominates the mobile channel characteristics and usage. For example, amplitude modulation at microwave frequencies is not feasible, because for a fast-moving mobile terminal the fading interferes directly with the modulation. Slow Fading. The dynamic range of the received signal is also affected by slow fading, also called long term fading or shadow fading. This is superimposed on the fast fading. It is caused by shadowing of the radio signal to the scatterers as the mobile terminal moves behind large obstacles such as hills and buildings. The rate of the slow fading therefore depends on the large scale nature of the physical environment. The basic short term multipath mechanism remains unchanged. The dynamic range of the slow fading is typically less than that of the fast fading, being confined to about ±10 dB for most of the time in urban and suburban environments. The total dynamic range for the fading therefore becomes about 70 dB. The distance-based path loss, as a mobile terminal roams near to and far from a base station, adds to this range. Narrowband and Wideband. In a typical mobile microwave signal link, the relative bandwidth is small. This means that the spectral extent of the signal is less than a few percent of the nominal carrier frequency. The fading within the frequency response of the transfer function is referred to as frequency-selective fading. If the bandwidth is sufficiently small so that all the frequency components fade together, then this is called a flat fading channel. In the mobile channel context, a narrowband channel has flat fading and a wideband channel has frequency-selective fading. The use of a single frequency, or CW, for channel characterization is the limiting case of the narrowband channel. Historically, fading has been the principal observed characteristic of the mobile channel. Fast fading is merely one manifestation of the reception of several replica signals.

The Effect of Fading on the Digital Channel: Irreducible Bit Error Ratio. Timing Errors From Random Frequency Modulation. The digital channel in Fig. 1. is in principle the simplest channel to characterize experimentally, since it concerns a BER measure. The fading in the mobile channel has a particular effect on the BER curves, namely, the “irreducible BER” mentioned above. The example in Fig. 2. (2) shows curves of BER against carrier-to-noise ratio (CNR) from simulations of the narrowband mobile channel with carrier frequency 920 MHz. The static (no fading) curve shows the classical waterfall shape of the Gaussian channel. But the fading channel curves, shown with fading rate f D , feature irreducible

MOBILE RADIO CHANNELS

5

Fig. 2. The irreducible BER for a digital mobile channel is attained when an increase of SNR does not improve the BER. The static (no fading) channel shows the classical waterfall shape of the Gaussian noise-limited channel, but as the fading rate increases, the form of the curve alters drastically. From 2.

BERs, which occur at lower CNRs with increasing fading rate. The fading rate of 40 Hz corresponds to a mobile speed of about 40 km/h and a carrier frequency of 900 MHz. This corresponds approximately with using a cellphone from a moving car. The curves hold their basic form independently of the type of angle modulation used. The mechanism for the bit errors is timing error caused by the random FM, discussed below, imposed on the signal by the fading channel. The random FM causes jitter on the symbols after they passed through the mobile channel. Intersymbol Interference From Multiple Time Delays. As the signaling rate increases, an analogous irreducible BER effect occurs as a result of the several signal replicas arriving at different times. This spread of delays causes intersymbol interference when one dispersed symbol overlaps with other, similarly dispersed symbols. In analog parlance, this is called dispersive distortion. In the mobile channel the situation is complicated by the dispersion changing with time. The effect is depicted in the experimental example of Fig. 3. (3), where for a fixed fading rate of f D = 40 Hz the increasing digital transmission rate experiences an increasing irreducible BER. As in Fig. 2, the effect is that the capacity of a given link cannot be increased by simply increasing the CNR, for example, by increasing the transmitted power. Signal processing is required. Signal Processing for Mitigation of the Multipath Effect. Several signal-processing techniques can be applied to the mobile channel to reduce distortion and recover the capacity relative to the static channel.

6

MOBILE RADIO CHANNELS

Fig. 3. The irreducible BER caused by intersymbol interference. As the fading rate increases relative to the spread of multipath propagation delay times, the irreducible BER increases. From 3.

Equalization and rake systems basically attempt to gather the delayed signal replicas and recombine them into a single signal, which, ideally, is no longer distorted or faded. Antenna diversity uses multiple antenna elements to receive the same signal but with different multipath degradations, and combines the signals so that the resultant channel has better capacity than any of the channels from the individual antenna elements. A combination of the equalization, or rake, and antenna diversity methods is called space-time processing. All these techniques can be effective in improving the mobile channel. In fact, the use of antenna diversity offers very large potential capacities by effectively reusing the frequency at different positions in space. The Mobile Channel as a Transfer Function. Figure 4 depicts a static mobile channel, which is taken as the baseband equivalent radio channel of Fig. 1. Recall that the effect of the antennas is included in the transfer function. The impulse response h(τ) and the transfer function H(ω) are related by Fourier transformation in the usual way, denoted h(τ) ⇐⇒ H(ω). Here τ is the delay time and ω is the angular baseband equivalent frequency. The impulse response indicates the dispersive nature of the channel, which causes distortion of the signals which are transmitted through it. This impulse response is modeled as a series of discrete delta functions below. The example of Fig. 1 is for an instant in time, t. As the mobile terminal moves, the delays and phases of the individual multipath contributions become functions of time. The impulse response and transfer function therefore become expressed mathematically as functions of time, that is, h(τ, t) and H(ω, t). If the scatterers in the multipath environment can be considered to be essentially stationary, then the time t and position z are related by the velocity V of the mobile: z=Vt. From now on the spatial variable z will be mostly used.

MOBILE RADIO CHANNELS

7

Fig. 4. The static mobile channel transfer function. x(τ) and y(τ) are electronic signals before the transmitting antenna and after the receiving antenna, respectively. The impulse response can be found by Fourier transformation of a swept frequency measurement, for example.

The following sections will develop, through the use of several assumptions about the channel, a double Fourier transform relation between the impulse response as a function of delay time and time (i.e., position) and the transfer function as a function of baseband angular frequency and Doppler frequency. Because of the variation of the transfer functions, the statistical parameters of the channel are relevant, and these also can be couched in terms of Fourier relations. The Receiving Antenna in Multipath Transmission. The moving antenna combines the radiowave contributions, which have continuously changing delays, amplitudes, and polarizations. Deterministic analysis is not feasible except in simplistic situations, and to be able to interpret the statistical description requires an appreciation of multipath phenomena. A base station transmitter is taken to emit power in a fixed radiation pattern. After multiple scattering, for example from many reflections, the polarization is changed in a random way and the electric (and magnetic) field has all three Cartesian components, independent of the transmitted polarization. These components can be independent functions of frequency and position. So the total incident electric field, at a point in space, can be written in baseband equivalent form [i.e., with a complex envelope, in which a factor of exp(jωC t) is suppressed, where ωC is the carrier frequency] as the complex vector

in which the components, such as Ex , are complex scalars.

8

MOBILE RADIO CHANNELS

The introduction of an antenna promotes a change to spherical coordinates referred to the antenna orientation and position. The position is denoted with the single spatial variable z. The incident fields are now written as

The open circuit voltage of an antenna depends on both the incident field and the receiving pattern, ha (ω θ, φ) = hθ (ω θ, φ) + hφ (ω θ, φ) . This notation for the receiving pattern should not be confused with the symbol for the impulse response, h(τ,z). The open circuit voltage is defined by

and represents the transfer function of the electromagnetic signal channel. By expanding the dot product, this transfer function is written in terms of the incident field components, which are now collectively detected as standing waves, and the receiving pattern components, as

This formula shows the inseparability of the antenna pattern and the incident fields in the definition of the mobile channel. The antenna pattern is recognized as a filter in the spatial (including polarization) domain. The frequency dependence of the antenna pattern also represents a filter in the more familiar frequency domain. The space–frequency filter of the antenna is the difference between the vector electromagnetic propagation channel and the scalar electromagnetic signal channel of Fig. 1. If terminating (i.e., matching) the antenna has a negligible effect over the band of interest, then Eq. (4) represents the mobile channel.

Channel Model using Discrete Effective Scatterers Modeling the incident waves as emanating from discrete directions allows the convenience of using effective point sources. These are referred to as effective scatterers, because their scalar contribution is the physical incident wave weighted by the receiving pattern. The transfer function is written as the sum of effective scatterers, which have an amplitude a, a phase ψ, and a delay time τ for the information carried, that is,

Here the radio frequency is the sum of the carrier frequency (the center frequency of the radio band) and baseband equivalent frequency, i.e.,

In the static situation, the terms in the transfer function containing the delays are constant and can be incorporated into the phases of the effective scatterers.

MOBILE RADIO CHANNELS

9

Fig. 5. Point source with moving receiver. In a model of the channel, Eq. (5), the point source is not necessarily a physical scatterer, but can rather be considered as a point representation (an “effective scatterer”) that produces the waves received from a given angular direction.

The effect of the moving terminal on the transfer function can be seen by considering an effective scatterer at a relatively large distance r0 from it. The geometry is shown in Fig. 5. The mobile terminal moves a distance z along the spatial axis in the positive direction. The electrical distance to the i th effective scatterer changes from kR rθi , where kR is the radio frequency wavenumber, to

where

is the spatial Doppler frequency in radians per meter. The Doppler frequency in radians per second is

10

MOBILE RADIO CHANNELS

Here ui is a scaled directional cosine to the ith effective scatterer, and a receiver movement z produces a phase shift uiz in the signal from the scatterer. The changing phase term of an effective scatterer at position z in Eq. (5) is

The first term is independent of the position and baseband frequency, and can be incorporated in the phase of the scatterer. The last term is negligible, because in microwave communications we normally have a small relative bandwidth (i.e., ω/ωC 1 ). So within the above approximations, the transfer function is

Fourier transformation with respect to the baseband frequency ω gives the position-dependent impulse response as a function of the delay time and position,

A further Fourier transformation, this time with respect to the position z, gives a function of delay time and spatial Doppler frequency, denoted

Fourier Transform Relations With Continuous Transfer Functions. The Fourier pair a(τ,u) ⇐⇒ H(ω,z) have the continuous form

Note the mixed signs of the exponents. Moving in the negative z direction instead of the positive z direction, for example, changes the sign of the exponent zu in Eqs. (14) and (15). From the double Fourier transform relation, there can be four complex functions that carry the same information for characterization of the mobile channel. These are denoted: a (τ, u), the scattering function in the time-delay–spatial-Doppler domain (referred to as the effective scattering distribution) h (τ, z), the impulse response in the delay–space domain (spatial spectrum) A (ω, u), the transfer function in the baseband-frequency–spatial-Doppler domain (frequency spectrum) H (ω, z), the transfer function in the baseband-frequency – domain (space–frequency spectrum)

MOBILE RADIO CHANNELS

11

The functions are related by the following single-dimensional Fourier transforms of the mobile channel:

The amplitudes, phases, delays, and directions of the effective sources are randomly distributed, and the transfer function consequently behaves randomly, so a statistical approach is called for their characterization. Averaging Across a Transfer Function For Channel Gain. In terms of an individual channel transfer function, the total power, or channel gain, is given by

where L is an averaging distance or locus covering the positional averaging, and ωB is an averaging bandwidth. Any of the above channel functions can be used to get the power in this way (Parseval’s theorem). Integrating single variables gives the frequency-dependent power transfer function averaged over position,

and the position-dependent- (time-dependent)- power transfer function averaged over the frequency band,

This quantity is approximated in a receiver by the position-varying (or time-varying) received signal strength indicator (RSSI) signal. However, in practice, the RSSI voltage is normally proportional to the logarithm of the channel power.

12

MOBILE RADIO CHANNELS

On averaging the power across a wideband channel, the total received power fades less than a narrowband component. This is the advantage of wideband modulation systems. Analogously, antenna diversity is used to reduce the fading by averaging the channel over samples of the spatial variable.

Statistical Basis of a Mobile Channel Power Spectra and Channel Correlation Functions. Assuming ergodicity so that the statistics remain second order, the autocorrelation function, denoted by R, of the effective scatterer distribution with respect to the delay times is written

where the angular brackets denote averaging over all relevant realizations of the effective scattering distribution. This contrasts with the averaging over frequency or space for a single channel realization as in the previous section. The average power in the effective scattering distribution is

Note that the averaging is of the powers, not of the complex values, of the a(τ, u) . This averaged power distribution can be expressed in several different statistical forms as seen below. Substituting Eq. (16) into Eq. (23) gives the Fourier transform

The inverse relation is

Wide Sense Stationarity in Frequency. The channel is now assumed to be wide sense stationary in the frequency domain. This means that the mean and correlation of A(ω, u) do not depend on the choice of frequency, ω, but on only the frequency difference, ω = ω2 −ω1 . This is a reasonable assumption for the frequencies within the small relative bandwidths of most mobile communications systems. Denote the autocorrelation of a wide sense stationary (WSS) process using S, for example, by

i.e., the autocorrelation of the transfer function in the frequency–spatial-Doppler domain is a power spectrum whose argument is the frequency difference. The symbols S and R are used to represent the correspondence of the power spectra S and the autocorrelation R of a process that is WSS. As a result of the

MOBILE RADIO CHANNELS

13

wide sense stationarity in ω, we can write Eq. (25) as

where

is the averaged power delay–Doppler-frequency distribution of Eq. (24). The delta function in the autocorrelation of Eq. (28) is referred to as the uncorrelated scattering (US), and here means that a fading signal received at a given delay time is uncorrelated (when averaged over the relevant realizations) with a fading signal received at any other delay time. The wide sense stationarity (via the ω factor) in the baseband frequency domain and the uncorrelated scattering in the delay time domain [the δ(τ) factor] are equivalent characteristics. Wide Sense Stationarity in Space. Similarly, wide sense stationarity in the spatial domain corresponds to uncorrelated scattering in the Doppler domain. This means that the fading signal at one spatial Doppler frequency u [or angle θ = cos − 1 (u/kC ) ] is uncorrelated with a fading signal received from any other spatial Doppler frequency. Denoting the spatial difference z = z2 −z1 , we have

where the averaged power of the effective scattering distribution is expressed as

Wide Sense Stationary Uncorrelated Scattering Channel. Combining the space and frequency wide sense stationarity conditions, we have

where now

14

MOBILE RADIO CHANNELS The inverse Fourier transform is

Thus the above wide sense stationarity conditions result in the frequency–space correlation function being the double Fourier transform of the average power density of the effective scatterer distribution. The term WSSUS was used by Bello (4) to describe tropospheric multipath channels containing scintillating scatterers being illuminated by static antennas. In the context of the mobile channel, the WSS refers to wide sense stationarity in position, which implies uncorrelated scattering in the spatial Doppler frequency. The US refers to the delta function in delay time (effective sources at different delays are mutually uncorrelated), which implies WSS in the frequency domain. The assumption of the WSSUS conditions in the channel allows the convenience of the double Fourier transform relations. However, in applying the Fourier relations for a given situation, the validity of the WSSUS model should always be questioned. The channel can often be arranged to be “sufficiently valid” for gaining useful insight and inferring channel behavior, by appropriately arranging the averaging. This averaging, denoted with the angular brackets, is often taken as several sampled records over short distances (tens of carrier wavelengths or several tens of fades) in order to stay within a given physical environment, followed by the power distribution averaging. Statistically, ensemble averaging implies many “realizations.” We can interpret this as several sampled records that should have uncorrelated data (e.g., well separated spatial paths) within the same physical environment, or else as several records in different (i.e., independent) physical environments. The two cases are different. One case averages within a single environment; the other case averages over many different environments. Strictly speaking, the presence of multiple uncorrelated records in the same immediate environment does not truly satisfy the hypothesis of statistically independent records, because the scattering distribution is the same, that is, the signal sources constituting the physical scatterers are common to all the data records. Key Relations For a Mobile Channel. Equations (14), (15) and ((33), (34) are key results for the mobile channel. They relate, respectively, by double Fourier transformation, a baseband channel transfer function H(ω, z) to an effective source distribution a(τ, u) that provides the incident multipath signals, and the average power spectral density of the channel SH (ω, z) to the average power distribution of the effective scatterers, P(τ, u) Figure 6 (1) depicts the relations between the functions. Averaged Power Profiles. The more familiar single transformations also are of interest. Mathematically, we can put z = 0 in the frequency correlation, that is,

from which Eq. (34) reduces to

where the average power delay profile,

is the average power at delay τ, found by integrating over all spatial Doppler frequencies ( u = kC to u = −kC ), that is, in all directions over the averaged power of the effective scattering distribution. In practice, the

MOBILE RADIO CHANNELS

15

Fig. 6. Fourier transform relations for the mobile channel functions and for their statistical representations under wide sense stationarity in frequency and position. u = kC cos θ is the spatial Doppler frequency, with θ the zenith angle with respect to the direction of motion z, and kC the wavenumber of the radio carrier frequency. From Ref. 3.

antenna performs this integration [recall that the effective scattering distribution P(τ, u) already includes the effect of the antenna] for example, an omnidirectional antenna will gather the waves from all the directions. However, a single measurement from an antenna only accounts for a single realization of the effective scattering distribution—that is, for one point in the space of one environment. To estimate P(τ) from measurements, the averaging of the profile needs to be done over several different positions [i.e., several z0 values in Eq. (35)], either in the same physical environment or in many different physical environments, as discussed above. The frequency correlation function SH (ω) is the Fourier transform of the average power delay profile P(τ) for the WSS channel with uncorrelated scattering. The inverse relation is

The Fourier relation in Eqs. (36) and (38) is identical to the relation between the transfer function and its impulse response, as in Fig. 4. Similarly to the average delay profile, the average spatial Doppler profile is averaged over all delays:

P(τ) and P(u) are sometimes called the delay spectrum and the Doppler spectrum respectively. Finally, the total power of the effective scatterers is given by

Many details, extending to situations outside the mobile channel, may be found in Ref. 3.

16

MOBILE RADIO CHANNELS

Spreads. The spread, or second centralized moment, of a distribution is a standard characterizing parameter. For an instantaneous (i.e., snapshot, or unaveraged) channel distribution function, the instantaneous spread is the standard deviation of that function. For example, for a channel with a snapshot transfer function h(τ), the definition of the instantaneous delay spread is

The (average) delay spread, denoted στ , follows the same definition but uses the averaged distribution P(τ) = |h(τ)|2  instead of |h(τ)|2 . The analogous definition for the Doppler spread is

It is important to note that it is the individual power distributions that are averaged to produce the power profiles, which are then used to produce the spreads. It is wrong to calculate the spreads of individual channels, average these, and call the result the average spread. Power Profile Examples. Two power profiles that are commonly used for modeling because of their simplicity, are the one-sided exponential,

and the two-path,

which are shown in Fig. 7. The exponential is the most commonly used model. The two-path model offers much insight into the mechanisms of the channel and is used in the following to develop the basic characteristics and parameters of interest of the mobile channel’s behavior. It is later extended to the many-path situation. Liberties are taken with the mathematical use of the delta functions to allow convenient modeling.

The Two-Path Model The two-path model and its “statistics” (the model is treated statistically despite the situation being deterministic) produce and explain nearly all of the behavior that can be found in real world mobile channels. Such a model is also used in point-to-point communications where there can be a direct wave with a single ground bounce. The term “two-path” refers to two effective sources. However, the introduction of the directions of the effective sources is delayed until later, since the directions have no bearing on the received signal while the

MOBILE RADIO CHANNELS

17

Fig. 7. Examples of the exponential and the two-path models for the power delay profile. The two-path model comprises discrete multipath contributions, whereas the exponential profile has a continuum of multipath contributions.

Fig. 8. The impulse response of the static two-path model amplitudes 1 and a2 , and a complex plane representation of the transfer function for the case a2 > 1.

receiver is static. The moving receiver introduces a changing frequency dependence, and the rate of change is determined by the directions. Understanding the behavior of the static model allows a smooth transition to understanding the many-path channel behavior. Static Model for Frequency-Selective Fading. The two-path scenario is shown with its variation with frequency in Fig. 8. The impulse response, on setting τ1 = 0 and α1 = 0 for the first path, is

and so represents a signal arriving with zero delay with normalized magnitude and zero phase, and a signal arriving at a delay of τ2 with magnitude a2 and phase α2 . The transfer function is minimum phase when a2 ≤ 1, and is maximum phase when a2 > 1 .

18

MOBILE RADIO CHANNELS

This model, being static in the sense that the two effective scatterers are constant in amplitude and phase, needs no averaging to obtain the power profile. So P(τ) =  |h(τ)|2  = |h(τ)|2 for the static case. The delay spread is thus the same as the instantaneous delay spread, and from Eqs. (44) and (41), is στ (2) = a2 τ2 /(1 + a2 2 ). The delay spread is not affected by time reversal or magnitude scaling of the power profile. In the two-path case, this means that a2 can be replaced by 1/a2 (i.e., a change from a minimum- to a maximum-phase channel) and the delay spread stays the same. Transfer Function. The transfer function is obtained by Fourier transformation of Eq. (45), and is (the factor 1/2π is omitted for brevity)

where the delay difference is τ = τ2 −τ1 = τ2 . The in-phase component is the real part of the transfer function, I(ω) = 1 + a2 cos(ωτ2 −α2 ), and similarly the quadrature part is Q(ω) = a2 sin(ωτ2 −α2 ). Apart from the dc term, these are simply quadrature sinusoids with different amplitudes. The phase of the second effective scatterer, α2 , is now set to zero for brevity. The power transfer function is |H (ω)|2 = 1 + a2 2 + 2a2 cos(ω τ2 ), and so the frequency fading behavior is periodic with period 1/τ2 (Hz). The phase of the transfer function is

which has a maximum rate of change when the power is a minimum. For the case a2 ≤ 1, the maximum and minimum values of the phase are ±sin − 1 a2 . When a2 = 1 and ωτ2 = nπ ( n is an integer), the phase changes by π. Group Delay. The group delay of a transfer function is the negative derivative of the phase with respect to frequency, τg (ω) = −∂φ(ω)/∂ω. It approximates the time delay of the envelope of a narrowband signal after it has passed through a transfer function with phase φ (ω) (5). Changes in the group delay mean changes in the expected arrival times of information, such as symbols, at the receiver. For a channel that contains many delay values, the received signal becomes distorted owing to the dispersion. For the two-path model, the group delay is found by differentiating Eq. (47) to be

For the minimum phase case, this varies between a2 τ2 /(a2 − 1) and a2 τ2 /(a2 + 1). If different frequencies were sent through the channel, then these values are the extrema of the group delays that would be experienced. Figure 9 shows the in-phase and quadrature signals, the envelope and phase, and the group delay for the transfer function of a static two-path model. Fearesf the Static Two-Path Model. The features from this deterministic model are frequency dependence with: • • • • •

Smoothly varying in-phase and quadrature components Fading envelope Sharp transitions of the phase of the transfer function, occurring when the envelope is at a minimum Possibility of both minimum-phase fades ( a2 ≤ 1) and non-minimum-phase fades (a2 > 1) Dispersive channel with sharp spikes in the group delay at the envelope minima

MOBILE RADIO CHANNELS

19

Fig. 9. The periodic frequency selective channel behavior for the static two-path model. The receiver is at a fixed position. The magnitude shows fading, the phase is changing quickly at the fade frequencies, and the group delay is correspondingly large (and negative for a2 ≤ 1) at the fade frequencies.

These transfer function variations are all periodic in the two-path model, but as seen below, the same effects occur also in the real world channel, but with a random frequency and space dependence. The reason for the phase behavior coinciding with the envelope is best seen from the locus of the signal in Fig. 8, where the envelope minima occur as the locus is passing closest to the origin, which is also when the phase is changing the quickest. For deep fades, the phase change is always nearly ±π (the sign depends on whether a2 is less than or greater than zero), and such phase jumps are also a characteristic of the many-path channel. Moving Receiver. In a moving receiver, we can fix the frequency to a CW for simplicity and get behavior as in the static channel of Fig. 8, but with spatial (i.e., time, for a given mobile speed), instead of frequency, dependence. For a CW channel, the transfer function is

where u = kC (cos θ2 −cos θ1 ) is the spatial Doppler frequency difference between the two effective sources. The transfer function now has spatial periodicity with a period (in meters) of 2π/u. For example, with sources exactly in front of (θ1 = 0) and behind (θ2 = π) the moving receiver, the periodicity is given by a spacing of exactly z = λC /2, that is, half the carrier wavelength. Random Frequency Modulation. The spatial analogy to the group delay is the random FM, given in radians per meter by the derivative of the phase with respect to position as ωR (z) = 2π ∂φ(z)/∂z. The random FM is an angle modulation in the channel and will be applied to a signal borne by the channel. It means that angle-modulation systems are affected as the receiver moves. In practice, the random FM is often too small to be noticed in a working system, but as carrier frequencies increase, the fading rate and the spectrum of random FM increasingly invades the signal band. In summary, the CW spatial mobile channel follows the same behavior as that in the frequency-dependent static channel, the transfer function signals shown in

20

MOBILE RADIO CHANNELS

Fig. 9. apply with the abscissa ωτ2 replaced with z u, and the group delay becomes the random FM (with the opposite polarity). Two-Dimensional Transfer Function. The frequency and spatial dependences can be combined to give the two-dimensional transfer function, again with α2 = 0 ,

which explicitly indicates the two-dimensional nature of the fading. The range of angles, u, determines the spatial fading rate, and the range of delay times, τ = τ2 , determines the rate of fading in the frequency domain. The statistical equivalents of these quantities, the Doppler spread and the delay spread, are used for describing the average fading rates found in the real world many-path situation.

Statistical Approach Using Two-Path Model The statistical approach is required when there are too many paths to determine the channel, which is normally the case in mobile communications. The statistical approach to the two-path model also offers insight into the statistical behavior of the many-path case. In the static case, the transfer function of the two-path model assumes all its possible values as the relative amplitude a2 and phase α2 are varied. In practice, averaging is over the phase-mixing process, so here we fix the amplitude and average over the changing phase only. In the static case, the phase of the frequency-dependent transfer function can be changed by changing the frequency. In a mobile channel, the fixed-frequency transfer function is averaged over the varying phase by averaging over many positions. Since the two-path transfer function has a symmetric, periodic envelope with half period π/τ2 (rad), equally likely frequencies are expressed by a uniform probability density function (pdf) over one of the periods,

The analogous expression for the moving receiver holds for equally likely positions, viz., pz (z) = u/π. These pdfs allow the pdfs of the channel function to be calculated below. Probability Density Function of Channel Power. For a2 < 1 and equally likely frequencies, the pdf for the power γ(ω) = |H(ω)|2 is, from function transformation of pω ,

where 1 + a2 2 and (2a2 )2 are the mean and variance respectively of the power in the two-path channel. Cumulative Probability Function of Channel Power. The cumulative probability function (cpf) is the integral of the pdf over its range of values (1−a2 )2 to (1 + a2 )2 , and is written

MOBILE RADIO CHANNELS

21

Fig. 10. The cpf for the power of the n = 2, 3, 4, 8 channels, where all the multipath amplitudes are the same. The n = 8 model is essentially the same, for the cpf range displayed, as Rayleigh (n → ∞) distribution, given in Fig. 14.

This probability approach is an alternative to the deterministic form H 2 (ω) for characterizing the two-path channel. The approach is needed when a deterministic form is not available. The cpfs for the n-path model with all the an = 1 are given in Fig. 10 for n = 2, 3, and 8. The 8-path case is very close, except at the tails of the distribution, to the Rayleigh distribution, which corresponds to the limiting case n → ∞, discussed further below. The pdf for the two-path case is centered at its mean, 1 + a2 2 , and is confined to its limits, that is, between (1 − a2 )2 and ( 1 + a2 )2 . At these limits, the pdf pγ 2 goes to infinity. The many-path pdfs can behave the same way. This does not cause interpretation problems, however, since the probability of the power being at these limits is infinitesimal and the integral of the function of course maintains its unity value. For example, for a2 = 1, the fades go exactly to zero in the transfer function. In the cpf of Fig. 10, the interpretation is that there is an infinitesimally small probability of the power being zero, that is,:

A similar situation holds for the power approaching its maximum value (1 + a2 )2 :

In the a2 = 1 two-path example, the cpf diagram shows that for 10% of the frequencies the power transfer function is more than 13 dB below its mean value. The cpf curves are arranged so that the mean power always corresponds to 0 dB. A flat channel ( a2 = 0 ) would be represented by a line at γ 0 = 0 dB. In summary, it is the phase difference between the source contributions that is the generic random variable for the statistical approach to the short-term variation of the power or envelope. In the static scenario, the averaging over the phase difference is implemented by varying the frequency. For the moving receiver case,

22

MOBILE RADIO CHANNELS

Fig. 11. The definition of a coherence bandwidth, C , in terms of the frequency correlation coefficient function, or coherence function, and a correlation value of CC . Narrowband channels, separated by a minimum frequency C , will display mutually uncorrelated fading in the sense that the correlation coefficient is Cc < ∼0.75.

the CW transfer function is averaged over space. In the general case, the transfer function is a two-dimensional distribution with phase mixing causing fading in both frequency and position. Coherence Bandwidth. An important parameter in a frequency-selective fading channel is the frequency separation for which the fading becomes effectively independent in the statistical sense. This frequency separation is determined by the autocorrelation of the channel transfer function. It is presented here as independent of frequency, that is, the channel is assumed to be WSS. The frequency correlation coefficient function, sometes referred to as the coherence function, is

and for the static two-path model with a2 = 1, the magnitude of this is

The coherence bandwidth C (rad/s) is defined as the frequency span from the maximum (unity) of the frequency correlation coefficient function to where the magnitude of the function first drops to a value CC , that is,

as illustrated in Fig. 11. CC is taken by various authors from 1/e = 0.37 to 0.9 (6,7,8). A change of CC scales the coherence bandwidth nonlinearly, so any results derived from some value of CC are also scaled in some way. The coherence function is periodic in ω for the two-path channel, since H (2) ω is periodic. (2) C minimum for a2 = 1, and for this case, the coherence bandwidth in hertz, (2) C /2π, can be written directly from Eqs. (57) and (58) as

MOBILE RADIO CHANNELS

23

The coherence bandwidth decreases with increasing delay difference between the two-path contributions, τ2 . Also, the coherence bandwidth decreases with increasing relative amplitude a2 . When a2 is small, the coherence bandwidth becomes undefined, as the coherence function does not drop down to CC . Product of Coherence Bandwidth and Delay Spread. While the delay spread is a measure of the channel time dispersion, the coherence bandwidth is a measure of the fading rate with changing frequency. The ideal communications channel has a zero delay spread and infinite coherence bandwidth. For the two-path model, the delay spread increases while the coherence bandwidth decreases for increasing relative delay τ2 and increasing relative amplitude a2 . The coherence bandwidth and the delay spread are thus inversely related, but the exact relationship is not simple in the many-path case. The product of these two parameters was taken for experimental channels using CC = 0.75 (7), and an empirical law was found that Bστ was constant and approximately equal to 1/8 (Gans’s law). The constancy of the product can also be viewed as an uncertainty principle (5,9). It gives a lower bound for the many-path channel as

The equality holds for the two-path case with equal powers, as in Eq. (59), which corresponds to maximum delay spread. For the two-path channel, the product BC (2) σT (2) does not exist. The dependence of this product on a2 , is weaker than its dependence on the choice of CC . The product BC (2) στ (2) is a minimum when a2 is 1, that is, when the frequency fades are the deepest. In this case and for the value CC = 0.75, the two-path product is in close agreement with Gans’s law, BC,Hz (2) στ (2) = (1/2π) cos − 1 0.75 ≈ 18 . In the two-path model, then, the virtually constant value of the product allows the delay spread to be calculated from a measured correlation bandwidth, or vice versa. However, in a general many-path case, the expression for the coherence-bandwidth–delay-spread product must be heeded as a lower limit. It should always be borne in mind that the choice of CC for the coherence bandwidth affects the value of the product. Because the delay spread is mathematically unbounded in the model (no limit is placed on τ2 ), there is no theoretical upper limit for Bστ in the many-path case, even though the coherence bandwidth can simultaneously remain essentially constant. In practice, physical and practical considerations such as the space loss described below are imposed on the model and the delay spread and the product become bounded through these. Correlation Distance. The correlation distance is the spatial counterpart of the coherence bandwidth. It is traditionally defined as the spatial displacement dd = z corresponding to when the spatial correlation coefficient, defined at a given frequency, decreases to some value. Instead of using the complex transfer function H(z), analogously to using H(ω) for the coherence function, the envelope correlation coefficient function,

has been used traditionally, and the coefficient value is taken as ρr(dd ) = 0.7. The correlation distance is a measure of the spatial fading rate and therefore depends inversely on the spatial Doppler spread σu. The product of these, dd σu, is lower bounded, but not with the same relationships as Bστ .

24

MOBILE RADIO CHANNELS

MAny-Path Model The above discussion has touched several times on the many-path model. Many channel parameters for the three-path model can be derived deterministically. The three-path model has been of interest in point-to-point links because it matches the physical situation of a direct, a ground bounce, and a single atmospherically diffracted ray. It has been also used to help “randomize,” relative to the two-path model, a transfer function for a more realistic-looking (over two or three fades), but tractable, model. However, it otherwise offers little more insight into the channel behavior than does the two-path model. The statistics for the few-path (less than about 10) less than about) model are rather complicated. When there are more than about 10 components of similar amplitude, however, the statistics follow, to a good approximation, the limiting case of a very large number of paths. The phase-mixing process of adding many random phasors gives, from the central limit theorem, the classical Rayleigh channel. The distribution functions are given below. Phase Mixing with Many Random Contributions. Equations (11) and (12) describe the model. For a narrow band channel, the in-phase and quadrature components are Gaussian distributed from the central limit theorem. It follows that: the distribution of the power is chi-square with two degrees of freedom (i.e., exponential), the envelope is Rayleigh-distributed, and the phase is uniformly distributed. The transfer function signals, as a function of position, are depicted in Fig. 12. The incident power is from all directions for this example. The figure can be compared with the signals from the two-path model, shown as a function of frequency in Fig. 9. The features of the channel are essentially the same as those in the two-path model, although the process is random. There are both minimum-phase and maximum-phase deep fades. Similarly, the random FM spikes have an associated polarity that is random. Rayleigh Envelope and Uniform Phase. The signal representing the channel transfer function is represented as a complex Gaussian process. The in-phase component and quadrature components are denoted x and y, the envelope r, and the phase θ, and these are related as

Here x and y are independent, zero mean Gaussians, so the pdf for each is (here for x)

where σ is the standard deviation of each component. The envelope and phase pdfs are established as independent with Rayleigh and uniform distributions respectively, through the steps

The pdf of the phase is 1/(2π), so the mean phase is π and the standard deviation is π/3. The averaged power is

MOBILE RADIO CHANNELS

25

Fig. 12. The signals of a many-path, narrowband channel as a function of position. As the mobile receiver moves, the narrowband signal quantities vary in a way similar to the behavior of the plots. The in-phase and quadrature components comprise complex Gaussians, the magnitude or envelope is Rayleigh-distributed, the phase is uniformly distributed, and the random FM is Student-t distributed.

and r2 is recognized as having a chi-square distribution with two degrees of freedom,

The Rayleigh statistics are included in the more general Rice statistics, below.

Rice Envelope and Phase. Sometimes, there is a single dominant effective source. This usually corresponds to a line-of-sight situation, which gives a single dominant effective scatterer. Multipath transmission still occurs, and the Rice distribution describes the statistics of the narrowband envelope. The Rice distribution results from one or both of the Gaussian processes having nonzero mean. These

26

MOBILE RADIO CHANNELS

Fig. 13. The Rice process has envelope rRi comprising the additive constant rs and the Rayleigh envelope r. The phase of the Rice signal is θ.

phase processes become

where the x and y are zero-mean Gaussian and xs and ys are the respective means representing the dominant component (sometimes called the specular, or coherent component, with x and y representing the diffuse, or incoherent component) of the signal. The phasor combination is shown in Fig. 13. in which φ = tan − 1 (yRi /xRi ) is the absolute phase of the Rice envelope rRi , and θRi is the phase difference between rRi (Rayleigh component plus dominant component) and the dominant component rs . The mean of the absolute phase of the process is (θRi + φ). A coordinate rotation allows the phase to bdefined as just θRi . From

the Rice pdf is

The envelope and phase are thus statistically dependent, unlike the Rayleigh case. The +π and −π transitions that occur in the phase of the Rayleigh signal as the locus passes near the origin are now reduced to smaller values, which depend on the length of the envelope phasor component rRi . The Rice channel can be purely minimum phase when the dominant component is large enough. The Rice k factor is the ratio of powers of the dominant component and the Rayleigh component,

When the dominant component rs approaches zero, kRi approaches 0, and the distribution reduces to Rayleigh. Similarly, when the dominant component becomes very large, the Rice distribution for the envelope approaches Gaussian with mean rs .

MOBILE RADIO CHANNELS

27

Rice Envelope. For convenience, the envelope is normalized by the Gaussian standard deviation:

The envelope pdf is

or

As kRi approaches infinity, the Rice pdf becomes a delta-like function, being a Gaussian with a variance approaching zero. The Rice distribution is sometimes called Nakagami–Rice, in recognition of its independent development by Rice (10) and by Nakagami (11), who reported it in English at a later time. Because of its physical justification for many situations, the Rice distribution is the preferred one for short-term fading. Review material covering aspects of Rice’s work is in Refs. 12,13. The distribution for the random FM and group delay for the Rice channel is given by the Student t distribution (14). The Rice envelope cumulative probability function (cpf) is expressed as

where Q is the Marcum Q function (15). Further worthwhile discussion on the Q function is given in Refs. 16,17. The Rice envelope cpf is sketched in Fig. 14. for values of the Rice k factor, including the Rayleigh case. Lognormal Shadow Fading. Shadow fading has been found experimentally to be well described by the lognormal distribution. Whereas the Gaussian distribution results from the addition of many random variables, the lognormal distribution results from the product of many positive random variables. It follows that when Gaussian variables are expressed in logarithmic units, they then follow a lognormal distribution. The transformation of variables between the distributions is z = ex , or ln z = x. (z here is a variable, not distance.) If x is Gaussian, then z is lognormal. Alternatively stated, if z is lognormal, then ln z is Gaussian. The pdf of the lognormal distribution is found from the Gaussian pdf, viz.,

28

MOBILE RADIO CHANNELS

Fig. 14. The Rice envelope cpf. For zero specular component, the distribution is Rayleigh, and approaches Gaussian (vertical line at 0 dB) for an asymptotically large specular component.

where mlz and σlz 2 are the mean and variance respectively of ln z. The lognormal signal representing the local mean of the envelope looks like one of the phase components of Fig. 14, except that the scale would be in decibels rather than linear. Typically σlz is 3 dB to 8 dB in urban environments. Suzuki: Lognormal and Rayleigh. Combining the short-term Rayleigh and long-term lognormal distributions provides a model for the stochastic component of the path loss of a narrowband signal in mobile communications. The lognormal distribution is over the mean of the envelope. This can be interpreted as Gaussian for the envelope mean in decibels. The Rayleigh envelope mean is linearly related to the Gaussian standard deviation, viz., r = , so the lognormal distribution can be applied to the σ (18). The distribution can be written

No closed form has been found for the integral, which is a practical inconvenience when applying the Suzuki distribution. However, the distribution has the advantage of being based on a physical model for the envelope, and thus offers good agreement with experimental results on large scale records of envelopes of narrowband signals. Many other distributions have been used to fit mobile channel fading (19). Some have various advantages for mathematical manipulations or for the fitting of experimental data. Two are noteworthy because of their versatility. The Nakagami m (11) distribution has a single parameter that allows the shape of the distribution to be altered, in particular for small values of r. The generalized gamma distribution (20) has effectively two parameters that can independently adjust the shape of the small and large values of r.

MOBILE RADIO CHANNELS

29

Path Loss and the Mobile Channel Much of the above discussion has been a statistical description of the behavior of the mobile channel. The interest in the envelope or power of the mobile transfer function is because this dominates the SNR of the received signal. The power is also referred to as the channel gain. How this ties in with the path loss is addressed in this section. In so doing, the discussion returns to the electromagnetic propagation and antenna issues of the opening sections. The path loss is a well-defined concept originating from point-to-point radio links. It comes from the Friis transmission equation, which relates the transmitted and received powers (PT , PR respectively), the antenna gains (GT , GR respectively), and the path loss L:

The path loss is seen from this equation to be the reciprocal of the path gain. For frequency-independent antenna gains, the free space path loss for a separation distance d and wavelength λ = c/f is

so that it varies as the frequency squared and the distance squared. The incident field strength is not dependent on frequency. In Eq. (78), the antennas are considered impedance- and polarization-matched. Mean Path Loss and Mean Antenna Gain. In a mobile channel, the classical point-to-point situation does not apply. The received power and the receiving antenna gain become statistical quantities. The antenna’s mean gain can be defined by the average gain into a well-defined distributed direction. The mean received power can be defined from a time average. The path loss is the time-varying quantity (because of the spatially dependent phase mixture of multipath propagation signals), and so the mean received power with Eq. (77) defines a mean path loss. Sometimes the term mean effective gain is used when comparing antennas by measuring their time-averaged received powers in the same environment. In this context, it must be assumed that the transmitting power and the mean path loss are both common to each measurement record used for the averaging. The mean effective gains are then proportional to the mean received powers and include polarization mismatches. What is being measured is how well, on average, the vector antenna pattern is directed towards the vector distribution of incoming power from the measurement environments. Scenario Models. Model distributions are used to approximate the average incident power directions for various applications. For a mobile vehicle, for example, the Clarke scenario (21,22), given by

is often used. This corresponds to a uniform source distribution at the horizon, surrounding the antea. Transforming to the spatial Doppler variable results in the pdf

30

MOBILE RADIO CHANNELS

and this spatial Doppler spectrum is for the incident fields or the electromagnetic propagation channel (for one polarization), and also for the mobile channel if an omnidirectional (in the θ = π/2 plane) antenna is used. The spatial Doppler spread is σu (C) = kC / (rad/m). The spatial correlation coefficient for the envelope is ρr (C) (z) ≈ J 0 2 (kC z), giving a 0.7-correlation distance of about 0.13 wavelengths and an average distance between fades of about 0.5 wavelengths. For a directional antenna, the spatial Doppler distribution corresponding to the pattern must be multiplied with Eq. (80) to get the spatial Doppler spectrum of the mobile chanl. This is how the antenna pattern can control the mobile channel behavior. A single-lobed, directional pattern acts as a spatial Doppler bandpass filter and results in a decreased (relative to an omnidirectional pattern) Doppler spread, and therefore a decreased spatial fading rate. This effect can be seen with laser speckle, where the dark areas are the deep fades of energy, and the interspeckle distance, even though the frequency is optical, is sufficiently large to be visible to the eye because the spatial Doppler spread of the illuminating beam is so small.

BIBLIOGRAPHY 1. R. G. Vaughan J. Bach Andersen, Principles of Propagation and Antennas in Mobile Communications, London: Peregrinus, 2000. 2. T. Miki M. Hata, Performance of 16 kbits/s GMSK transmission with postdetection selection diversity in land mobile radio, IEEE Trans. Veh. Technol., VT-33 (3): 128–133, 1984. 3. K. Sakoh, et al. Advanced radio paging service supported by ISDN, Proc. Nordic Seminar on Digital Land Mobile Radiocommunication, Espoo, Finland, February 1985, pp. 239–248. 4. P. A. Bello, Characterization of randomly time-variant linear channels, IEEE Trans. Circuits Syst., CS-11: 360–393, December 1963. 5. A. Papoulis, Signal Analysis, New York: McGraw-Hill, 1977. 6. P. A. Bello B. D. Nelin, The effect of frequency selective fading on the binary error probabilities of incoherent and differentially coherent matched filter receivers, IEEE Trans. Circuits Syst., CS-21, 170–186, June 1963. 7. M. J. Gans, A power spectral theory of propagation in the mobile-radio environment, IEEE Trans. Veh. Technol., VT-21 (1): 27–38, February 1972. 8. D. C. Cox R. P. Leck, Correlation bandwidth and delay spread multipath propagation statistics for 910 MHz urban mobile radio channels, IEEE Trans. Commun., Com-23 (11): 1271–1280, 1975. 9. B. H. Fleury, An uncertainty relation for WSS processes and its application to WSSUS systems, IEEE Trans. Commun., Com-44 (12): 1632–1635, December 1996. 10. S. O. Rice, Mathematical analysis of random noise, Bell Syst. Tech. J., 1944, No. 3; 1945, No. 1. 11. M. Nakagami, The m-distribution—a general formula of intensity distribution of rapid fading, in W. C. Hoffman (ed.), Statistical Methods in Radio Wave Propagation, Oxford: Permagon, 1960. 12. W. B. Davenport W. L. Root, An Introduction to the Theory of Random Signals and Noise, New York: McGraw-Hill, 1958; reprinted, Piscataway, NJ: IEEE Press, 1987. 13. D. Middleton, An Introduction to Statistical Communications Theory, New York: McGraw-Hill, 1960; reprinted, Piscataway, NJ, IEEE Press, 1997. 14. J. Bach Andersen, S. L. Lauritzen, C. Thommesen, Distributions of phase derivatives in mobile communications, IEE Proc., 137 (4): 197–201, 1990. 15. J. I. Marcum, A statistical theory of target detection by pulsed radar, IRE Trans., IT-6: 59–267, April 1960. 16. S. Stein, M. Schwartz, W. R. Bennett, S. Stein, Communications Systems and Techniques, New York: McGraw-Hill, 1966, Part III; reprinted, Piscataway, NJ: IEEE Press, 1996. 17. J. G. Proakis, Digital Communications, New York: McGraw-Hill, 1983. 18. H. Suzuki, A statistical model for urban radio propagation, IEEE Trans. Commun., Com-25 (7): 673–680, July 1977. 19. J. Griffiths J. McGeehan, Interrelationship between some statistical distributions used in radio-wave propagation, IEE Proc., 129, Part F, No. 6, 411–417, December 1982.

MOBILE RADIO CHANNELS

31

20. A. J. Coulson, A. G. Williamson, R. G. Vaughan, Improved fading distribution for mobile radio, IEE Proc. Commun., 46: 494–502, 1998. 21. R. H. Clarke, A Statistical theory of mobile radio reception, Bell Syst. Tech. J., 47: 957–1000, 1968. 22. W. C. Jakes (ed.), Mobile Microwave Communications, New York: AT&T, 1974; reprinted, Piscataway, NJ: IEEE Press, 1989.

READING LIST W. C. Jakes (ed), Mobile Microwave Communications, New York: AT&T, 1974; reprinted Piscataway, NJ: IEEE Press, 1989. W. C. Y. Lee, Mobile Communications Engineering, New York: McGraw-Hill, 1982. R. C. V. Macario, Personal and Mobile Radio Systems, IEE Telecommunications Series, 25, London: Peregrinus, 1991. J. D. Parsons, The Mobile Radio Propagation Channel, London: Pentech Press, 1992. T. S. Rappaport, Wireless Communications, Principles and Practice, New York: IEEE Press, 1996. S. O. Rice, Statistical properties of a sine wave plus random noise, Bell Syst. Tech., 27: 109–157, 1948. R. Steele, Mobile Radio Communications, London: Pentech Press, 1992. ¨ G. Stuber, Principles of Mobile Communications, Boston: Kluwer, 1996. R. C. Vaughan J. Bach Andersen, Principles of Propagation and Antennas in Mobile Communications, London: Peregrinus, 2000.

RODNEY G. VAUGHAN Industrial Research Limited, New Zealand

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELECTRICAL...NICS%20ENGINEERING/62.%20Vehicular%20Technology/W7708.htm

}{{}}

●

HOME ●

ABOUT US ●

CONTACT US ●

HELP

Home / Engineering / Electrical and Electronics Engineering

Wiley Encyclopedia of Electrical and Electronics Engineering Mobile Satellite Communication Standard Article John Lodge1 1Communications Research Centre, Ottawa, Canada Copyright © 1999 by John Wiley & Sons, Inc. All rights reserved. DOI: 10.1002/047134608X.W7708 Article Online Posting Date: December 27, 1999 Abstract | Full Text: HTML PDF (445K)

●

●

● ●

Recommend to Your Librarian Save title to My Profile Email this page Print this page

Browse this title ●

Search this title Enter words or phrases

Abstract The sections in this article are Mobile Satellite Links The Signal-Processing Path Present and Planned Systems Trends in Mobile Satellite Systems About Wiley InterScience | About Wiley | Privacy | Terms & Conditions Copyright © 1999-2008John Wiley & Sons, Inc. All Rights Reserved.

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELE...ENGINEERING/62.%20Vehicular%20Technology/W7708.htm15.06.2008 12:44:56

❍

❍ ❍

Advanced Product Search Search All Content Acronym Finder

366

MOBILE SATELLITE COMMUNICATION

MOBILE SATELLITE COMMUNICATION Mobile satellite systems provide communications services to mobile and portable terminals using a radio transmission path between the terminal and the satellite. An example of such a system, illustrating its typical components, is shown in Fig. 1. The mobile terminal may be installed in any one of a number of platforms including cars, trucks, rail cars, aircraft, and ships. Alternatively, it could be a portable terminal with a size ranging from that of a hand-held unit up to that of a briefcase, depending upon the system and the provided service. Yet a third class could be small but fixed remote terminals serving functions such as seismic data collection and pipeline monitoring and control. A mobile satellite system requires one or more satellites with connectivity to the terrestrial infrastructure (e.g., the public switched telephone network and to the various digital networks) being supplied by one or more earth stations. Typically, most of the communications traffic is between the mobile terminal and another terminal or application outside of the mobile satellite system. However, most mobile satellite systems allow for mobile-tomobile communications within the system. The earth stations J. Webster (ed.), Wiley Encyclopedia of Electrical and Electronics Engineering. Copyright # 1999 John Wiley & Sons, Inc.

MOBILE SATELLITE COMMUNICATION

367

Figure 1. The major components of a mobile satellite system. Lines terminated with arrowheads indicate communications links.

are coordinated by a control center in a way that shares the satellite transmission resources efficiently. Also, the control center may issue commands to the satellites via the earth stations. A number of radio links are required for such a system. Communication from the earth station to the mobile terminal is said to be in the forward direction, whereas communication from the mobile terminal to the earth station is said to be in the return direction. In both the forward and return directions, an up-link to the satellite and a down-link from the satellite are required, for a total of four radio links. The links between the earth station and the satellite are sometimes referred to as feeder links, whereas the links between the mobile terminal and the satellite are typically referred to as service links or mobile links. In some of the more advanced satellite systems with multiple satellites, there are radio links between adjacent satellites called intersatellite links. A wide variety of services and applications are being supported by mobile satellite systems, with many more being proposed. First- and second-generation systems are limited to data rates ranging from a few hundred bits per second (bps) to several tens of kilobits per second (kbps) and have concentrated their efforts on providing services that fall within categories such as telephone-quality speech, packet data communications, facsimile, generic asynchronous stream data, and paging. Third-generation systems are expected to be capable of transmission at rates up to several hundred kilobits per second and will be capable of delivering moderate-quality video and high-quality audio services. Increasingly, the services delivered by these systems will appear to be an exten-

sion of those available to users over the converging terrestrial systems. Many satellites isolate selected frequency bands from the composite up-link signal using filtering, translate these selected bands to their down-link frequency band, amplify them, and then transmit them toward the earth in the appropriate antenna beam. The term transparent satellite is used in this case. As an extension of this concept, some of the newer satellites use digital processing to select the up-link signal in a given frequency band, time slot, and antenna beam, and then ‘‘switch’’ it to the desired down-link frequency band, time slot, and antenna beam. The most sophisticated satellites demodulate the up-link transmissions and then process the resulting data signals in the same manner as a digital switch prior to modulation for down-link transmission. This type of satellite is sometimes referred to as a regenerative satellite. A wide variety of mobile satellite terminals is commercially available. Here, we give only a few examples. Figure 2 shows a receive-only unit, manufactured by Skywave Mobile Communications Inc., that can be used to receive alphanumeric messages sent to a personal computer-based terminal, over the Inmarsat-D system. This system is a high-penetration system and can receive messages even when moderate blockage of the satellite signal is occurring. The receiver is the small black rectangular object beside the laptop computer. The white disk-shaped object is the antenna, which has a magnetic base allowing it to be temporarily mounted on the roof of a vehicle. At other times, any flat surface will suffice. A Mitsubishi MSAT telephone transceiver, mounted on the front wall of the trunk of a car, is shown in Fig. 3. The corre-

368

MOBILE SATELLITE COMMUNICATION

Figure 2. A receiver and antenna for the Inmarsat-D high-penetration messaging system. The receiver is shown connected to a laptop computer. Reprinted with permission from Skywave Mobile Communications, Inc.

sponding antenna subsystem, mounted on the car’s roof, is shown in Fig. 4. A third subsystem, which is not shown, is the users interface unit in the passenger compartment, including the telephone handset. The major subsystems of the CAL Corporation’s satellite telephone terminal, for telephone communications to aircraft

Figure 3. A Mitsubishi MSAT telephone transceiver mounted on the front wall of the trunk of a car.

via MSAT, are shown in Fig. 5. Most of the terminal’s electronics are contained in the black box on the right-hand side. This box would normally be mounted inside the pressurized cabin of the aircraft. The antenna subsystem is shown on the left-hand side, with its radome placed behind it. For this particular antenna, two short helices are used as the transducing

MOBILE SATELLITE COMMUNICATION

369

Figure 4. The antenna for a Mitsubishi MSAT telephone terminal, mounted on the roof of a car.

elements in order to achieve the required amount of antenna gain while keeping the profile of the antenna low. The antenna is often mounted on the top of the fuselage, as is shown in Fig. 6. However, on some aircraft the top of the tail fin is a preferred location for antenna mounting.

MOBILE SATELLITE LINKS We will start by considering the path of the radio signal as it travels from the satellite to the mobile terminal, that is the down-link in the forward direction. The detailed discussion of

Figure 5. The major subsystems of the satellite telephone terminal, intended for use by aircraft with the MSAT system. Reprinted with permission from CAL Corporation.

370

MOBILE SATELLITE COMMUNICATION

Figure 6. A Cessna Citation jet aircraft, operated by the Ontario Air Ambulance Service, equipped with a mobile satellite communications terminal. The antenna subsystem can be seen mounted on the top of the fuselage.

this link will introduce the concepts necessary to understand more concise discussions pertaining to the other links of interest in a mobile satellite system. Radio frequency bandwidth and electrical power are two scarce resources that tend to constrain the design of mobile satellite systems. In this section, the focus is primarily on power, with efficient bandwidth utilization being partially addressed in subsequent sections. Clearly, down-link power will be limited because most satellites use solar power as their primary source of electrical power. Also, up-link power from the mobile terminal tends to be limited because such a terminal receives electrical power from either its own battery or that of the vehicle. Line-of-Sight Transmission At the satellite, the signal is amplified so that its average signal power is Pt dBW, at the input to the transmitting antenna. It is the transmitting antenna’s function to spread that signal power as uniformly as possible over the desired coverage area on the Earth’s surface, while wasting as little power as possible outside this coverage area. This is directly analogous to the ability of the reflecting surface of a flashlight to focus the light from the bulb into a beam of light. A measure of the ability of the antenna to focus the radiation is its gain, which is the ratio of the flux density at the center of the coverage area to that value that would occur if the power had been radiated equally in all directions (i.e., isotropic radiation). This gain is a function of the size of the antenna, and for a circular parabolic antenna it is given by G = 10 log10 (π 2 D2 /λ2 ) dBi

(1)

where ⍀ is the efficiency of the antenna (typically between 50% and 70%), D is the diameter in meters, and ␭ is the wave-

length of the radio frequency signal in meters. Other types of antennas will have differing gains, but Eq. (1) provides an order of magnitude estimate of the required antenna size to achieve a prescribed gain. This discussion assumes that a single beam is used to cover the desired area. For reasons that will be discussed later, it may be advantageous to cover the desired area with multiple overlapping beams, but using the same antenna superstructure. An example of one way to achieve this is to use a single large reflector with multiple feeds (i.e., source transducers) in different locations near the focal point of the reflector. Of course, increasing the number of beams increases the complexity of the satellite. The size and weight of the satellite’s antennas is constrained by the need to maintain reasonable costs for the satellite and its launch. Nevertheless, advanced technology allows for surprisingly large antennas to be deployed in space. For example, the North American MSAT satellites have two elliptical antennas, measuring 6 m by 5 m, and provide five beams covering all continental North America, the Caribbean Sea, and Hawaii. Some later systems have significantly larger antennas and can support more than 100 beams. As the signal travels from the satellite to the earth, its flux density decreases as the square of the distance traveled. This power loss is referred to as the free space path loss and is given by Lp = 10 log10 [(4πd)2 /λ2 ] dB

(2)

where d is the distance traveled between the satellite and the mobile terminal. A geostationary orbit is a circular orbit for which the orbital radius, position, and velocity are such that the satellite remains in approximately the same location above the equator as the earth rotates. For a geostationary

MOBILE SATELLITE COMMUNICATION

orbit, like that of MSAT, the radius is about 42,163 km resulting in a typical propagation delay of greater than an eighth of a second to traverse from the satellite to the surface of the earth. At MSAT frequencies, the corresponding path loss is about 188 dB! The great altitude of a geostationary satellite allows it to view about a third of the surface of the earth. Consequently, global coverage (with the exception of the polar regions) is possible with only three satellites. A larger number of satellites, in circular orbits at lower altitudes, can be used to provide global service with the advantages of lower path loss, shorter propagation delay, and cheaper launch costs on a per satellite basis. For reasons of satellite longevity, altitudes that avoid the Van Allan radiation belts are usually selected. The low earth orbits (LEO) are located beneath the primary belt and have altitudes between 500 km and 2000 km. Similarly, the medium Earth orbits (MEO) are located between the primary and secondary belts and have altitudes between 9000 km and 14,000 km. The medium earth orbits are sometimes referred to as intermediate circular orbits (ICO). Unlike systems that use geostationary orbits, these other systems typically use several distinct orbital planes, each of which is inclined with respect to the equator. A number of proposed systems have planned to use highly elliptical orbits (HEO) instead of circular ones. The potential advantage of a HEO-based system is that it can provide high angle-of-elevation coverage to selected areas in the temperate zones (i.e., those parts of the world for which the demand for communications services is the greatest) with a moderate number of satellites. Despite this advantage, it does not appear that HEO systems will play a significant role in mobile satellite communications. Upon reaching the terminal, the signal energy is collected by the receiving antenna and is converted by a transducer to an electrical signal. A typical example of a mobile satellite antenna designed for the MSAT system is shown in Fig. 7. Here, the transducing element is a short helical structure, similar to the element used for the land mobile satellite terminal and to each of the two elements for the aircraft mobile satellite terminal shown previously in this article. The white dome is a radome that is placed over the antenna to protect it. This antenna must be steered in azimuth but has a wide enough beam that steering in elevation is not necessary. Many mobile terminals use closed-loop antenna steering mechanisms, based upon the received signal strength. The gray box shown in Fig. 7 contains a self-calibrating electronic compass that can be used to improve the antenna steering achievable using signal strength alone. Because of the fact that the physical rules describing the propagation of transmitting and receiving display a reciprocal relationship, an appropriate measure of the antenna’s ability to collect the energy is the antenna gain, as described in the text near Eq. (1). Therefore, the average power of the received signal from the line-of-sight propagation path at the output of the receiving antenna is given by Pr = 10 log10 C = Pt + Gt − Lp + Gr dBW

(3)

where Gt is the gain of the transmitting antenna and Gr is the gain of the receiving antenna. The signal’s radio frequency plays a major role in Eq. (3), with Gt, Lp, and Gr increasing with the square of the frequency. The net result is that the received power also in-

371

Figure 7. A prototype antenna system designed for the North American MSAT system. On the right-hand side of the foreground is the short helical antenna element. It can be steered in azimuth but is fixed in elevation. The box on the left-hand side is the antenna steering unit, and the radome is shown in the background.

creases with the square of the frequency. Alternatively, if the received power is treated as the fixed parameter, smaller antennas could be used at higher frequencies. Some of this benefit for higher frequencies is offset by other propagation effects. For example, the lower frequencies (i.e., longer wavelengths) are more robust in the presence of blockage by collections of small obstacles such as foliage and rain. A second factor that is very important is the availability of an otherwise unused radio spectrum. At the international level, spectrum usage is determined by the International Telecommunications Union (ITU) at an on-going series of World Administrative Radio Conferences (WARC). Then national bodies, such as the Federal Communications Commission (FCC) in the United States, license specific service providers to offer the corresponding services within each country. A wide variety of frequency bands have been allocated for mobile satellite systems, typically with the larger allocations being at the higher carrier frequencies as a result of availability. Consequently, the systems that offer low data rate services are generally allocated lower frequency bands than those offering high data rate services. For example, a number of systems offering lowrate store-and-forward messaging services communicate between the satellite and the mobile terminal in frequency bands between 100 MHz and 400 MHz, although most of the systems offering medium-rate mobile satellite telephone ser-

372

MOBILE SATELLITE COMMUNICATION

vices use bands between 1.5 GHz and 2.5 GHz, and many of the proposed systems for providing high-rate multimedia services plan to operate in bands between 20 GHz and 30 GHz. Because of the large path loss that is typical of satellite transmissions, the received power is very low. In fact, it is so low that the thermal noise in the receiving antenna and front end of the receiver must be accounted for. The resulting carrier-to-noise-spectral-density ratio is given by 10 log10 (C/N0 ) = Pr − Tr − k dB-Hz

(4)

where Tr is the composite noise temperature of the receiver expressed (dBK) and k is Boltzmann’s constant (⫺228.6 dBW/ K-Hz). If the transmission is digital with a rate of R bps, the energy-per-bit-to-noise-spectral-density ratio is given by Eb /N0 = C/(N0 · R)

(5)

Of course, thermal noise is not the only impairment that needs to be considered. Some of the other common impairments that are encountered by mobile satellite transmissions will be addressed in the following sections. Multipath Propagation and Shadowing In addition to the line-of-sight path, the signal can reach the receiving antenna by reflected paths from objects that are usually located nearby. Often, several distinct reflecting objects are in the field of view of the receiving antenna. If the differences in the propagation times for the various propagation paths (reflected and line-of-sight) are much less than the reciprocal of the bandwidth of the transmitted signal, the effect of the multipath propagation can be viewed as non-timedispersive. This type of multipath propagation will affect the power and carrier phase of the received signal according to the nature of the superposition of the paths, but it will not distort its frequency content or introduce intersymbol interference in the case of a digital transmission. For land mobile satellite applications, measurements (1) taken in a frequency band near 1.8 GHz indicate that the difference in propagation times rarely exceeds 600 ns. Consequently, for signal bandwidths up to several hundred kilohertz, the multipath propagation can be considered non-time-dispersive. The following discussion is based on this assumption being valid. If the geometry of the paths change with time as a result of terminal motion, satellite motion, or motion of the reflecting objects, the power and carrier phase of the received signal will vary with time. This time-varying phenomenon is referred to as fading, or more specifically as flat fading for the non-timedispersive case. For the purpose of evaluating the performance of candidate transmission techniques, it is frequently desirable to model the propagation environment in a way that is suitable for numerical analysis and simulation. An approximation that is often made is to assume that the reflecting objects are adequately numerous and independent in nature for the central limit theorem to apply. Consequently the fading can be represented by a Gaussian process that is completely statistically characterized by its power spectral density. The power spectral density will be nonzero only over a bandwidth equal to the difference in frequency between the path with the greatest Doppler frequency shift and that with the least (2). This

type of fading model is referred to as Rayleigh fading. The combination of the line-of-sight path with the Rayleigh fading reflected path is referred to as Rician fading, which has the additional parameter called the carrier-to-multipath ratio (C/M), defined to be the ratio of the average signal power received over the line-of-sight path to that received over the reflected paths. Another effect that can greatly affect the availability and performance of a mobile communications link is shadowing, the term given to blockage of the line-of-sight path. Such blockage occurs naturally in terrestrial mobile satellite environments as the moving vehicle passes by obstacles such as buildings, trees, and bridges. Many obstacles result in such severe attenuation of the line-of-sight signal that it is weaker than the reflected paths and can be ignored. A useful but simple model for shadowing is to switch between a good state (unshadowed) and a bad state (shadowed) with the typical time period for enduring each state being determined by the parameters of a two-state Markov model (3,4). A transmission model corresponding to this discussion is shown in Fig. 8. A simple shadowing model is to apply a fixed attenuation selected on a shadowing event by shadowing event basis, using a lognormal distribution. If the shadowing is predominantly caused by foliage, the line-of-sight path may also be included with it being subjected to attenuation according to another lognormal distribution (5). Of course the values selected for the model’s parameters depend upon many issues, including angle of elevation to the satellite, type of terminal (e.g., land mobile, aircraft, marine, hand-held), antenna gain pattern, vehicular velocity, satellite velocity, environment (e.g. urban, suburban, highway), and terrain. Other Sources of Degradation Degradation to the received signal caused by thermal noise, multipath propagation, and shadowing have already been discussed. In many systems, these are the dominant sources of degradation, but there are a number of other ones that should be appreciated. Perhaps the next most important source of degradation is interference to the desired signal from other signals within the same system. If the interference is cause by another signal that is located in the same frequency channel as the desired signal, the interference is referred to as co-channel interference. For narrowband signals, co-channel interference is generally caused by interferers in other antenna beams for which the out-of-beam attenuation provided by the satellite antenna is not sufficiently great to render the interfering signal negligible. For spread spectrum signals, some of the co-channel interference may be due to other signals within the same beam. Interference to the desired signal can occur from signals in the adjacent frequency channels as a result of the fact that some of their transmitted energy falls outside of their allotted frequency channel. This type of interference is known as adjacent-channel interference. For some mobile satellite systems, the ratio of the carrier frequency to the bit rate is many orders of magnitude. When this is the case, a nonnegligible amount of degradation can occur because the phase of the radio frequency carrier differs significantly from its ideal value in a time-varying nature, which is the result of the electronic components in the system. This phenomenon is called phase noise. Common sources of phase noise include imperfect oscillators and frequency syn-

MOBILE SATELLITE COMMUNICATION

373

√c X +

Unshadowed +

Transmitted signal

Rayleigh fading

X

Shadowed

Received signal Thermal noise

White Gaussian noise (power spectral density = N0/2)

√m Shadowing model

2-state Markov model

thesizers, vibration of the mobile terminal’s electronic circuitry (known as microphonics), and electronic steering of the mobile terminal’s phased array antenna. Nonlinear power amplification, at several locations in a mobile satellite system, can cause degradation. In the case of the transmitting power amplifier in the mobile terminal, typically only a single carrier (signal) is present, and the distortion of that signal by the amplifier’s nonlinear behavior has two effects. First, there will be a small reduction in the power efficiency of the desired transmission. For example, if the signal is digital, a little more transmit power will be necessary to achieve the required bit-error rate. Second, the distortion will often broaden the power spectrum of the transmitted signal resulting in increased interference in the adjacent channels. Nonlinear distortion will also occur in the transmit power amplifiers of the earth stations and the satellites. Usually, there will be many carriers being amplified simultaneously. In this case, the result is a broadband noiselike signal caused by the intermodulation of the many carriers present in the amplifier. Depending on the frequency band used by the given mobile satellite system, it may be necessary to account for effects such as ionospheric scintillation, tropospheric scintillation, gaseous absorption, and rain attenuation. In general, these effects become more severe for lower angles of elevation.

Figure 8. A useful model of fading and shadowing for the evaluation of mobile satellite transmission schemes. Here, c is the average power for the line-of-sight path and m is the average power for the reflected paths.

the receiving side to recover the transmitted information. Here, we will discuss the blocks in this processing chain only to the level necessary for understanding their role in a mobile satellite context. More detailed treatment of many of these processing stages can be found elsewhere in this encyclopedia. The first block in the chain is the Information Source. Examples include telephone-quality speech, data representing text, and multimedia signals representing a composite of audio, video, and data components. Regardless of the type of information that is to be transmitted, it is important to minimize the number of bits required to represent the information subject to constraints such as delay, processing complexity, and quality of the representation. This is the objective of the second block in the chain, entitled ‘‘Source coding.’’ Using telephone-quality speech as an example, the analog waveform can be accurately represented using a 64 kbps stream of data, by sampling the waveform at 8 ksamples/s and giving each sample 8 bits of precision. However, using recently developed speech-coding standardized techniques, the bit rate can be reduced a full order of magnitude to 6.4 kbps without a significant reduction in speech quality (6). Very efficient standardized low-rate video-coding techniques also exist (7). Of course, the same techniques as are used for computer storage can be used to reduce the size of data and text files for mobile satellite transmission. Error Control Coding

THE SIGNAL-PROCESSING PATH In this section we discuss some of the signal-processing techniques that can be used to increase the efficiency with which the scarce resources of radio frequency spectrum and electrical power are used. Figure 9 shows a high-level block diagram of the processing stages for the transmitting side of the communications chain. The inverse operations are performed on

Information source

Source coding

Error control coding

Error control coding introduces redundancy into the bit stream by increasing the total number of bits in such a way that each original bit influences several bits in the error-control-coded bit stream. This redundancy can then be used to correct (forward error correction coding) or detect (error detection coding) transmission errors at the receiver. We will consider forward error correction first. Even though the additional bits do result in an increase in the required number of

Interleaving

Data modulation

Channel sharing (multiple access)

To receiver processing

Figure 9. A high-level block diagram of the processing stages for the transmitting side of the communications chain.

MOBILE SATELLITE COMMUNICATION

bits to be transmitted, appropriate coding and decoding schemes will generally result in a net reduction in the transmitted power required to meet a given bit-error rate. For firstgeneration mobile satellite systems, rate-1/2 constraintlength-7 convolutional coding has been a fairly standard choice. Note that the rate is the ratio of the number of bits into the coder to those out of the coder. In some cases, punctured versions of this code have been used to achieve a higher coding rate, thereby improving bandwidth efficiency at the expense of power efficiency. A predominant reason for the popularity of this code is that it was one of the first fairly powerful error correction codes for which decoder integrated circuits, capable of processing soft decisions, were commercially available. For decoding in fast fading and shadowing conditions, the soft decision should incorporate channel state information so that the decoder assigns relatively less importance to bits that were received when the signal was faded or blocked. The achievable coding gain is a strong function of the block length over which coding is performed, with larger blocks allowing for greater gains. For applications for which the packet or frame length is quite short (e.g., most packet data and low-rate speech applications) convolutional coding is still a good choice although constraint lengths greater than 7 can be implemented now. Tail biting (i.e., encoding the input data in a circular buffer) can be performed to eliminate the overhead of transmitting extra bits to terminate the code’s decoding trellis (8). For applications for which the frame length is longer than a couple of hundred bits, turbo coding will be a strong candidate for future systems (9). The performance of turbo coding improves as the block length increases. However, the end-toend delay of the transmission system increases with increasing block lengths. Consequently, only services that are tolerant of fairly large delays can benefit from the most power efficient error control coding. For rate-1/2 coding, Fig. 10 shows the performance for constraint-length-9 convolutional coding (80-bit block) and turbo coding (512-bit block and

10–1

Average bit error rate

Uncoded

10–2 80 bits 512 bits

10–3 10,000 bits 10–4 0.5

1

2 1.5 Eb/N0 (dB)

2.5

3

Figure 10. The performance of various rate-1/2 codes in an additive white Gaussian noise environment. Shown are simulation results for a constraint-length-9 convolutional code with tail biting and a block size of 80 bits, and turbo coding with block sizes of 512 bits and 10,000 bits.

10–0

10–1 Average bit error rate

374

Uncoded

10–2 Rate-3/4

10–3

10–4

10–5

Rate-1/2

1

2

3 4 Eb/N0 (dB)

Rate-1

5

6

Figure 11. The performance of codes of differing rates in an additive white Gaussian noise environment. Shown are simulation results for the constraint-length 9 rate-1/2 code; the rate-3/4 code, which is a punctured version of the rate-1/2 code; and the rate-1 code, which is a pragmatic trellis-coded modulation with the rate-1/2 code being mapped into a 4-level constellation.

10,000-bit block). This turbo code uses 16-state recursive systematic convolutional codes as its component codes. These performance results assume antipodal signaling (e.g., ideal coherent binary phase-shift keying) with the only channel impairment being Gaussian noise. In general, the benefit that can be achieved by error correction coding increases with increasing decoding complexity and block (i.e., code word) size, and with decreasing code rate. One way to achieve a higher code rate, for a fixed decoding complexity, is to use puncturing (10). Puncturing increases the code rate by selectively deleting some of the coded bits prior to transmission. In order to increase the code rate beyond 1 bit per symbol, it is necessary for the coder to map the input sequence of bits into a sequence of symbols for which the size of the symbol alphabet is greater than 2. A well-known technique for doing this is trellis-coded modulation (11,12). Some forms of trellis-coded modulations are designed in such a way that standard convolutional decoder integrated circuits can be used to perform the decoding. These forms are referred to as pragmatic trellis-coded modulations (13). An example of the trade-off between power and bandwidth efficiency can be seen in Fig. 11. Here, all three codes are based upon the same convolutional code, with the rate-1/2 code being a constraintlength-9 code, the rate-3/4 code being a punctured version of the rate-1/2 code, and the rate-1 code being a pragmatic trellis-coded modulation with the rate-1/2 code being mapped into a 4-level constellation. Error detection coding is useful for services that are message or frame based, and it is important to know whether a given message or frame has been received correctly. In these cases a small field of parity bits (e.g., 16 parity bits) is appended to the message, with the parity bits being generated using a cyclic redundancy code. At the receiver, if the parity bits computed from the received data bits do not agree with the received parity bits, the message is known to be in error.

MOBILE SATELLITE COMMUNICATION

Modulation After interleaving, the sequence of coded symbols is modulated. Here, we restrict our consideration to linear modulation schemes. For a linear modulation scheme, the transmitted signal is given by

 s(t) = Re =

N−1  i=0

N−1 

−

ai g(t − iT ) e

jω 0 t

g(t − iT )Re(ai ) cos(ω0t)

i=0





N−1 

(6)

g(t − iT )Im(ai ) sin(ω0t)

i=0

where ai; i ⫽ 0, . . ., N ⫺ 1 is the sequence of complex modulation symbols, T is the symbol period, g(t) is the unit pulse response of the pulse-shaping filter and is assumed to be real, and 웆0 is the radian carrier frequency. In the second line of Eq. (6), the term inside the square brackets prior to ‘‘cos’’ is referred to as the in-phase component of the signal and the term inside the square brackets prior to ‘‘sin’’ is referred to as the quadrature component of the signal. For M-ary signaling, each ai is selected from an alphabet of M complex numbers, with the modulus of each complex number representing the amplitude of the given symbol and the phase of each complex number representing the phase of the given symbol. The majority of mobile satellite communications systems uses one or more forms of phase modulation. In the case of phase modulation, each ai is selected from a symbol alphabet for which all elements have a modulus of one. Therefore, only the phase of the symbol varies. Binary phase shift keying (BPSK) is popular for low rate systems because of its robustness. For BPSK, each ai is select from the alphabet 兵1, ⫺1其 which is purely real, and consequently a BPSK waveform has no quadrature component. A variation of BPSK, that is used in aeronautical satellite communications, is 앟/2-BPSK for which subsequent symbols experience a relative phase shift of 앟/2 radians. For example, each ai is select from the alphabet 兵1, ⫺1其 when i is even and from 兵j, ⫺j其 when i is odd. When used with an appropriate choice of pulse-shaping filter, such as a 40% square-

root raised-cosine filter, the result is a waveform that suffers less spectral spreading when passed through a nonlinear amplifier, but enjoys all the robustness of standard BPSK. For systems requiring some additional spectral efficiency, some form of quadrature phase shift keying (QPSK) is usually selected. Standard QPSK can be thought of as two BPSK signals being transmitted in parallel; one as the in-phase component and the other as the quadrature component. A variation of QPSK that is of some interest is 앟/4-QPSK, for which subsequent symbols experience a relative phase shift of 앟/4 radians. The advantages of selecting 앟/4-QPSK are similar to those described previously for 앟/2-BPSK. Another variation of QPSK that is even more robust to nonlinear amplification is offset-QPSK for which the symbol timing for the in-phase component is offset by half a symbol period relative to that of the quadrature component. Multiple Access Next we consider how the satellite resources of bandwidth and power can be efficiently shared between many users. The sharing of the transmission medium between several users is referred to as multiple access (see MULTIPLE ACCESS MOBILE COMMUNICATIONS). We start from a highly idealized point of view, considering the case where there is only a single beam, perfect synchronization in both time and frequency have been achieved, and no interference is permitted between users. First, let power be the only constraint. Each user can have as much bandwidth as he wishes but cannot exceed some fixed maximum value of transmit power. Under this constraint, each user attempts to maximize his throughput (i.e., bit rate) subject to the requirement that the average bit error rate is better than some specified value. In general, lowering the coding rate allows for greater power efficiency and consequently a higher throughput for a given amount of power. The achievable region is illustrated by the area under the curve labeled ‘‘Power Constraint’’ in Fig. 12. Note that the coding rate is expressed in bits per dimension, which takes into account the modulation and error control coding. This is the ratio of the number of bits into the error correction coder to the number of dimensions out of the modulator, over a fixed period of time. In Eq. (6), Re兵ai其 and Im兵ai其 can be considered as examples of dimensions in the signal space. It is well

Capacity (composite bite rate)

In some systems, a request will then be sent to the transmitter to retransmit the message. Returning to error correction coding, many forward error correction codes are much better suited to correcting randomly distributed single errors than long bursts of errors, assuming that the average bit error rate is fixed. However, some impairments such as multipath fading cause error patterns that are bursty in nature. To the extent allowed by constraints such as message length and delay restrictions for the service, interleaving can be used between the coder and the modulator in an attempt to eliminate error bursts prior to decoding. Interleaving permutes the order of the coded symbols according to a rule that is known at both the transmitter and the receiver. After demodulation at the receiver, the deinterleaver performs the inverse permutation prior to passing the soft decisions to the decoder. By so doing, sequences of soft decisions corresponding to poor bursts of signal are broken up and mixed with soft decisions that were received under more favorable conditions.

375

Bandwidth constraint Power constraint

Bandwidth Power limited limited region region Ropt Coding rate (bits per dimension)

Figure 12. The tradeoff-between capacity and coding rate subject to a power constraint and a bandwidth constraint. Ropt is the coding rate that maximizes the capacity.

376

MOBILE SATELLITE COMMUNICATION

known that for a bandwidth of B and a time duration of Ts the number of available dimensions is 2BTs (14). Now let bandwidth be the only constraint being considered. Clearly, the composite bit rate will increase linearly as the users increase their coding rate. The achievable region is illustrated by the area under the curve labeled ‘‘Bandwidth constraint’’ in Fig. 12. If both the power and bandwidth constraints are taken into account, there is an optimal code rate (assuming block size and decoding complexity are fixed) Ropt that maximizes the throughput of the system. If the system is operating at a lower rate, it is said to be bandwidth limited, and if it is operating at a higher rate, it is said to be power limited. With most of the early mobile satellite systems, the satellites were comparably weak, the demand for spectrum was low, and few devices were available to support coding rates below rate-1/2. Consequently, most early systems were operating in the power-limited region. With newer systems, much more emphasis is being placed on achieving nearly optimum capacity in the system design. One example of a set of dimensions (i.e., a basis) for the signal space is the time sample representation of the composite signal, with sampling being performed at the Nyquist rate. If sequential groups of these time samples are apportioned between the users, the sharing arrangement is called time division multiple access (TDMA). Here, the mobile terminals must be fairly accurately synchronized in time so that bursts arriving at the satellite from different terminals can be tightly packed without interfering with each other. Typically, the required timing accuracy is achieved when the terminal requests to initiate communication by sending a short burst on a random access channel, for which accurate timing is not necessary. Then along with an assignment of a set of time slots, the system sends the terminal an accurate clock correction that was calculated by the Earth station based upon the measured time-of-arrival of the burst. Of course many other potentially useful bases exist. If nonoverlapping portions of the total bandwidth are apportioned between the users the arrangement is called frequency division multiple access (FDMA). In this case, timing accuracy is no longer important but narrower band filtering is necessary and the lower data rates present on each carrier tend to make the system more susceptible to phase noise. If orthogonal codes are used to form the basis of the signal space, the sharing is called code division multiple access (CDMA). In a synchronous CDMA system, the carriers must be synchronized in time to within a small fraction of a chip period so that orthogonality is maintained. In the forward direction, this is fairly straightforward to achieve if all the signals are originating from a single Earth station. In an asynchronous CDMA system, time synchronization is not required with the result that the signals are no longer truly orthogonal, resulting in some interference. In the return direction, achieving sufficiently accurate time synchronization amongst all of the mobile terminals is quite challenging so asynchronous CDMA could be preferred over synchronous CDMA. Of course, combinations of these approaches are possible. Most of the mobile satellite systems to date have used FDMA. However, systems based upon narrowband TDMA, which is a combination of FDMA and TDMA, are beginning to appear, even though CDMA is a strong candidate for systems with many beams or where there are severe power spectral density limitations.

3 1 2

1 2

3 1

3 1

2 3

2 3

1

Figure 13. Total coverage area being covered by multiple beams. In this case, there are 14 beams and the total frequency band is subdivided into three subbands. Two of the subbands are used in five beams, whereas the remaining subband is used in four. The resulting frequency reuse factor is 4.667.

More efficient use of both bandwidth and power can be achieved if the satellite’s antenna system covers the desired area of the Earth’s surface with several smaller beams instead of one large one. The power efficiency of the link is improved as a result of the higher antenna gain associated with the smaller beams. With respect to frequency, the total allocated system bandwidth is divided into a number of distinct subbands, which need not be of equal bandwidth. As illustrated in Fig. 13, each beam is assigned a subband in such a manner that some desired minimum distance between beams with the same subband is maintained. The frequency reuse factor is the ratio of the number of beams to the number of distinct frequency subbands. For most of the preceding discussion, it was assumed that no interference between users is permitted. In reality, some interference is unavoidable and may even be desirable to decrease system complexity and possibly to improve system capacity. For example, some CDMA systems allow each transmission to be completely asynchronous in chip timing and carrier phase relative to that of other users occupying the same frequency band and period in time. In this case, each transmission appears to be low-level broad band noise to the other users. Unlike the FDMA and TDMA systems for which the interference tends to be dominated by a small number of dominant interferers, the interference experienced by a user is the result of a very large number of other users resulting in a level of interference that is much less variable. Full statistical advantage can be take of voice activation without the need for sophisticated dynamic channel assignment strategies. Powerful error correction coding allows for high levels of both intra- and interbeam interference. This results in the ability to reuse the same frequency bands in every beam and a corresponding high level of capacity in a multibeam satellite system (15). Because interference is unavoidable, interference mitigation techniques are of interest. One example of such a technique is power control, for which the power of each user terminal is dynamically adjusted with the goal of providing it with just enough power to meet the required grade of service. Allowing terminals additional power would only serve to exacerbate the interference levels experienced in the system. A

MOBILE SATELLITE COMMUNICATION

second example is the use of multiuser detection schemes (16).

PRESENT AND PLANNED SYSTEMS Here the intent is to provide some examples of systems that are presently offering mobile satellite communications services and of those that are planned for the future. The systems discussed represent only a sampling and not an exhaustive summary. Global mobile satellite communications got its start in 1976 when 3 Marisat satellites where launched and positioned at approximately equal intervals in geostationary orbits. In 1979, Inmarsat was formed to offer global maritime satellite communications services. Inmarsat is a multinational organization that was created by the United Nations affiliated International Maritime Organization. Even though its original charter restricted its operation to maritime services, its charter was later extended to include aeronautical as well as land mobile and portable services. The nature of the Inmarsat organization continued to evolve with the goal of allowing it to offer an increasing array of mobile satellite services in a commercially competitive environment. Inmarsat-A was the first system to offer commercial service on a global basis. Its terminals are relatively large and expensive, with the typical antenna being a 1-m diameter parabolic dish and a terminal weight of around 35 kg being representative. Consequently, the majority of the customers are large commercial users with most of the marine terminals installed on ocean-going ships and most of the portable terminals belonging to governments or news gathering organizations. Voice transmission was accomplished using analog frequency modulation, which is neither bandwidth nor power efficient by today’s standards. Inmarsat has introduced several new voice and data systems that are based on more recent digital technologies. All Inmarsat’s systems operate over geostationary satellites. The first of these new systems is the Inmarsat aeronautical system, which is based upon the work of the International Civil Aviation Organization and the Airlines Electronic Engineering Committee. The purpose of the aeronautical system is to provide comprehensive aeronautical communications services, including basic air traffic services, aeronautical operational control, and cabin telephone. Inmarsat began by providing the cabin telephone service, with other services to be phased in later. This system is unique in that it is the only mobile satellite system that has been designed in a manner consistent with Open System Interconnect (OSI) principles. The Inmarsat-M and -B systems were developed in parallel and share a common protocol. The M system offers lower-cost and reduced weight (typically about 10 kg) terminals, which provide communications-quality voice (4.2 kbps voice coding rate with the addition of error control coding bringing the rate up to 6.4 kbps), low-speed data (2.4 kbps), and facsimile services. In addition to marine and land mobile terminals, portable terminals the size of a small briefcase (including the antenna) are available. Telephone booths based on Inmarsat-M technology, that are powered using solar panels, are used in underdeveloped parts of the world. Inmarsat-B is the designated successor to Inmarsat-A for providing high-quality professional communications services.

377

For operation within the global beam of a satellite, the mobile antenna requirements for the A and B systems are identical, with a typical gain of 20 dBi. Inmarsat-M terminals have smaller antennas, with gains of 14 and 12 dBi for marine and land mobile terminals, respectively. Also available are still smaller ‘‘mini-M’’ terminals that operate only in the higher gain beams provided by the Inmarsat-3 series of satellites, launched in 1996 and 1997. Inmarsat-C was introduced in 1990 to support store-andforward packet data services such as telex, electronic mail, messaging, and position reporting. Even though only low-bit rates (600 bps) are supported, the terminals are small and inexpensive relative to those for the other Inmarsat sytems. An antenna with a gain as low as 1 dBi will suffice. A number of regional systems offer terminals and services similar to those of the mini-M system. One example is the North American MSAT system, for which Canada and the United States each launched a geostationary satellite. A number of future regional systems are planned for Asia and the Middle East, using extremely large geostationary satellites, which should be capable of delivering these services to handheld terminals, or higher data rate services to larger terminals. These systems are alike in that they all use geostationary satellites, and the mobile terminals receive their signals in a band around 1,550 MHz and transmit their signals in a band around 1,650 MHz. Systems exist that use completely different frequency bands and in some cases orbits. We will begin with brief discussions of two systems that offer two-way messaging and position determination. These systems have targeted truck fleet management and cargo position reporting as primary application areas. In 1990, the OmniTRACS system began full operation, providing two-way communications and position reporting services. It was licensed to operate on a secondary basis, which implies that it must not interfere with primary users, in the 12/14 GHz bands using existing geostationary satellites. The early start of service has allowed the OnmiTRACS system to build up a large customer base. A number of novel spread spectrum techniques are employed to safeguard against interfering with other systems. The Orbcomm system plans to operate with a full constellation of 36 LEO satellites. The mobile terminals will receive their signals at about 138 MHz and transmit their signals at about 150 MHz. The system operators hope to achieve a competitive cost advantage by having small inexpensive satellites, low launch costs (as a result of the small satellites and low orbits), and lower terminal costs caused by the lower-frequency electronics. A number of planned systems expect to offer hand-held telephone services on a global basis. Three systems that deserve particularly close attention are Globalstar, Iridium, and ICO. Globalstar and Iridium are LEO systems with 48 and 66 active satellites in a full constellation, respectively. The ICO system will use 10 active MEO satellites. The multiple access technique selected for ICO and Iridium is narrowband TDMA, whereas Globalstar will use CDMA. Iridium and Globalstar should be offering global services before the turn of the century, whereas ICO is expected to be a couple of years later. Early in the next century, a number of satellite systems are planned to offer a broad range of services, including higher rate services which should effectively extend the digi-

378

MOBILE SATELLITE COMMUNICATION

tal network capabilities that will be available terrestrially. The highest profile of these is the Teledesic system. Originally, this system planned to use 840 LEO satellites! This has now been scaled back to a planned initial constellation of 288 LEO satellites. For a number of reasons, position determination can be very important for a mobile satellite communications user. In fact, position determination is an integral part of many of the services such as vehicle fleet management and cargo tracking. Some terminals may use position information for antenna steering and to aid in the satellite and antenna beam handoff algorithms. Also, accurate position information is required for obtaining a license to offer service in some countries because the national authority insists on knowing if a call is being made within its territory. Some mobile satellite communications systems are capable of providing fairly coarse position estimation using the signals and satellites within the systems itself. However, accurate position determination is usually done by taking advantage of the Navstar Global Positioning System (GPS) (17). The GPS system employs 24 satellites distributed in 6 orbital planes, each inclined by 55⬚ with respect to the equator. These satellites are in 12 h medium earth orbits. Even though the system is financed by the US Department of Defense, it is used globally for both civilian and military applications. In addition to the signals generated aboard the Navstar satellites, the Inmarsat-3 satellites have transponders that can relay ground-generated GPS-type signals. These additional signals can be used to improve the accuracy and reliability of the position estimates. A GPS receiver estimates the range to several satellites and then uses these estimates to determine its position by triangulation. Range estimates to three satellites are sufficient to provide two-dimensional position (i.e., on the surface of the earth or if the altitude is known) plus accurate time, whereas four satellites are required to provide three-dimensional position plus accurate time. Each Navstar satellite transmits in two frequency bands; the L1 carrier is centered at 1,575.42 MHz and the L2 carrier is centered at 1,227.60 MHz. Frequency-dependent range estimates can be used to compensate for the effect of the ionosphere. The L1 carrier is modulated with a short coarse/acquisition code (C/A code) at a chip rate of about 1 MHz and a longer precision code (P code) at a chip rate of about 10 MHz. The L2 carrier is modulated with the P code only. The P code is dithered in a pseudorandom fashion so that the precision is limited for users other than those in the US military. In addition to the previously mentioned ranging codes, the carriers are modulated by a low-rate data stream carrying a navigation message that includes satellite position and satellite clock correction information. Typical civilian GPS receiver sets achieve a position accuracy of about 100 m and a time accuracy of about 10 ns. It is expected that the dithering of the P code will be eliminated within several years, allowing the accuracy for civilian sets to improve to better than 30 m.

to mobile satellite users will be an extension of those that are available from terrestrial systems, with the result that mobile satellite service offerings will be pulled along by the expansion and convergence that is occurring terrestrially. The upward trend in the data rates will necessitate increased use of the higher frequency bands by mobile satellite systems. In order to achieve the large numbers of users predicted by market studies, the trend toward smaller and less-expensive terminals will need to continue. Small and simple antennas for the mobile terminals will be essential to achieve this goal. New systems must find ways to provide the extra power needed to offer the combination of higher data rates to smaller terminals. For systems based on geostationary satellites, this will require very powerful satellites with extremely large antennas. Because of reduced path loss, for systems using satellites in lower orbits, the size and power of the satellite can be traded off with the altitude of the orbit. Of course, as the altitude of the orbit decreases, the number of satellites needed to provide global coverage increases. A large number of systems are in the planning stage, and one can expect fierce competition based upon cost to the user, range of services, quality of services, and availability. Because it is usually not feasible to overcome blockage, satellite diversity to offer improved availability may become an important issue. Systems based upon geostationary satellites will have an advantage for services requiring broad area coverage, such as point-to-multipoint communications, broadcasting, and wide-area paging. On the other hand, systems based upon lower Earth orbits will have an advantage for global point-to-point communications services, particularly if large transmission delays are undesirable. An example of such a service is global hand-held telephony. From the wide range of technologies and service offerings that characterize planned systems, it is clear that the field of mobile satellite communications is far from being mature.

BIBLIOGRAPHY 1. A. Jahn, et al., Narrow- and wide-band channel characterization for land mobile satellite systems: Experimental results at L-band, Proc. 4th Int. Mobile Satellite Conf., 1995, pp. 115–121. 2. W. Jakes, Multipath interference, in W. Jakes (ed.), Microwave Mobile Communications, New York: Wiley, 1974. 3. E. Lutz et al., The land mobile satellite channel;—Recording, statistics, and channel model, IEEE Trans. Veh. Technol., 40: 375– 386, 1991. 4. R. Barts and W. Stutzman, Modeling and simulation of mobile satellite propagation, IEEE Trans. Antennas Propag., 40: 375– 381, 1992. 5. C. Loo, A statistical model for a land mobile satellite link, IEEE Trans. Veh. Technol., VT-34: 122–127, 1985. 6. R. Cox and P. Kroon, Low bit-rate speech coders for multimedia communication, IEEE Commun. Mag., 34 (12): 34–41, 1996. 7. K. Rijkse, H.263: Video coding for low-bit-rate communication, IEEE Commun. Mag., 34 (12): 42–45, 1996.

TRENDS IN MOBILE SATELLITE SYSTEMS

8. H. Ma and J. Wolf, On tail biting convolutional codes, IEEE Trans. Commun., COM-34: 104–111, 1986.

Increasingly a broader range of services is being offered, with many of the new services requiring data rates that are higher than those currently available. Ultimately the services offered

9. C. Berrou and A. Glavieux, Near optimum error correcting coding and decoding: Turbo-codes, IEEE Trans. Commun., 44: 1261– 1271, 1996.

MOBILE TELECOMMUNICATIONS STANDARDS 10. Y. Yasuda, K. Kashiki, and Y. Hirata, High rate punctured convolutional codes for soft Viterbi decoding, IEEE Trans. Commun., COM-32: 315–319, 1984. 11. G. Ungerboeck, Trellis-coded modulation with redundant signal sets: Part I. Introduction, IEEE Commun. Mag., 25 (2): 5–11, 1987. 12. G. Ungerboeck, Trellis-coded modulation with redundant signal sets: Part II. State of the art, IEEE Commun. Mag., 25 (2): 12– 21, 1987. 13. A. Viterbi et al., A pragmatic approach to trellis-coded modulation, IEEE Commun. Mag., 27 (7): 11–19, 1989. 14. C. Shannon, Communications in the presence of noise, Proc. IRE, 37: 10–21, 1949. 15. K. S. Gilhousen et al., Increased capacity using CDMA for mobile satellite communications, IEEE J. Selec. Areas Commun., 8: 503– 514, 1990. 16. A. Duel-Hallen, J. Holtzman, and Z. Zvonar, Multiuser detection for CDMA systems, IEEE Personal Commun., 2 (2): 46–58, 1995. 17. M. Kayton (ed.), Navigation: Land, Sea, Air and Space, New York: IEEE Press, 1990.

Reading List J. Lodge and M. Moher, Mobile satellite sysetms, in J. D. Gibson (ed.), The Communication Handbook, Boca Raton, FL: CRC Press, 1997, pp. 1015–1031. T. Logsdon, Mobile Communications Satellites, New York: McGrawHill, 1996. S. Kato, Personal communication systems and low earth orbit satellites, Proc. Space Radio Sci. Symp., U.R.S.I., Brussels, Belgium, 1995, pp. 30–42. W. Wu et al., Mobile satellite communications, Proc. IEEE, 82 (9): 1431–1448, 1994. J. Lodge, Mobile satellite communications systems: Toward global personal communications, IEEE Commun. Mag., 29 (11): 24–30, 1991.

JOHN LODGE Communications Research Centre

379

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELECTRICAL...NICS%20ENGINEERING/62.%20Vehicular%20Technology/W7711.htm

}{{}}

●

HOME ●

ABOUT US ●

CONTACT US ●

HELP

Home / Engineering / Electrical and Electronics Engineering

Wiley Encyclopedia of Electrical and Electronics Engineering Mobile Telecommunications Standards Standard Article Girish Patel1 1NORTEL, Richardson, TX Copyright © 1999 by John Wiley & Sons, Inc. All rights reserved. DOI: 10.1002/047134608X.W7711 Article Online Posting Date: December 27, 1999 Abstract | Full Text: HTML PDF (246K)

●

●

● ●

Recommend to Your Librarian Save title to My Profile Email this page Print this page

Browse this title ●

Search this title Enter words or phrases

Abstract The sections in this article are Mobile Communications Standards Organizations Air Interface Standards Network Standards Regulation In The United States About Wiley InterScience | About Wiley | Privacy | Terms & Conditions Copyright © 1999-2008John Wiley & Sons, Inc. All Rights Reserved.

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELE...ENGINEERING/62.%20Vehicular%20Technology/W7711.htm15.06.2008 12:45:24

❍

❍ ❍

Advanced Product Search Search All Content Acronym Finder

MOBILE TELECOMMUNICATIONS STANDARDS

379

In regard to mobile communications, a standard is simply a document that establishes technical requirements for the design, function, interoperability, and interworking of mobile stations, base stations, and mobile telecommunications networks. Basically, there are two different types of standards: de jure and de facto. A de jure standard is formed by committee. These standards can take many years to develop because the process used in committees tends to be long, bureaucratic, and political. Nevertheless, most of the mobile communications standards in use today are the result of standards committees. A de facto standard is the result of a manufacturer or service provider dominating a market. A good example of a de facto standard is Microsoft Windows. Most personal computers (PCs) on the market today uses Microsoft Windows, yet there is no standard in existence that specifies the use of this software on all PCs. Voluntary Standards. De jure and de facto standards can be either voluntary or regulatory in nature. Voluntary standards are adopted by companies on a voluntary basis. There are no rules that say all manufacturers must comply with a voluntary standard. However, the advantages of compliance are many. In regard to mobile communications, a voluntary standard helps ensure that everyone who develops products will design/build their products for interoperability. Without this, only a few subscriber/infrastructure equipment manufacturers would win the market—those with the largest installed base. Other advantages offered by complying with voluntary standards are:

MOBILE TELECOMMUNICATIONS STANDARDS What Is a Standard Although there is no widely accepted and quoted definition of the term standard, the following definition from the 1979 National Policy on Standards for the United States encompasses the essential concept (National Standards Policy Advisory Committee, National Policy on Standards for the United States, 1979):

A prescribed set of rules, conditions, or requirements concerning the definition of terms; the classification of components; the specification of materials, performance, or operation; the delineation of procedures; or the measurement of quantity and quality in describing materials, products, systems, services, or practices.

• Increased productivity and efficiency in industry because of larger-scale, lower-cost production (e.g., the exponential reduction in the cost of hand-held mobile phones in the United States and the rest of the world is the direct result of the fact that the mobile communications standards-making bodies have published a limited number of voluntary standards). • Increased competition by allowing smaller firms to market products. Voluntary standards are quite often at the forefront of technology development, and they provide a valuable source of knowledge to smaller carriers or subscriber/infrastructure equipment manufacturers in emerging markets. • Quality control. By specifying minimum performance requirements for subscriber/infrastructure equipment, voluntary standards enable quality control to be maintained at a high level. Regulatory Standards. Regulatory standards are created by government agencies and must be conformed to by the industry. In general, regulatory standards do not offer any major advantage to the carrier or the subscriber/infrastructure equipment manufacturer, but instead are in place (in most cases) to protect the consumer. Regulatory standards are monitored by government agencies such as the Federal Communications Commission (FCC). These agencies ensure the protection of the public by enforcing standards covering safety, interconnectivity, electromag-

J. Webster (ed.), Wiley Encyclopedia of Electrical and Electronics Engineering. Copyright # 1999 John Wiley & Sons, Inc.

380

MOBILE TELECOMMUNICATIONS STANDARDS

netic emissions, and in some cases health. A good example is the FCC Code of Regulations (Title 47), Part 15—‘‘Radio Frequency Devices,’’ Subpart A—General, Subpart B— Unintentional Radiators, Subpart C—Intentional Radiators, and Subpart D—Unlicensed Personal Communication Service Devices.

The mobile communications standards-making process involves cooperation between hundreds of diverse organizations at many levels—both national and international. Cooperation exists between industrial concerns within a country (e.g., carriers and subscriber/infrastructure equipment manufacturers), between industrial concerns and their national governments (e.g., carriers and regulatory agencies), and between nations. The mobile communications standards-making process is hierarchical in nature and can be categorized into three tiers: • Base standardization • Functional standardization • Conformance standardization The base standardization process requires the participation of international standards-making bodies. Base standards typically incorporate alternative specifications in their requirements in order to satisfy varying regional/national needs. They offer the implementor a degree of flexibility. Adopting the alternative specifications means that, even though the implementation will be in compliance with the base standard, there is no guarantee that equipment based on separate alternative specifications will be interoperable. The problem of interoperability is tackled to a large extent by the functional standardization process, which requires the

International

IS-54-B IS-95 (TDMA) (CDMA)

IS-95 -A

1983

1989 1990

1991 1992 1993

1994

1995

1996

Figure 2. Timeline for North American mobile communications standards.

participation of regional/national standards-making bodies. Confronted with a much more specific set of requirements than international standards-making bodies, regional/national standards-making bodies tend to adopt base standards as functional standards that incorporate only a limited subset of the permissible base standard alternative specifications. The conformance standardization process is the most narrowly focused. Conformance standards prescribe test specifications and methods, which specialized test houses must use when they conduct conformance tests to certify products. Figure 1 shows an example of the mobile communications standards-making process in the United States. What all of this means is that, due to the amount of cooperation required in the standards-making process and its hierarchical nature, standardization takes a long time. How long? Figures 2 and 3 illustrate North American and European mobile communications standards timelines, respectively. The process of migrating from purely analog cellular systems to analog/digital cellular systems has taken approximately 10 years in both North America and Europe. MOBILE COMMUNICATIONS STANDARDS ORGANIZATIONS Mobile communications standards are the result of years of research and development conducted by many different international and national standards organizations. Within these standards organization there are a myriad of study groups

ANSI

TIA

IS-54 (TDMA)

First amps call

Mobile Communications Standards-Making Process

National

EIA/TIA 553 (amps)

JTC air-interface standards IS-136 (TDMA DCCH)

ATIS ITU-R

NMT (900 MHz) launched in Sweden NMT (450 MHz) launched in Sweden

US State Department

European GSM900 begin commercial operations

GSM becomes an ETSI technical committee

UK and Germany DCS1800 begin operations

TACS (900 MHz) launched in UK

ITU-T Carriers and subscriber/infrastructure equipment vendors Figure 1. Example mobile communications standards-making process in the United States. ANSI: American National Standards Institute; TIA: Telecommunications Industry Association; ATIS: Alliance for Telecommunication Industry Solutions; ITU: International Telecommunications Union.

1982

1984

GSM created within CEPT

1986 1989 1990 1991 1992

GSM spec produced

1994 1996 1995 1993

DCS1800 adaptation starts

Figure 3. Timeline for European mobile communications standards.

MOBILE TELECOMMUNICATIONS STANDARDS

and working groups that come and go as the need arises. Many independent forums also evolve. Typically, they are composed of government agencies, carriers, and subscriber/ infrastructure equipment manufacturers representing many different countries. International Standards Organizations ITU. The ITU (International Telecommunications Union) is a United Nations (UN) treaty organization whose purpose is to develop standards that will allow end-to-end compatibility between international radiocommunications and telecommunications networks. Founded in 1865 as the Union Te´le´graphique, the ITU is organized into three sectors: the Radiocommunication Sector (ITU-R), the Telecommunication Standardization Sector (ITU-T), and the Telecommunication Development Sector (ITU-D). Figure 4 shows the organizational structure of the ITU. Even though the ITU has not developed any significant mobile communication standards, it is likely to play a significant role in development of thirdgeneration mobile communication standards. The primary responsibility of the ITU-R (known as the CCIR prior to 1993) is to develop technical standards in the radiocommunication field. The ITU-R is composed of eight study groups (SGs). The key SG for mobile communications standards development is SG 8 (mobile, radiodetermination, amateur, and related satellite services). Under SG 8 is Task Group 8/1 (TG 8/1). TG 8/1 is responsible for International Mobile Telecommunications-2000 (IMT-2000), formerly known as Future Public Land Mobile Telecommunication Systems (FPLMTS). IMT-2000 is the ITU-R name for third-generation digital cellular systems. The primary responsibility of the ITU-T (known as the CCITT prior to 1993) is to develop technical standards in the telecommunications field. The ITU-T is composed of fourteen SGs. The key SGs for mobile communications and IMT-2000 standards development are SG 2 (network and service operation), SG 4 [Telecommunications Network Management (TNM) and network maintenance], SG 7 (data networks and open system communications), SG 11 (signaling requirements and protocols), SG 13 (global infrastructure), and SG 16 (multimedia services and systems). The primary responsibility of ITU-D is to promote and offer technical assistance to developing countries in the field of telecommunications and also to promote the mobilization of the material and financial resources needed for implementation. The ITU-D consists of two SGs. ISO. Established in 1947, the ISO (International Standards Organization) is responsible for many data communications standards, including the open systems interconnection (OSI) model. The ISO is divided into more than 150 technical committees. The key technical committee for mobile communications standards development is the Joint ISO/IEC (International Electrotechnical Commission) Technical Committee 1 (JTC1). Established in 1987, its scope is information technology, and it collaborates closely with ITU-T. Under JTC1 is Subcommittee 6 (SC 6). JTC1/SC 6 deals with telecommunications and information exchange between systems. The ISO is composed of standards bodies from various countries—mostly government agencies responsible for set-

381

ting communications standards within their own governments. The United States representative is ANSI. North American (United States) Standards Organizations ANSI. At the forefront of the United States telecommunications standardization process is the American National Standards Institute (ANSI). Founded in 1918, ANSI is responsible for accrediting other U.S. standards-making bodies. The two key ANSI accredited standards-making bodies for mobile communications standards development are the Telecommunications Industry Association (TIA) and the Alliance for Telecommunications Industry Solutions (ATIS). Figure 5 depicts the U.S. standards organizations. TIA. The TIA was formed in 1988 from the combination of the Information and Telecommunications Technology Group of the Electronics Industries Association (EIA) and the US Telecommunications Suppliers Association. The charter of the TIA is the formation of new telecommunications standards. Specifically, it develops standards for technologies as diverse as telecommunications networks, fiber optics, mobile communications, and satellite communications. The TIA is still associated with the EIA and is an ANSI-accredited standardsmaking body composed primarily of subscriber/infrastructure equipment manufacturers (carriers also participate in the standards committees, but are not full members of the TIA). The TIA has developed most of the mobile communications standards used in the United States today. TIA has changed its charter in 1998 to allow membership of organizations in Canada and Mexico. The TIA primarily develops what are known as interim standards (ISs) (e.g., IS-95-A and IS-136-A). These standards have a limited life span of 3 years. An IS can eventually become a full ANSI standard if it is agreed upon by the larger membership of ANSI. The TIA is composed of many committees that develop mobile communications and other telecommunications standards. Many of the committees are designated as TR committees (the designation is a relic of the term transmission, which was the original technology being standardized in the early days of the EIA). The key TR committees for mobile communications standards development are TR-45 (Public Mobile and Personal Communications Standards) and TR-46 (Public Mobile and Personal Communications Standards). TR-45. TR-45 is responsible for the development of mobile communications and personal communications services (PCS) standards in the licensed 800 MHz and 1900 MHz bands [e.g., Advanced Mobile Phone Service (AMPS), D-AMPS, and CDMA]. Specifically, TR-45 develops ANSI standards, IS standards, Telecommunications Systems Bulletins (TSBs), and technical reports pertaining to the performance, compatibility, interoperability, and service descriptions of mobile communications and PCS. TR-45 also develops and recommends positions on related subjects under consideration by other domestic and international standards forums. TR-45 is composed of six subcommittees (see Fig. 5): • TR-45.1 (Analog Technology—Mobile and Personal Communications Standards) works on analog standards (e.g., EIA/TIA-553).

382

MOBILE TELECOMMUNICATIONS STANDARDS

ITU Plenipotentiary Conference

Council

ITU-R

SG 1

ITU-T

SG 8

ITU-D

SG 2

SG 9

SG 1

SG 2

TG 8/1 SG 3

SG 9

SG 3

SG 10

SG 4

SG 10

SG 4

SG 11

SG 7

SG 11

SG 5

SG 12

SG 6

SG 13

SG 7

SG 15

SG 8

SG 16

ITU-R Study Groups

ITU-T Study Groups

ITU-D Study Groups

Spectrum management Radiowave propagation Fixed-satellite service Science services Mobile, radiodetermination, amateur, and related satellite services SG 9 — Fixed service SG 10 — Broadcasting service (sound) SG 11 — Broadcasting service (television)

SG 2 — Network and service operation SG 3 — Tariff and accounting principles SG 4 — Telecommunications network management (TNM) and network maintenance SG 5 — Protection against electromagnetic environment effects SG 6 — Outside plant SG 7 — Data networks and open system communications SG 8 — Characteristics of telematic systems SG 9 — Television and sound transmission SG 10 — Languages and general software aspects for telecommunications systems SG 11 — Signaling requirements and protocols SG 12 — End-to-end transmission performance of networks and terminals SG 13 — General network aspects (including, GII) SG 15 — Transport network, systems, and equipment SG 16 — Multimedia services and systems

SG 1 — Telecommunication development strategies and policies SG 2 — Development, harmonization, management, and maintenance of telecommunication networks and services, including spectrum management

SG SG SG SG SG

1 3 4 7 8

— — — — —

Figure 4. ITU structure.

MOBILE TELECOMMUNICATIONS STANDARDS

383

ANSI Other organizations

Other TIA technical committees

TIA

TR-45

ATIS

T1

TR-46

TR-45.1

TR-46.1

TR-45.2

Other ATIS technical committees

T1A1

T1E1

T1M1

T1P1

T1S1

T1X1

T1P1.1

T1P1.2

T1P1.3

T1P1.5

T1P1.6

T1P1.7

TR-46.2

TR-45.3 TR-46.3 TR-45.4 TR-46.5 TR-45.5 TR-46.6

JTC

(Disbanded)

TR-45.6

ATIS

TIA Analog Technology Wireless intersystem technology Time division digital technology Radio switching technology Spread spectrum digital technology TR-45.6 — Adjunct wireless packet data TR-45.1 TR-45.2 TR-45.3 TR-45.4 TR-45.5

— — — — —

TR-46.1 — Wireless ISDN multimedia services TR-46.2 — Personal communications services intersystem operation TR-46.3 — Air interface TR-46.5 — PCS 1900 TR-46.6 — Composite CDMA/TDMA

T1A1 — Performance and signal processing T1E1 — Interfaces and protection for T1M1 — Networks T1P1 — Internetwork operations, administration, maintenance, and provisioning (OAM&P) T1S1 — Wireless/mobile services and signaling T1X1 — Digital hierarchy and synchronization

T1P1.1 — International personal communications services T1P1.2 — coordination Personal communications service descriptions and network architecture T1P1.3 — Personal advanced communications system (PACS) T1P1.5 — PCS 1900 T1P1.6 — CDMA/TDMA T1P1.7 — Wideband CDMA

Figure 5. North American mobile communications standards organization.

• TR-45.2 (Wireless Intersystem Technology—Mobile and Personal Communications Standards) develops standards for intersystem operations (e.g., ANSI TIA/EIA41). • TR-45.3 (Time Division Digital Technology—Mobile and Personal Communications Standards) develops TDMA air interface standards (e.g., TIA/EIA/IS-136). • TR-45.4 (Radio to Switching Technology—Mobile and Personal Communications Standards) develops stan-

dards for the interface of base stations to mobile switching centers (MSCs) (e.g., TIA/EIA/IS-634). • TR-45.5 (Spread Spectrum Digital Technology—Mobile and Personal Communications Standards) is responsible for the development of CDMA standards (e.g., TIA/EIA/ IS-95-A). • TR-45.6 (Adjunct Wireless Packet Data Standards) is developing an industry standard based on the cellular digital packet data (CDPD) specifications, which were devel-

384

MOBILE TELECOMMUNICATIONS STANDARDS

oped by the CDPD industry forum. The subcommittee will also work on other data standards for cellular and PCS. Committee TR-45 also has a number of ad hoc groups that address specific technical topics but do not develop standards. The main ad hoc groups at present are the Authentication Ad Hoc Group, the Network Reference Model Ad Hoc Group, the Operations, Maintenance, Administration, and Provisioning (OMA&P) Ad Hoc Group, the Wireless Local Loop Ad Hoc Group, and the International Standards Development Ad Hoc Group. TR-46. TR-46 develops and maintains performance, compatibility, interoperability, and service standards for PCS operating in the licensed 1900 MHz band and the unlicensed 1900 MHz band. TR-46 is now primarily limited to standards that are based on GSM and DCS-1800, which have been standardized in ETSI. TR-46 is comprised of four subcommittees (see Fig. 5): • TR-46.1 (Wireless ISDN Multimedia Services) develops standards for wireless multimedia services. Current efforts include definition of a wideband CDMA technology. • TR-46.2 (Personal Communications Services Intersystem Operation) addresses cross technology issues including RF interference mitigation. TR46.2 published standards for interworking/interoperability between DCS 1900 (GSM) and IS-41-based mobile application parts (MAPs) for 1800 MHz PCS in 1996. • TR-46.3 (Air Interface) is inactive but listed here because of the standards that were developed during the liaison between TR-46 and Committee T1 via the JTC. • TR-46.5 (PCS 1900) concentrates on the enhancement and evolution of the PCS 1900 MHz family of standards [e.g., J-STD 007, An Air Interface for Personal Communications (GSM-based) for 1.8 GHz to 2.2 GHz]. • TR-46.6 (Composite CDMA/TDMA) develops and enhances the standards for the composite CDMA and time division multiple-access (TDMA) mobile communications systems. When the FCC started the process of spectrum allocation in the PCS band, a Joint Technical Committee (JTC) was formed with the ATIS Committee T1 (see Fig. 5) to develop standards for operations in the new spectrum at 1.9 GHz. Several air interface standards were successfully completed and are still awaiting publication. Future revisions of these documents will not be handled by JTC, but rather by a lead organization, as per agreement between the TIA and Committee T1. Now that the work on the PCS standards has been completed, the JTC has been disbanded and the standards work on the air interfaces has been moved back to the TIA and T1 parent subcommittees. ATIS. The ATIS was created in 1983 as part of the breakup of the Bell System. Initially, ATIS was known as the Exchange Carriers Standards Association (ECSA) and sponsored Committee T1 to give the exchange carriers a voice in the creation of interconnection standards, which had previously been done de facto by AT&T.

Committee T1. Established in February 1984, Committee T1 develops standards regarding the interconnection and interoperability of telecommunications networks with end-user systems, carriers, customer premises equipment (CPE), and information/enhanced-service providers. Committee T1 is composed of six technical subcommittees (TSCs) (see Fig. 5) that are advised and managed by the T1 Advisory Group (T1AG): • T1A1 (Performance and Signal Processing) • T1E1 (Interfaces, Power, and Protection for Networks) • T1M1 (Internetwork Operations, Administration, Maintenance, and Provisioning) • T1P1 (Wireless/Mobile Services and Systems) • T1S1 (Services, Architectures, and Signaling) • T1X1 (Digital Hierarchy and Synchronizations) Each technical subcommittee develops standards and technical reports in its designated area of expertise. The key TSC for mobile communications standards development is T1P1. T1P1. Subcommittee T1P1 is composed of six working groups (see Fig. 5): • T1P1.1 (International Personal Communications Services Coordination). • T1P1.2 (Personal Communications Service Descriptions and Network Architectures) addresses reference models and Stage 1 service descriptions. • T1P1.3 (Personal Advanced Communications Systems, PACS) developed J-STD 014, PAC-UA, and PAC-UB. • T1P1.5 (PCS 1900) addresses all aspects of PCS 1900 standards (e.g., J-STD 007). The subcommittee is currently working closely with ETSI Special Mobile Group (SMG) on developing services and features for the Groupe Spe´cial Mobile (GSM) specifications required for North America. • T1P1.6 (CDMA/TDMA) developed TIA/EIA/IS-661. • T1P1.7 (Wideband-CDMA) developed TIA/EIA/IS-665. European Standards Organizations In 1990 the European Community (EC) set up the European Standardization Organization (ESO) to oversee the activities of European standards-making bodies such as the Comite´ Europe´en de Normalisation (CEN), the Comite´ Europe´en de Normalisation Electrotechnique (CENELEC), and the European Telecommunications Standard Institute (ETSI). ETSI. The Confe´rence Europe´ene des Administrations des Postes et des Te´le´communications (CEPT) was formed in 1958 by the European postal, telephone, and telegraph (PTT) authorities to harmonize the development of European telecommunications standards. It covers all of the countries of the EC and the European Free Trade Association (EFTA) in addition to PTTs from various European countries. In 1988, CEPT set up an independent body called the ETSI to conduct all telecommunications standards meetings on its behalf. Although the ETSI is an independent organization funded by its members, who decide on its work program, the EC and

MOBILE TELECOMMUNICATIONS STANDARDS

385

ETSI

Technical committees

Technical committees

Special committees

SMG TC EE

TC ERM

TC HF

TC ICC

TC MTS

TC NA

TC SES TCSTQ

SMG1

SMG7

SMG2

SMG8

SMG3

SMG9

SMG4

SMG10

ECMA TC32C

EBU/ CENELEC/ ETSI JTC

GMM CG

CN

DTA

NBDG

DECT

PTS

MTA

TETRA JEEC

TCSTQ

TC TMN

CTM

SAGE

TC SES

TC TMN

ATA

TIPHON

EASI OCG

BRAN SMG6

USER GROUP

SMG11 SMG1 — SMG2 — SMG3 — SMG4 — SMG6 — SMG7 — SMG8 — SMG9 — SMG10— SMG11—

Service and facilities Radio aspects Networtk aspects Data services GSM operations and maintenance Mobile station testing Base station system (BSS) testing SIM aspects Security group Speech aspects

Figure 6. ETSI structure.

EFTA can also fund the ETSI to produce specific standards of interest to the EC. The ETSI’s main interest lies in the area of telecommunications, although it also participates in issues relating to information technology, for which it cooperates with CEN/CENELEC. The principal role of the ETSI is telecommunications, information technology, and broadcasting standardization. The ETSI membership includes administrations, national standards organizations, subscriber/infrastructure equipment manufacturers, subscriber groups, research bodies, and carriers. Under the ETSI are technical committees (TCs), ETSI projects (EPs), and special committees (SCs). Figure 6 shows the organizational structure of the ETSI, with emphasis on the EP SMG. Within the ETSI, there are a total of twelve TCs (with two joint technical committees), twelve EPs, and six SCs. The key EP for mobile communications standards development is the EP SMG, which develops standards for the GSM family of digital mobile communications systems (i.e., GSM 900, DCS 1800, and UMTS) with a built-in capability for unrestricted world-wide roaming of users and/or terminals between any networks belonging to this family. ETSI SMG is made up of

10 SMG technical committees (STCs) and a Project Team (PT) (see Fig. 6): • • • • • • • • • •

SMG1 (Services and Facilities) SMG2 (Radio Aspects) SMG3 (Network Aspects) SMG4 (Data Services) SMG6 (GSM Operations and Maintenance) SMG7 (Mobile Station Testing) SMG8 [Base Station System (BSS) Testing] SMG9 (SIM Aspects) SMG10 (Security Group) SMG11 (Speech Aspects)

Recently, the EP SMG has instituted a set of formal working procedures with T1P1 for the establishment, evolution, and maintenance of common GSM specifications. Japanese Standards Organizations Wireless standards in Japan are set by two organizations: the Association of Radio Industry Business (ARIB) and the Telecommunication Technology Committee (TTC).

386

MOBILE TELECOMMUNICATIONS STANDARDS

The ARIB was established by the Ministry of Posts and Telecommunications in June 1995 as ‘‘the Realization Center for Efficient Use of Radio Spectrum.’’ The objectives of the ARIB are to develop standards for mobile communications systems and conduct research on the utilization of various wireless technologies. The TTC was established as a private standardization organization in October 1985 to contribute to telecommunications standardization by establishing protocols and standards for interconnections between telecommunications networks. Industry Forums As the cellular industry matured, industry forums composed of carriers and subscriber/infrastructure equipment manufacturers began to evolve. These forums are primarily for the purpose of promoting a particular technology; however, they can also provide a means for specifying technical requirements, which may then be introduced into formal mobile communications standards-making bodies such as the TIA for the purpose of generating a new standard or for incorporation into an existing standard. In regard to mobile communications standards development in North America, the key industry forums are: Cellular Telecommunications Industry Association (CTIA), Personal Communications Industry Association (PCIA), Universal Wireless Communications Consortium (UWCC), CDMA Development Group (CDG) and GSM North America (GSM NA). CTIA. The CTIA established in 1984, represents all of the players in the wireless industry in North America—all commercial mobile radio service providers including cellular, personal communication services, enhanced specialized mobile radio, and mobile satellite services, as well as manufacturers of wireless devices and infrastructure equipment. CTIA has provided a forum for resolution of industry-wide issues and has guided the development of standards in TIA TR-45 and TIA TR-46. PCIA. The PCIA represents wireless communication industry in North America. The association has been at the forefront of advancing regulatory policies, legislation and technology standards for PCS. UWCC. The UWCC is a limited liability corporation established in 1995 to support an association of carriers and subscriber/network infrastructure equipment manufacturers that develop, build, and deploy products and services based on IS-136 TDMA and IS-41 WIN standards. The primary work of the UWCC is accomplished through general membership activities and the principal UWCC forums. The UWCC is composed of three forums: the Global TDMA Forum (GTF), the Global WIN Forum (GWF), and the Global Operators Forum (GOF) (see Fig. 7). CDG. The CDG (CDMA Development Group) is an industry consortium of companies that have come together to develop the products and services necessary to promote the adoption of CMDA mobile communications systems around the world. The CDG is composed of the world’s leading carriers and subscriber/network infrastructure equipment manufacturers.

UWCC board of governors UWCC management: •Executive director •Marketing manager •Administrative assistant

Global TDMA forum (GTF)

Global WIN forum

Global operators forum (GOF)

Figure 7. UWCC structure.

By working together, the members will help ensure interoperability among systems while extending the availability of CDMA technology to consumers. The CDG is composed of several technical groups (TGs), a Promotions/Conferences/Education Group, a Next Generation Systems Group, and an International Working Group (see Fig. 8). GSM NA. GSM North America (GSM NA) is the North American interest group (NAIG) of the GSM MoU Association. This forum has been created to promote PCS1900 service providers’ business interests. The membership consists of the GSM/PCS1900 operators; however, manufacturers are allowed to participate. GSM NA has a number of working groups, which are primarily responsible for addressing and resolving issues from a high-level end user prospective (see Fig. 9). AIR INTERFACE STANDARDS First-Generation Analog Cellular Systems AMPS. The development of mobile communications systems started during the 1930s. The first two-way mobile telephone system was placed into service by the New York City Police Department in 1933. In 1947, a major breakthrough occurred when the concept of cellular mobile communications was discovered. Cellular mobile communications replaces a single large cell (consisting of a single high-power transmitter which provides coverage to a large service area) with many small cells (consisting of many low-power transmitters, each of which provides coverage to only a small portion of a large service area). By dividing a large service area into small cells and using low-power transmitters, frequencies assigned in one small cell can be reused in another small cell. This significantly increases the traffic capacity. Unfortunately, due to technological limitations at the time (e.g., the transistor had not even been invented yet), the implementation of the cellular concept was delayed until the 1980s. In 1975, AT&T was granted a license to operate a developmental cellular radio system in Chicago. Subsequently, AT& T formed a separate subsidiary known as Advanced Mobile Phone Service (AMPS) and on October 13, 1983, in Chicago, AMPS became the first cellular radio system to be put into operation. The modulation scheme for AMPS was FM (frequency modulation) for voice and FSK (frequency shift keying) for signaling. The multiple access scheme was FDMA (frequency division multiple access).

MOBILE TELECOMMUNICATIONS STANDARDS CDG executive board

387

CDG leadership council

CDG Steering committee

Technical groups

Promotions/ Conferences/ Education groups

Next generation system group

International working groups

Data Advanced systems group Interoperability analysis team International roaming Mobile station certification Subscriber interface System test Mobile software download Tiered hierarchical services WLL

Figure 8. CDG structure.

EIA/IS-3. In order to make AMPS commercially viable, the development of an AMPS mobile communications standard was essential. This would ensure that an AMPS subscriber would be able to obtain service from any wireline or nonwireline carrier with a license to operate an AMPS cellular system. The first mobile communications standard in North America was the FCC Office of Engineering and Technology Bulletin No. 53, which included EIA/IS-3 and its revisions. Control Channel Signaling. Signaling information is transferred on the EIA/IS-3 control channel via FSK. The data rate is 10 kbit/s. It is Manchester-encoded to allow the receiver to track the phase and to prevent any dc bias from creeping into the signal via a long series of ones or zeros in the baseband data. On the forward control channel (FCC), 10 words follow a dotting/sync sequence. Words are alternated A, B, A, B, etc. and repeated 5 times for diversity, with the A words designated for mobile stations with even number MINs (mobile station identification numbers) and the B words designated for mobile stations with odd number MINs. An FCC word is 40 bits long. Each word has BCH error correction/detection

included, so that the data content is 28 bits and parity is 12 bits. The FCC is interleaved with busy/idle (B/I) bits. On the reverse control channel (RCC), 5 words follow a dotting/sync sequence. An RCC word is 48 bits long, and each word has BCH error correction/detection included, so that the data content is 36 bits and parity is 12 bits. Each word is repeated 5 times for diversity. Messages are sent by the mobile station on the RCC as coordinated with the base station via the B/I bits sent on the FCC. Voice Channel Signaling. Signaling on the voice channel is divided into in-band and out-of-band signaling. In-band signaling occurs when control signals between 300 Hz and 3000 Hz replace the voice signal. In-band signals on the voice channel are blank-and-burst FSK digital messages. To inform the receiver that a control signal is coming, a 101 bit dotting sequence, which produces a 5 kHz tone, precedes the message. Blank-and-burst signaling on the voice channel differs between the forward and the reverse direction. On the forward voice channel (FVC), messages are repeated 11 times for diversity. Words contain 40 bits and have 12 bits of BCH error correction/detection included. On the reverse voice channel

GSM NA interest group

Security WG: •E911 •Lawful intercept •Fraud

Services and data/vocoder WG

Terminals WG: •Type approval

Numbering WG

International roaming WG

Figure 9. Structure of GSM NA working groups.

Standards WG

Billing and roaming WG

388

MOBILE TELECOMMUNICATIONS STANDARDS

Table 1. Key North American Analog Standards a Standard

Publication Date

Authentication

MWI

CLI

SMS

EIA/TIA-553 (Basic AMPS) TIA/EIA/IS-88 (NAMPS) TIA/EIA/IS-94 (Private PBX) TIA/EIA/IS-91 (553 ⫹ 88) TIA/EIA/IS-91-A (91 ⫹ 94 ⫹ WRE) EIA/TIA-553-A Core Standard ANSI TIA/EIA-691

Sept. 1989 Jan. 1993 Dec. 1993 Dec. 1993 1998 b 1998 b 1998 b

NA NA NA Yes Yes Yes Yes

NA EP NA EP MWN ⫹ EP MWN MWN ⫹ EP

NA EP NA EP AWI/FWI ⫹ EP AWI AWI ⫹ EP

NA EP NA EP AWI-SMS ⫹ EP NA AWI-SMS ⫹ EP

a Definitions: MWI, message waiting indicator; CLI, calling line identification; SMS, short-message service; EP, extended protocol (an optional mechanism to deliver the optional features (MWI, CLI, SMS); MWN, message waiting notification (an order that is used to indicate the number of messages in the mailbox); AWI, alert with information (an order that is used to alert the user that certain information has been delivered with caller ID); FWI, flash with information (an order that is used to send certain information to the mobile without alerting the user); AWI-SMS, alert with information–short-message services (an optional order that is used to alert the user that certain SMS information has been delivered); NAMPS, narrow AMPS (an optional system that users a 10 kHz/V channel); WRE, wireless residential extension (an optional mode of operation); ANSI, American National Standards Institute. b Expected.

(RVC), messages are repeated only 5 times. Words contain 48 bits and have 12 bits of BCH error correction/detection included. Out-of-band signaling consists of control signals above the 300 Hz to 3000 Hz range, which may be transmitted without alternation of the voice signal. Out-of-band control signals sent on the voice channel include the supervisory audio tone (SAT), signaling tone (ST), and dual tone multifrequency (DTMF). EIA/TIA-553. As AMPS cellular systems evolved, there was a need for more sophisticated call procedures and system features. EIA/IS-3 was ill equipped to accommodate this growth. This meant that new mobile communications standards had to be developed. However, although the need to standardize new features was great, it was imperative that the fundamental signaling compatibility requirements for both the mobile station and the base station (which were specified in EIA/IS3 and its revisions) remain unchanged. Therefore, the new standards had to be backward compatible. Table 1 lists all of the major North American analog mobile communications standards that postdate EIA/IS-3 and the new capabilities/system features that they incorporate. In addition to the sophisticated new features that are incorporated in the analog mobile communications standards listed Table 1, interoperability and interconnectivity between AMPS cellular systems and digital cellular systems such as IS-136 (TDMA) and IS-95 (CDMA) soon became a major issue in North America. In order to address these issues, the Telecommunication Industrial Association (TIA) developed EIA/ TIA-553-A, a new revision of the basic AMPS standard. EIA/ TIA 553-A is defined as the ‘‘core’’ analog standard that is common to all TIA analog standards and analog specifications of digital dual-mode standards. With its development the AMPS technology is expected to remain a viable cellular technology in North America for years to come. NMTS and TACS. While the AMPS cellular system was being developed in the United States, several analog cellular systems were being developed in Europe. The Nordic Mobile Telephone System (NMTS) was developed jointly by the telecommunications administrations of Denmark, Finland, Norway, and Sweden. The NMTS system was put into operation

at the end of 1981. The original NMTS system operated at 450 MHz, but it was updated to 900 MHz in 1986. Compared to the AMPS system, NMTS systems has different channel spacing, control channel modulation, and data coding (see Table 2). Another cellular system was also developed in United Kingdom and put into operation in 1985—the Total Access Communication System (TACS). The basic requirements of the TACS system are similar to those of the AMPS system except for channel spacing and additional supplementary facilities such as call forwarding, message waiting, three-way conferencing, and call holding (see Table 2). Second-Generation Digital Cellular Systems North American TDMA. In September 1988, the CTIA (Cellular Telecommunications Industry Association) created a UPR (user performance requirements) document that described a new generation of cellular equipment that would meet the growing needs of the cellular industry. The UPR did not specify the use of either analog or digital technology; it only specified the system capacity requirements and the need for new features. In response to this, major cellular carriers and manufacturers developed a digital cellular system for North America known as the IS-54 system or digital AMPS (D-AMPS). DAMPS is a TDMA (time division multiple-access) system offering 3 times the capacity of AMPS. It provides dual-mode service, which features analog control channels and analog/ digital voice channels. A single D-AMPS RF carrier can support up to three full-rate digital voice channels using the same amount of bandwidth as an analog voice channel (30 kHz). Subsequently, with the development of the IS-136 system in 1995, the digital control channel (DCCH) was introduced, thereby providing D-AMPS users with a platform for implementing personal communications services (PCS) into their existing networks. IS-54. During 1989 the TIA subcommittee TR-45.3 formulated the interim digital cellular standard EIA/TIA/IS-54. Soon thereafter, field trials were conducted to verify that base stations and mobile stations manufactured by different companies met the requirements of this standard. After the publi-

MOBILE TELECOMMUNICATIONS STANDARDS

389

Table 2. Comparison of Three Major Analog Cellular Systems a Specification Mobile TX freq. (MHz) Mobile RX freq. (MHz) Channel separation (kHz) Duplex spacing (MHz) Modulation (voice channel) Modulation (control channel) Bit rate (kbit/s) (control channel) a

AMPS (North America)

NMTS-900 (Nordic Countries)

TACS (UK)

824 to 849 869 to 894 30 45 FM FSK 10

890 to 915 935 to 960 25 or 12.5 45 FM FFSK 1.2

890 to 915 935 to 960 25 45 FM FSK 8

Definitions: FM, frequency modulation; FSK, frequency shift keying; FFSK, fast frequency shift keying.

cation of EIA/TIA/IS-54-A in 1991, a limited number of systems were then put into operation; however, widespread commercial deployment did not occur until after the publication of EIA/TIA/IS-54-B in April 1992. EIA/TIA/IS-54-B became an official ANSI standard in June 1996. It is now known as ANSI TIA/EIA-627. See Table 3. In September and October 1993, the minimum performance requirements for IS-54-B-compatible mobile stations and base stations were standardized in EIA/TIA/55-A and EIA/TIA/56-A, respectively. Both of these interim standards became official ANSI standards in September 1996, and they are now known as ANSI TIA/EIA-628 and ANSI TIA/EIA629. The major distinction between EIA/TIA/IS-54-B and the analog cellular specification EIA/TIA-553-A is the addition of the digital traffic channel (DTC). On the DTC, 1944 bits (or

972 앟/4 DQPSK symbols) are transmitted every 40 ms, yielding a channel rate of 48.6 kbit/s. The 40 ms DTC frame is divided into six slots. Each slot consists of 324 bits. A fullrate user requires two slots every 40 ms. IS-54-B specifies the use of the VSELP vocoder. Every 20 ms the vector-sum-excited linear predictive (VSELP) vocoder produces 159 bits of compressed speech data (7.95 kbit/s). 77 bits are designated as Class 1, and 82 bits are designated as Class 2. The Class 1 bits are given a 7 bit CRC, 5 tail bits are added, and then they are passed through a rate 
, constraint length 5 convolutional encoder. This produces 178 coded Class 1 bits. Together, the coded Class 1 bits and uncoded Class 2 bits add up to 260 data bits every 20 ms (13 kbit/s). The 260 data bits are interslot interleaved. The forward DTC frame is offset from the reverse DTC frame by 207 앟/4 DQPSK symbols (assuming no timing ad-

Table 3. Key North American TDMA Standards Publication ID

Publication Date

EIA/TIA IS-54-B

April 1992

Cellular System Dual-Mode Mobile Station–Base Station Compatibility Standard

TIA/EIA IS-136.1 Rev. 0

December 1994

800 MHz TDMA Cellular–Radio Interface, Mobile Station—Base Station Compatibility—Digital Control Channel

TIA/EIA IS-136.2 Rev. 0

December 1994

800 MHz TDMA Cellular—Radio Interface, Mobile Station— Base Station Compatibility—Traffic Channels and FSK Control Channel

TIA/EIA IS-130

July 1995

800 MHz Cellular System, TDMA Cellular–Radio Interface— Layer-Two Logical Link Control—Radio Link Protocol 1

TIA/EIA IS-135

July 1995

800 MHz Cellular System, TDMA Services for Asynchronous Data and Fax

TIA/EIA IS-641

May 1996

TDMA Cellular/PCS Radio Interface, Enhanced Full-Rate Speech Codec

ANSI TIA/EIA-627

June 1996

800 MHz Cellular System, TDMA Radio Interface, Dual-Mode Mobile Station–Base Station Compatibility Standard

TIA/EIA TSB-73

July 1996

IS-136/IS-136-A Compatibility Issues

TIA/EIA IS-136.1-A

October 1996

TDMA Cellular/PCS Radio Interface, Mobile Station–Base Station Compatibility—Digital Control Channel

TIA/EIA IS-136.2-A

October 1996

TDMA Cellular/PCS Radio Interface, Mobile Station–Base Station Compatibility—Traffic Channels and FSK Control Channel

ANSI TIA/EIA-136

June 1998 a

TDMA Cellular/PCS Radio Interface, Mobile Station–Base Station Compatibility Standard

a

Expected.

Title

390

MOBILE TELECOMMUNICATIONS STANDARDS IS–130 and IS–135

IS–136–A IS–54–B

IS–55–A and IS–56–A

ACC

Minimum performance requirements

Async data and G3 fax IS–641

AVC

ACELP vocoder

DTC

IS–137–A and IS–138–A Minimum performance requirements

VSELP vocoder

DCCH

ACC AVC DTC VSELP

Analog control channel Analog voice channel Digital traffic channel Vector–sum–excited linear predictive ACELP — Algebraic–code–excited linear predictive DCCH — Digital control channel

1900 MHz operation (DCCH and DTC)

Figure 10. Scope of North American TDMA standards.

SMS, OATS. ect.

vance). Signaling on the DTC is a blank-and-burst 앟/4 DQPSK digital message—FACCH (fast associated control channel)—which is passed through a rate 
, constraint length 5 convolutional encoder and interslot interleaved. IS-136. Dissatisfied with the analog control channel’s inefficiency and lack of features, TR-45.3 standardized the digital control channel (DCCH) in December 1994 by issuing the interim standard TIA/EIA/IS-136.1 Rev. 0. The DCCH contains the structure necessary for providing D-AMPS users with the platform for implementing PCS features such as short-message service (SMS) and over-the-air activation teleservice (OATS) into their existing networks. The DCCH did not become available for widespread commercial deployment until the publication of TIA/EIA/TSB-73 in July 1996 (which was composed of IS-136 Rev. 0 functionality along with selected IS-136-A functionality). Shortly thereafter, TIA/EIA/IS-136-A was published in October 1996. It is estimated that TIA/EIA/ IS-136-A will become an official ANSI standard in June 1998 after undergoing major architectural changes to the document itself (the unlayered approach will no longer be maintained). It will then be known as ANSI TIA/EIA-136. In addition to IS-136-A requirements, ANSI TIA/EIA-136 will incorporate the minimum performance requirements for IS-136-A-compatible mobile stations and base stations, which were standardized in July 1996 in TIA/EIA/IS-137-A and TIA/EIA/IS138-A. Figure 10 illustrates the scope of IS-136-A. IS-136.1-A. The major distinction between TIA/EIA/IS136.1-A and EIA/TIA/IS-54-B is the addition of the digital control channel (DCCH). The DCCH has a logical channel structure that adds the functionality of digital control to existing AMPS and D-AMPS systems. The DCCH retains the frame structure and timing used for the DTC that was introduced in IS-54-B, while at the same time providing a complete DCCH structure.

— — — —

DCCH superframes are composed of TDMA time slots 1 and 4. They are used specifically for DCCH messaging. Each superframe contains 32 TDMA time slots (0.64 s). The purpose of the superframe is to provide a structure that will enhance the sleep mode in the mobile station. Each mobile station is assigned a paging channel (PCH) to monitor on the DCCH. The mobile station may go into sleep mode during the remaining portion of the superframe. When the mobile station is in sleep mode, it only wakes to monitor its PCH, thereby allowing it to receive calls while at the same time conserving battery life. Hyperframes are composed of two superframes (1.28 s). The first superframe is called the primary superframe, and the second is called the secondary superframe. Every PCH in the first Superframe is always repeated in the second Superframe. This provides time diversity for the PCH while maintaining sleep mode efficiency. The DCCH contains several subchannels, which will be discussed under two main headings: forward DCCH and reverse DCCH (see Fig. 11).

DCCH Forward

BCCH

SPACH

F–BCCH E–BCCH S–BCCH

SCF

Reverse

Reserved

PCH ARCH SMSCH

Figure 11. DCCH logical channel structure.

RACH

MOBILE TELECOMMUNICATIONS STANDARDS

Forward DCCH. Forward DCCH messages are sent from the base station, mobile switching center (MSC), and interworking function (BMI) to the mobile station (MS). They are composed of three main types: broadcast control channel (BCCH); SMS, paging, and access response channel (SPACH); and shared channel feedback (SCF). A fourth logical channel, marked as reserved, has been designated for future use. The BCCH is composed of the fast BCCH (F-BCCH), extended BCCH (E-BCCH), and SMS BCCH (S-BCCH). These three logical subchannels are generally used to carry system parameters, DCCH frame structure, neighbor cell DCCH list, and global SMS messages. The SPACH is divided into three logical subchannels: paging channel (PCH); access response channel (ARCH); and SMS channel (SMSCH). A PCH is assigned to an MS after it completes the initialization process. This assigned PCH is a specific time slot within the whole PCH subchannel structure. The MS monitors its assigned PCH time slot for messages or incoming calls. When an MS originates a call, it sends an origination request on the random access channel (RACH). The MS then reads the complete ARCH structure and locates the specific message intended for it. This ARCH message assigns an AVC or DTC to the MS. The SMSCH is used to deliver alphanumeric short messages to a specific MS. The SCF logical subchannel is used to provide the MS with the real-time status of the RACH. Reverse DCCH. An MS sends a message to the BMI exclusively on the RACH. Messages sent by the MS are part of the initialization or call origination process. IS-136.2-A. TIA/EIA/IS-136.2-A is a modified version of EIA/TIA/IS-54-B with no major architectural changes to the document itself (the unlayered approach is maintained). One key difference does exist in the forward DTC slot structure, though. IS-136.2-A specifies that 11 of the 12 RSVD bits on the forward DTC must be used for the coded digital control channel locator (CDL) field at the end of the slot. The CDL contains a coded version of the DCCH location (DL) values and provides information that may be used by a mobile station to assist in the location of a DCCH. Another key difference is the explicit support of dual-band operation (800 MHz and 1900 MHz) on the IS-136.2-A DTC. IS-641. The first D-AMPS vocoder was the VSELP (vector– sum-excited linear predictive) vocoder. It was standardized in 1990 as part of EIA/TIA/IS-54-B. Dissatisfied with the voice quality that it achieved, TR-45.3 standardized the ACELP (algebraic-code-excited linear predictive) vocoder in May 1996 by issuing the interim standard TIA/EIA/IS-641. The IS-641 vocoder is commonly referred to as the EFRC (enhanced fullrate vocoder). Widespread commercial deployment commenced in late 1996. Like the VSELP vocoder, the ACELP vocoder specified in TIA/EIA/IS-641 requires a full-rate DTC from IS-136.2-A. However, the ACELP vocoder is more robust to errors (there are 96 protected bits for ACELP versus 77 for VSELP), handles background noise better (ACELP uses predictive vector quantization for the short-term predictor, while VSELP uses scalar quantization), and handles female voices better (ACELP quantizes the long-term predictor in 
-sample increments over the mid-to-high pitch range, while VSELP quantizes in 1-sample increments). Every 20 ms the IS-641 vocoder produces 148 bits of compressed speech data (7.4 kbit/s). 48 bits are designated as

391

Class 1A, 48 bits are designated as Class 1B, and 52 bits are designated as Class 2. The Class 1A and 1B bits are given a 7 bit CRC, 5 tail bits are added, and then they are passed through a rate 
, constraint length 5 convolutional encoder. This produces 216 coded Class 1A and 1B bits. 8 of these bits are punctured, leaving 208 bits. Together, the coded/punctured Class 1A and 1B bits and uncoded Class 2 bits add up to 260 data bits every 20 ms (13 kbit/s). The 260 data bits are interslot interleaved. IS-136-A Services. TIA/EIA/IS-136-A provides the framework for the implementation of several new services and features: • • • • •

• • • • • •

IS-54-B compatibility Short message service (SMS) Over-the-air activation teleservice (OATS) Sleep mode for extending mobile station battery life Hierarchical cell structure, which allows cells of different sizes (macro, micro, and pico) to coexist within the same geographical area Intelligent rescan, which provides an efficient control channel selection process for mobile stations Identity structures to support caller ID and improve authentication Public, private, and residential wireless telephony support Data and fax services Explicit support of dual-band operation (800 MHz and 1900 MHz) Provision for future expansion

Many of these new services and features are more advanced than those offered by other mobile communications standards. North American CDMA. The decision by TR-45.3 in 1989 to adopt TDMA as the air interface access scheme for secondgeneration cellular technology was made prior to the introduction of CDMA (Code Division Multiple Access) digital cellular technology. Shortly thereafter, Qualcomm (along with a number of carriers and subscriber/infrastructure equipment manufacturers) presented CDMA concept to CTIA, April 1990, and conducted successful field trials on a CDMA digital cellular system. On December 5, 1991, the results were presented at the Cellular Telecommunications Industry Association’s (CTIA) ‘‘Presentations of the Results of the Next Generation Cellular Field Trials.’’ At this meeting CDMA garnered more support, and on January 6, 1992, the CTIA Board of Directors unanimously adopted a resolution to prepare ‘‘structurally’’ to accept contributions regarding wideband digital cellular systems. IS-95. In early 1992 the TIA subcommittee TR-45.5 convened to develop spread spectrum digital cellular standards. This effort was initially driven by several US service providers such as Airtouch and Bell Atlantic Mobile, which were interested in CDMA’s superior channel traffic capacity and performance. In July 1993, TR-45.5 adopted TIA/EIA/IS-95 (Mobile-Base Station Compatibility Standards for Dual-Mode Wideband Spread Spectrum Cellular System).

392

MOBILE TELECOMMUNICATIONS STANDARDS Table 4. Key North American CDMA Standards Publication ID

Publication Date

Key Additions

TIA/EIA/IS-95

July 1993

Basic CDMA operations

TIA/EIA/IS-95-A

May 1995

Subscriber access control IMSI support SMS support

TIA/EIA/TSB-74

December 1995

Rate Set 2 (14.4 kbit/s) Service negotiation Status request/response

ANSI J-STD-008

Awaiting publication

1.8 GHz (PCS) support TMSI support

ANSI TIA/EIA-95-B

June 1998 a

Dual-mode, dual-band support High-speed data (up to 115.2 kbit/s) Enhanced soft handoff algorithm Enhanced access procedures

a

Expected.

Several revisions of the IS-95 standard were later released in order to support enhanced system features and capabilities. The latest revision of the IS-95 family is the ANSI TIA/ EIA-95-B standard, which is expected to be published in September, 1998. Table 4 lists all of the IS-95 revisions and the key capabilities/system features that they support. Digital cellular systems based on TIA/EIA/IS-95 support dual-mode (analog and digital) and dual-band (800 MHz and 1900 MHz) operation. Also, mobile stations can hand off from CDMA digital cellular systems both to narrowband analog (TIA/EIA/IS-88) and AMPS cellular systems and to 800 MHz and 1900 MHz CDMA digital cellular systems. On the network side, CDMA is supported by TIA/EIA/IS41 (Cellular Radio Telecommunications Intersystem Operations) and TIA/EIA/IS-634 (the A interface between the MSC and the BS for public 800 MHz operation). IS-95 Techniques. The major distinction between CDMA and the other narrowband technologies such as AMPS or DAMPS (i.e., TDMA) is that signals from different mobile sta-

tions in a CDMA digital cellular system share the same spectrum (1.25 MHz) and are distinguished from each other by unique codes. Many techniques specific to CDMA technology are used to achieve higher channel traffic capacity, better performance, and superior quality of service (QoS). These techniques include channel spreading, power control, soft–softer handoff, and variable-rate speech coding. Forward Link Channels. The IS-95 forward link is channelized by spreading each channel with a unique code. These codes are mutually orthogonal and permit the separation of 64 logical channels on the forward link. There are four types of forward link channels: pilot channel, sync channel, paging channel, and forward traffic channel (see Fig. 12). The pilot channel carries an unmodulated signal and is used to identify unique CDMA coverage area (i.e. cell or sector). There is only one pilot channel per IS-95 forward link. The pilot channel provides nearly perfect phase reference for the coherent demodulation of the other 63 forward channels. It is also used by the mobile station for acquisition and

Forward CDMA channel (1.23 MHz radio channel transmitted by base station)

Figure 12. IS-95 forward link channels.

Pilot ch.

Traffic ch.

Traffic ch. 1

W0

W32

W1

... up to

Traffic ch. 7

Traffic ch. 1

W7

W8

...

Traffic ch. N

... up to

Mobile Traffic power data control subchannel

Traffic ch. 24

Traffic ch. 25

W31

W33

Traffic ... ch. 55 up to W63

MOBILE TELECOMMUNICATIONS STANDARDS

393

Reverse CDMA channel (1.23 MHz radio channel received by base station)

Access ch. 1

...

Access ch. N

Traffic ch. 1

..................................................................... Addressed by long code PNs

tracking of new base stations when the mobile station moves from one coverage area to another. The sync channel conveys information that can be used by the mobile station to synchronize its timing with the IS-95 network. There is only one sync channel per IS-95 forward link. The data rate on the sync channel is 1.2 kbit/s. The paging channels carry information including system parameters, short messages, pages, etc. There can be up to seven paging channels per IS-95 forward link, depending on the paging capacity requirements. The paging channels are also monitored by the mobile station to determine whether or not the serving base station is reliable and whether or not to select another system—IS-95 or AMPS. The data rate on the paging channels can be either 4.8 kbit/s or 9.6 kbit/s. Three types of traffic information can be multiplexed and carried on the traffic channels: primary, secondary, and signaling traffic. Data rates on the traffic channels are flexible in order to support variable-rate vocoding and to reduce the mutual interference between channels. The two sets of data rates supported are Rate Set 1 (9.6 kbit/s, 4.8 kbit/s, 2.4 kbit/ s, and 1.2 kbit/s) and Rate Set 2 (14.4 kbit/s, 7.2 kbit/s, 3.6 kbit/s, and 1.8 kbit/s). The rates can change dynamically on a frame-by-frame basis (i.e., every 20 ms). Reverse Link Channels. The IS-95 reverse link is also distinguished by spreading with a unique code. However, on the reverse link these codes are pseudorandom long codes and are not mutually orthogonal. There are two types of reverse link channels: access channel and reverse traffic channel (see Fig. 13). The access channels are used by the mobile station to initiate transmission with the base station. There can be 32 access channels per paging channel in a cell or sector. Predefined long codes are used for the access channels. The data rate on the access channel is 4.8 kbit/s. The reverse traffic channels, similar to the forward traffic channel, can carry up to three types of traffic: primary, secondary, and signaling traffic. Pseudorandom long codes are generated, based on the mobile station’s electronic serial number (ESN), in order to guarantee uniqueness. Data rates are variable, and both Rate Set 1 and Rate Set 2 are supported. Power Control. On the reverse traffic channel, the mobile station transmit power is tightly regulated to keep each mobile station transmitting at the minimum power level necessary to ensure acceptable QoS. Since CDMA capacity is inter-

Traffic ch. M

Figure 13. IS-95 reverse link channels.

ference-limited, precise control of mobile station transmit power maximizes CDMA channel traffic capacity. Two types of power control are used: open loop and closed loop. Open loop power control allows the mobile station to estimate its transmit power based on the received base station power. Closed loop power control allows the base station to correct the mobile transmit power by sending power control bits on the forward link. On the forward traffic channel, the base station transmit power can also be controlled to maximize system performance. For Rate Set 2, the mobile station sends an erasure indicator bit (EIB) to the base station every 20 ms to indicate whether or not a bad forward link frame (erasure) was received. The base station can use the EIB to adjust its transmit power on the forward traffic channel. Also, both Rate Sets 1 and 2 support signaling messages to convey forward link error statistics, which can be used by the base station to adjust the forward link transmit power. Soft and Hard Handoffs. IS-95 supports seamless soft handoffs. This is accomplished by two or more base stations transmitting the traffic signals for the mobile station. The mobile station combines the signals from these base stations using its RAKE receiver. This provides spatial diversity, thereby improving QoS and coverage. Soft handoffs are assisted by the mobile station, which measures and reports the received signal strengths of the neighboring pilot signals. When the mobile station detects a neighbor pilot of sufficient signal strength, it immediately reports the detected pilot to the base station. The base station can then assign a forward traffic channel to the mobile station on the neighboring cell or sector and direct the mobile station to perform a soft handoff. Hard handoffs are supported during situations when the mobile station is transferred between base stations of different frequency bands and different CDMA frame offsets. Hard handoffs are also supported to transfer a mobile station between CDMA digital cellular systems and analog cellular systems. Variable-Rate Speech Coding. While most speech coding algorithms restrict their design to fixed-channel-rate applications, the IS-95 speech codecs (TIA/EIA-96-B, TIA/EIA-IS733, and TIA/EIA/IS-127) employ algorithms that exploit the time-varying nature of the speech signal in order to encode speech more efficiently in a variable-rate manner. For typical conversational speech, this variable-rate approach enables

394

MOBILE TELECOMMUNICATIONS STANDARDS

Table 5. North American CDMA Services Standards Publication ID

Publication Date

Service

TIA/EIA-96-B

July 1996

8 kbit/s speech service

CDG-27

May 1995

13 kbit/s speech service

TIA/EIA/IS-127

January 1997 a

8 kbit/s EVRC speech service Radio link protocol AT command sets Circuit-switched data Digital fax Packet data STU-III secure voice service

TIA/EIA/IS-707

March 1998

TIA/EIA-637 TIA/EIA/IS-683

December 1995 February 1997

Short-message services Over-the-air service provisioning

TIA/EIA-126

December 1994

Mobile station loopback

a

Expected.

the algorithm to achieve an average speech coding rate less than half that for the fixed-rate vocoders. This lower average rate is one of the factors that gives the IS-95 an overall higher channel traffic capacity. IS-95 Services. IS-95 supports various services such as digital voice, data, and short message services. Table 5 lists the different standards and the services supported by IS-95 digital cellular systems. GSM. The specification of the Groupe Spe´cial Mobile (GSM)—now known more familiarly as the Global System for Mobile Communications for marketing reasons—started in 1982 in Europe with the CEPT. The goal of GSM was to specify a common mobile communications system for Europe in the 900 MHz band. Commercial deployment was planned for 1991. The early years of GSM were devoted primarily to the selection of a radio technology for the air interface. In 1986, GSM began to perform field trials on the different candidate systems that had been proposed. Considerable debate took place over which of these radio technologies represented by the candidate systems (i.e., FDMA, TDMA, or CDMA) was the most suitable. The final decision to adopt a TDMA approach was made in April 1987. On September 7, 1987, in Copenhagen, carriers representing 12 different countries prepared and signed a memorandum of understanding, agreeing on procedures and schedules to procure, build, and test GSM systems. This ensured that different national markets would evolve simultaneously and that international roaming could be successfully implemented. GSM Technical Specifications. GSM mobile communications standards are not structured in the same manner as the North American ones. Therefore, it is not practical to list them all in a table. The structure of the entire set of GSM Phase 1 technical specifications is shown in Fig. 14. In regard to the GSM air interface standard, the key series are 04, 05, and 06. GSM Channel Types. There are two types of GSM logical channels: traffic channels (TCHs) and control channels (CCHs) (see Fig. 15). TCHs carry digitally encoded speech or data and have identical functions and slot formats on both the forward and reverse links. CCHs carry signaling information

between the base station and the mobile station, with certain types of CCHs being defined for only the forward or the reverse link. GSM Air Interface Techniques. GSM uses a TDMA air interface access scheme in conjunction with FDMA. The GSM specifications define a 200 kHz RF channel spacing that uses GMSK modulation and can support either 8 full-rate users or 16 half-rate users. The GSM TDMA frame is 4.615 ms long and is composed of eight 577 애s timeslots. These frames are grouped into multiframes, which are constructed differently for CCHs and TCHs. A group of 26 TCH frames is 120 ms long and defined as a 26 TDMA frame multiframe. For CCHs, 51 frames are grouped into a multiframe of 51 TDMA frames. A superframe is common to both CCHs and TCHs, lasting 6.12 s (i.e., 51 TCH multiframes or 26 CCH multiframes). A group of 2048 superframes forms a hyperframe, which forms the basis for frame numbering in GSM. A hyperframe is 12,533.760 s long (or just under 3
 h). Third Generation Digital Cellular Systems Standards International Mobile Telecommunications-2000 (IMT-2000)— known as Future Public Land Mobile Telecommunication Systems (FPLMTS) prior to 1994—is the ITU name given to third generation (3G) digital cellular systems, which aim to unify today’s diverse digital cellular systems into a common, flexible radio infrastructure capable of offering a wide range of services in many different operating environments around the year 2000. Background. The study of FPLMTS began with the establishment of CCIR Interim Working Party 8/13 (IWP 8/13) in 1985. IWP 8/13 made a significant contribution in the area of FPLMTS spectrum requirements with Recommendation ITU-R 205. In 1987, the ITU World Administrative Radio Conference for Mobile Services (WARC MOB-87) adopted Recommendation 205, which dealt with the need to designate suitable frequency bands for ‘‘international use by future public land mobile telecommunication systems.’’ WARC-92 followed the technical advice of the CCIR and identified 230 MHz of global spectrum in the 1885 MHz to 2025 MHz and

MOBILE TELECOMMUNICATIONS STANDARDS

01 series: General (2 specifications)

05 series: Physical layer on the radio link (7 specifications)

09 series: Network interworking (9 specifications)

02 series: Service aspects (21 specifications)

06 series: Speech coding specification (7 specifications)

11 series: Equipment and type approval specification (7 specifications)

03 series: Network aspects (24 specifications)

07 series: Terminal adaptor for mobile stations (3 specifications)

12 series: Operation and maintenance (7 specifications)

04 series: MS–BS Interface and protocols (16 specifications)

08 series: BS to MSC interface (13 specifications)

Figure 14. GSM Phase 1 technical specifications.

2110 MHz to 2200 MHz bands for FPLMTS. This global spectrum also had a satellite component (1980 MHz to 2010 MHz and 2170 MHz to 2200 MHz bands). Today, studies on IMT-2000 continue in ITU-R (known as the CCIR prior to 1993) Study Group 8 (SG 8) under the auspices of Task Group 8/1 (TG 8/1). This work has resulted in the publication of 15 recommendations covering all technical aspects of IMT-2000, such as requirements for radio interfaces, satellite operation, security mechanisms, and network architecture. In 1997, the ITU approved Recommendation ITU-R M.1225 (Guidelines for Evaluation of Radio Transmission Technologies for IMT-2000). A formal request by the ITU-R director for submission of candidate radio transmission technologies (RTTs) for IMT-2000 was distributed on April 4, 1997, with a closing date of June 30, 1998. Independent evaluations of these proposals, based on Recommendation ITU-R M.1225, will be carried out by various ITU-R members (e.g., TIA TR-45 ISD and ATIS T1P1/TIA TR-46 IAH in the United States) and submitted to the ITU-R by September 1998. Con-

sensus on the key characteristics of the IMT-2000 radio interfaces is planned for March 1999, with the objective of completing detailed ITU-R IMT-2000 standards in time for service to begin shortly after the year 2000. In addition to the ITU-R activities, related IMT-2000 studies are underway in the ITU-T. To facilitate coordination of the activities between the ITU-R and ITU-T, an Intersector Coordination Group (ICG) on IMT-2000 has been formed. 3G Efforts in North America. The North American 3G activities have been concentrating in large part on the evolution of the CDMA IS-95 technology as the RTT for IMT-2000. In 1997, the CDMA Development Group (CDG) agreed on a requirements document that emphasized backward compatibility as one of the key requirements for a 3G system. Soon thereafter, discussions started on a W-cdmaOne proposal, which is a natural evolution of IS-95 and meets all CDG requirements. In the meantime, the UWC also initiated similar standards efforts, which are leading to a proposal for evolution of IS-136 systems to UWC-136 3G RTT submission. Since

GSM

Traffic channel

Signaling channels

Broadcast channels

Common control channels Dedicated control channels

Half rate Full rate Downlink

TCH/H

395

TCH/F BCCH FCCH

Downlink

SCH

PCH

AGCH

Uplink

Slow

Fast

RACH

SACCH

SDCCH FACCH

Figure 15. GSM logical channel structure.

396

MOBILE TELECOMMUNICATIONS STANDARDS

Direct spread

0 MHz

1 MHz

2 MHz

3 MHz

4 MHz

5 MHz

4 MHz

5 MHz

Multicarrier

0 MHz

1 MHz

2 MHz

3 MHz

Figure 16. W-cdmaOne forward link.

most of the 3G discussions in North America have been on WcdmaOne, it is described below. W-cdmaOne Forward Link. The W-cdmaOne proposal uses multiples of the IS-95 1.228 Mchip/s. As illustrated in Fig. 16 there are two approaches being considered for the WcdmaOne forward link. One is the direct spread approach, which scales up the 1.25 MHz IS-95 channels to form a wideband channel. The other is the multicarrier approach, which bundles up multiple 1.25 MHz IS-95 channels to form a wideband channel. Currently, both approaches are being analyzed and compared. In terms of performance and technical parameters, there are no major differences between the two. In terms of deployment, however, there is one notable difference. A multicarrier forward link can be overlaid on top of a narrowband IS-95 forward link because orthogonality is preserved. This feature is particularly attractive for D, E, or F block PCS carriers, because without overlay they may not be able to deploy any 3G systems without having to disable one or more of their existing second-generation (2G) systems. Major enhancements to the forward link include: QPSK modulation and QPSK spreading, which doubles the number of Walsh codes available; an auxiliary pilot to provide spot coverage and beamforming, which further reduces interference; continuous transmission, which reduces interference with medical devices; and variable- length Walsh codes (shorter Walsh codes for higher-data-rate users). Faster forward power control at 800 bit/s is also provided. W-cdmaOne Reverse Link. Figure 17 illustrates the WcdmaOne reverse link. The reverse link is mainly just a scaled-up version of the IS-95 reverse link. The most significant modification is the use of four Walsh code channels as a dedicated pilot channel, fundamental channel, supplemental channel, and control channel.

Carrier A Guard

Carrier B

Carrier C Guard

Wideband carrier 1.25 MHz

1.25 MHz

5 MHz Figure 17. W-cdmaOne reverse link.

The user-specific pilot channel is used for coherent detection, which makes the reverse link demodulation similar to the forward link. The fundamental channel is used mainly for voice and signaling traffic (same as the IS-95 reverse link). The supplemental channel is the high-speed data channel, with a different number of repetitions giving different data rates. Finally, the control channel is designed to help the base station efficiently schedule forward and reverse link resources and realize fast forward power control. Other enhancements include continuous transmission for reducing interference, and for allowing different types of traffic (e.g., voice/signaling versus data) to be sent at different power levels, which will substantially maximize the system capacity. 3G Efforts in Europe. 3G activities in Europe have also been very intense. At an ETSI SMG2 meeting held on January 28 to 29, 1998 in Paris, France, a consensus was reached on the radio interface for universal mobile telecommunications systems (UMTS). The solution, called UTRA (UMTS terrestrial radio access), draws on both W-CDMA and TD/CDMA technologies. 316 delegates representing carriers, subscriber/infrastructure equipment manufacturers, government administrations, and research bodies agreed on the following: 1. In the paired band [i.e., frequency division duplex (FDD)] of UMTS, UTRA adopts the radio access technique proposed by the ETSI Alpha group, that is, WCDMA (wideband code division multiple access). 2. In the unpaired band [i.e., time division duplex (TDD)] of UMTS, UTRA adopts the radio access technique proposed by the ETSI Delta group, that is TD/CDMA (time division/code division multiple access). 3. During the process, it was agreed that the following technical parameters shall be objectives: • Low-cost terminal • Harmonization with GSM • FDD–TDD dual-mode operation • Fit into 2 ⫻ 5 MHz spectrum allocation According to ETSI SMG2 #24 Document TDoc SMG 903/97, some key performance enhancement features of the W-CDMA proposal include: downlink antenna diversity; transmitter diversity; receiver structures; adaptive antennas; and support of relaying and ODMA (Opportunity Driven Multiple Access, an intelligent relaying protocol). The document also indicates that the W-CDMA system supports interfrequency handover for operation with hierarchical cell structures and intersystem handover with 2G systems such as GSM. Another SMG2 document—ETSI SMG2 #24 Document Tdoc SMG 897/97—points out that the TD/CDMA proposal was specifically designed for the purpose of building the IMT2000 system on top of the proven GSM technology (see Fig. 18). Key performance enhancement features of the proposal include: base transceiver station (BTS) antenna hopping; frequency and time slot hopping; directive and/or adaptive antennas; faster and quality-based power control; relaying and advanced relay protocols such as ODMA. Hierarchical cell structure and handovers are also supported. 3G Efforts Elsewhere and the Final Convergence. At least three standards are foreseen as potential members of the

MOBILE TELECOMMUNICATIONS STANDARDS

397

One time slot

8 TCH per time slot

Codes 1–8

1.6 MHz

Frequency

Wb–TDMA/CDMA

1

2

3

4

5

6

7

8

Time

Frame with 8 time slots

One time slot

One TCH per time slot

GSM

200 kHz

Figure 18. TD/CDMA and GSM.

IMT-2000 family. These are shown in Table 6. In principle, any of these four regional standards could be recognized as IMT-2000 family members with the understanding that a regional standard can also be deployed outside the region where it originated from. In other parts of the world, 3G proposals are also being discussed and evaluated, especially in Asian countries such as Japan, Korea, and China. For example, Japan has been making progress in consolidating proposals from NTT DoCoMo, which is aligned with the ETSI W-CDMA proposal, and the W-cdmaOne proposal. Currently, the key remaining difference is the chip rate, which is set to 4.096 Mchip/s in W-CDMA and 3.6864 Mchip/s (3 times the 1.2288 Mchip/s IS-95 chip rate) in W-cdmaOne. It is anticipated that major progress will be made down the road, especially in the REVAL process, for converging different regional proposals into a possible global standard.

The basic mobile communications network architecture consists of a number of network elements and interfaces (see Fig. 19). The principal network elements include: • Mobile station (MS) • Radio base station system (BSS) • Mobile switching center (MSC) • Home location register (HLR) • Visitor location register (VLR) • Authentication center (AC) • Short-message service–service/message center (SMS-SC or SMS-MC) The principal network interfaces include: • Um air interface between the MS and BSS • A interface between the BSS and MSC

NETWORK STANDARDS

• B interface between the MSC and VLR

In regard to mobile communications systems, network elements are not directly related to the air interface between mobile stations and base stations. Network elements provide switching system functions, call and services control, mobility management, and data-based functions, while the air interface provides the means for radio communication between mobile stations and base stations.

• C interface between the MSC and HLR • D interface between the VLR and HLR • E interface between MSCs (In the majority of network configurations, the VLR functionality is integrated or colocated with the MSC.)

Table 6. Possible 3G Family Members Standard:

Europe—UMTS

U.S.—TDMA

U.S.—CDMA

Radio:

W-CDMA for FDD TD/CDMA for TDD

UWC-136

W-cdmaOne

Network:

GSM Map evolution

IS-41 evolution

IS-41 evolution

Standards body:

ETSI

ANSI TIA

ANSI TIA

a

Expected.

398

MOBILE TELECOMMUNICATIONS STANDARDS

SMS–SC/MC

AC

HLR

C Transceiver

D B VLR

Mobile switching center

A interface

Base transceiver station

Base transceiver station

Um MS

E VLR

Mobile switching center

Base station system Public network Figure 19. Mobile communications network architecture.

Network Elements Base Station System. The mobile station communicates to the network via the air interface. The base station system (BSS) includes a base station controller (BSC) and one or more base transceiver stations (BTSs). The BSC manages radio channels on the radio interface and handovers. Radio transmission and reception devices, including antennas, and all radio interface signal processing are contained in the BTS. The BSS terminates the air interface from the mobile station and then connects the network signaling and user traffic to the mobile switching system over the A interface. The BSC manages the air interface to the mobile station and usually any associated air interface encryption protection. The BSC, along with the MS and MSC, manages the handoff between cells. When cells span MSCs, the handoff signaling for the cells is exchanged between MSCs. This type of handoff is referred to as intersystem handoff. Mobile Switching Center. The MSC provides switching and call control for basic and supplementary services (call origination and termination, call transfer, call hold, etc.), connection to the public network, and connection to the user’s home system during roaming. The originating MSC accesses location and routing information from the HLR, and routes the call to the destination/serving MSC. Location Registers. The home location register (HLR) and visitor location register (VLR) store user subscription and location information. The HLR provides for user registration, user location, and user profile information (features and services subscribed to by the user). Each service provider has an HLR for its subscribers. The HLR can support multiple MSCs and can either be integrated into a MSC or exist on a separate, centralized platform. In a roaming scenario the serving system MSC/VLR communicates with the user’s home HLR to obtain user profile and user authentication information. The information is then stored in the VLR. Authentication Center. The authentication center (AC) provides for user verification and manages security data used in authentication. The HLR communicates with the AC to validate users of the network.

Short-Message Service–Service/Message Center. The SMSSC or SMS-MC is responsible for the transmission of a short message to/from a mobile station. Network Interfaces Network interfaces carry user traffic and network signaling between the network entities. This signaling can be classified as call and service control, mobility control, or resource management. Call signaling includes origination and termination requests and user signaling such as busy tone and ringing. Service control includes user requests such as transfer, forward, hold, and conference. Mobility control includes signaling to register a user, authenticate a user, provide user profile information in a roaming environment, and locate a user. Resource management includes management of radio resources (air traffic and signaling channels), and traffic and signaling channels between the BSC and MSC. A Interface. The A interface provides the communication capability between the BSC and MSC and carries user traffic, call and service control signaling, mobility control signaling, and radio resource management signaling. This interface is used for: • Allocation and maintenance of terrestrial channels (voice trunks) that connect the MSC and the BSS • Control of operations and resources that are a shared responsibility of the MSC and BSC • Transparent passing of information from the mobile set (MS) to the MSC. GSM A Interface. In GSM standards the A-interface signaling protocol is also referred to as the BSSAP (BSS application part) protocol as a user function of Signaling System No. 7 (SCCP and MTP); see Fig. 20. BSSAP is further subdivided into two separate functions: • DTAP (direct transfer application subpart) is used to transfer call control and mobility management messages to and from the MS. DTAP messages are not interpreted by the BSS; they pass transparently through to the MSC. • BSSMAP (BSS management application subpart) is used to coordinate procedures between the MSC and BSS re-

MOBILE TELECOMMUNICATIONS STANDARDS BSS

A interface

MSC

DTAP: MM, CC, SS, ect.

BSSMAP

SS7:SCCP SS7:MTP CC: Call control MM: Mobility management RR: Radio resource SS: Supplementary service SS7:Signaling system No. 7

Trunk

Figure 20. GSM A interface protocol stack.

lated to the MS (e.g. resource management or handover control), and coordinate connection between the MSC and BSS. The latest revision of the GSM standards, GSM ’96, was released in 1997. Table 7 lists the standards related to the A interface. North American SS7-Based A Interfaces. The specification that defines an open A interface for North American mobile communications systems is called IS-634 (MSC-BS Interface for Public Wireless Communications Systems). It uses a standard SS7 carriage protocol. IS-634 is developed by the TIA subcommittee TR-45.4 and supports AMPS/TDMA/CDMA/ NAMPS, (narrow AMPS) through common messages and procedures, thereby providing some air interface isolation from the MSC. Table 8 lists the revisions of the IS-634 standard. The IS-634 standard is based on the GSM A-I/F standard and includes most of the GSM BSS Management Application Part in addition to a subset of GSM Call Control and Mobility Management. It also includes modifications needed to support North American air interface standards and uses a modified subset of the GSM DTAP-like messages with interworking to create/convert messages in the BSS. Figure 21 gives a comparison between IS-634 and GSM A interface architectures.

Table 7. GSM A Interface Standards Revisions GSM A Interface Specifications

1997 Revision

Title

GSM 4.07

5.2.0

Mobile radio interface signaling layer 3: general aspects

GSM 4.08

5.7.0

Mobile radio interface layer 3 (DTAP)

GSM 8.02

5.1.0

BSS-MSC interface principles

GSM 8.06

5.1.0

Signaling transport mechanism specification for BSS-MSC I/F

GSM 8.08

5.6.3

MSC-BSS layer 3 specification (BSSMAP)

399

MAP Interfaces. The protocol that is used for communication between the network interfaces in a mobile communications system is referred to as the mobile application part (MAP). There are two main MAP standards: ANSI-41 (used by AMPS-based systems), and GSM MAP (used by GSMbased systems). The B interface between the MSC and VLR, the C interface between the MSC and HLR, the D interface between the VLR and HLR, and the E interface between the MSCs all use the MAP protocol. ANSI-41. The network standard that is used by AMPS family of mobile communications systems (AMPS, TDMA, CDMA) for intersystem signaling is referred to as ANSI-41 (or ANSI TIA/EIA-41). It was developed and is maintained by the TIA subcommittee TR-45.2. ANSI-41 specifies the protocol used by interfaces among elements in the cellular core network (e.g., HLR, VLR, MSC, AC). It supports all radio access technologies—AMPS, TDMA, and CDMA. ANSI-41 provides intersystem signaling support for cellular features and services as described in ANSI TIA/ EIA-664. Examples of the services and capabilities include: • • • • •

Call origination and termination while roaming Inter-MSC handoff Call forwarding Remote feature activation/deactivation Mobile station ‘‘flash’’ capability (three-way calling, call waiting, call transfer) • SMS (short message service) • CNIP (calling number presentation) • MWN (message waiting notification) and voice mail retrieval

The ANSI-41 protocol uses ANSI TCAP of Signaling System No. 7 (SS7) for intersystem operations. The ANSI-41 messages may be transported via X.25 or SS7 signaling links. Interim Standard 41 (IS-4) Revision 0 was published in 1988 and has gone through many revisions since then. IS-41 Revision C was published in February 1996 as a TIA interim standard. IS-41 Revision C has now become a full ANSI standard, ANSI/TIA/EIA-41 Revision D, and was published in December 1997. The next revision of ANSI/TIA/EIA-41, Revision E, is planned to go to ballot in December 1998. This revision will include the contents of the numerous standalone feature documents (e.g., TDMA, CDMA, Data, E911, IMSI, WIN) that have been developed by the TIA subcommittee TR45.2 over the last two years. Table 9 lists the revisions of the IS-41 standard. GSM MAP. The standard that is used by GSM mobile communications systems for intersystem signaling is also called the mobile application part (MAP). It was developed and is maintained by ETSI SMG with input from ANSI T1P1 for North American requirements. GSM MAP is specified in ETSI GSM 09.02 specification. The MAP is the GSM protocol used for mobility management signaling between network subsystem nodes. It is used to exchange location and service information between the MSC, HLR, and VLR. The MAP protocol is implemented on top of the ITU-T Transaction Capabilities Application Part (TCAP) and is transported over ITU-T SS7 and ANSI SS7 (for PCS 1900). The MAP uses protocol version and application

400

MOBILE TELECOMMUNICATIONS STANDARDS

Table 8. Revisions of the IS-634 Specification Reference

Publication Year

Key Functions/Additions

IS-634 Revision 0

1995

Call delivery Supplementary services Authentication/voice privacy Inter-BS/intersystem hard handoff

TSB-80

1996

Corrections to IS-634 Message waiting indication (MWI) Short-message services (SMS) Inter-BS/Intersystem soft handoff voice Frame format based on subrate circuit, via the MSC

IS-634 Revision A

1998 a

An alternative architecture (architecture B) Over-the-air service provisioning (OTASP) Circuit data (async. and G3 fax) Simplified call setup procedures Service negotiation Test calls ATM transport Packet-mode-based voice frame format for CDMA Direct BS–BS connection for inter-BS/intersystem soft handoff (IBSHO/ISSHO)

ANSI/TIA/EIA-634

1998 a

Promotion of IS-634 to full ANSI standard

a

Expected publication date.

context negotiation mechanisms for forwards and backwards compatibility. The latest revision of the GSM MAP (i.e., MAP ’96) was released in 1997. It is based on the Phase 2 MAP with specific enhancements to support GSM ’96 network features and IN capabilities. MAP ’96 uses a new Protocol Version, V3. In addition, new extensibility mechanisms were introduced to ease the burden of introduction of standardized and proprietary protocol elements. Some of the most important network features in GSM ’96 are: • • • • •

CAMEL Phase 1 (GSM IN) High-speed circuit switched data Support of optimal routing Enhanced data services and SMS Explicit call transfer

The ETSI standard body is currently developing the MAP ’97 (or ’98) revision. New parameters for new features will be included via the extensibility mechanism introduced in the MAP ’96 release. Table 10 lists the standard for the GSM MAP protocol. REGULATION IN THE UNITED STATES There are two aspects to the regulation of mobile communications. First, mobile communication carriers may be regulated in terms of the service they offer, the tariffs they charge, and the interconnections that they make with other carriers—just like any wireline carrier. Second, due to their wireless transmission medium, mobile communications carriers are dependent upon how the usage of the frequency spectrum is regulated and allocated. Since regulation and policies vary from country to country, the focus of this section is on the United States, with some

A

GSM Vocoder BTS

Vocoder control

MSC

BSC

Air interface GSM

GSM

DTAP BSSMAP A

N.A. air I/F’s Vocoder BTS

Vocoder control

MSC

BSC

Air interface

Figure 21. Architecture comparison between IS-634 and GSM.

AMPS CDMA TDMA NAMPS

DTAP-like Voice Signaling

BSMAP

IS-634

MOBILE TELECOMMUNICATIONS STANDARDS

401

Table 9. Revisions of the IS-41 Specification Reference

Publication Year

IS-41 Revision 0

1988

Inter-MSC handoff

IS-41 Revision A

1991

Automatic roaming Call delivery Call forwarding Remote feature control ANSI TCAP message encoding

IS-41 Revision B

1992

Dual-mode (TDMA) handoff MS flash after handoff MIN-to-HLR global title translation

IS-41 Revision C

1996

Authentication/voice privacy Short-message service Calling number presentation Message waiting notification Support of CDMA mobiles Flexible alerting groups Subscriber PIN access Voice message retrieval

ANSI/TIA/EIA-41 Revision D

1997

Promotion of IS-41 to full ANSI standard

ANSI/TIA/EIA-41 Revision E

1999 a

PCS multiband support Digital control channel support Circuit mode data Enhanced emergency services Wireless number portability Wireless intelligent network Internalization of ANSI-41

a

Expected.

discussion of issues regarding spectrum allocation for the IMT-2000 spectrum. This will give the reader a general idea about some of the issues surrounding the regulation of mobile communications. Telecommunication Regulation Deregulation of Telecommunication. Until the 1960s, telecommunications in the United States was regulated as a common carrier activity with AT&T and a few independents being classified as common carriers. Regulation at the state level was administered by Public Utility Commissions (PUCs), and at the federal level by the Federal Communication Commission (FCC). The FCC was established by the US government after the enactment of the Communications Act of 1934. The structure of telecommunications regulation was such that competition was not allowed in either the local and long-distance services or the rental of customer premise equipment (CPE)—phones, key systems, etc. The FCC decision on ‘‘Carterfone’’ in 1968 established the consumers’ right to interconnect equipment provided by noncommon-carriers that was technically compatible with the ex-

Table 10. GSM MAP Protocol Standards Revisions GSMMAP Specification GSM 9.02

Key Functions/Additions

1997 Revision

Title

5.7.0

Mobile Application Part (MAP)

isting network. The decision created a new competitive market. In 1969, Microwave Communications Inc. (MCI), a small startup company with a dream, filed an application with the FCC to carry voice and data on dedicated private lines between Chicago and St. Louis (principally for large businesses that needed private-line long-distance connections between their branches in the two cities). Even though this was a specialized carrier application (i.e., MCI would not be competing with AT&T’s long-distance services offered to the public), AT&T opposed the application on the basis that it was ‘‘cream skimming.’’ AT&T was required to maintain an entire nationwide network, and in order to subsidize the extension of phone lines to rural customers, it charged urban and business users more for telephone service. MCI would not have to do this. The FCC granted the application and soon thereafter more applications poured in from other companies wanting to offer similar services. Inundated with applications, the FCC created a new class of carrier for private line service—the specialized common carrier. In 1971, the FCC decided to permit the specialized common carriers to interconnect with the eighteen Bell operating companies’ (BOCs) local exchanges to allow users of their networks to reach any telephone outside the network. Sensing that competition might be possible in the long-distance service market, the US Department of Justice filed its third antitrust suit against AT&T in November 1974, alleging monopolization of both long-distance service and the manu-

402

MOBILE TELECOMMUNICATIONS STANDARDS

facture of telecommunications equipment. The trial began in 1981 and resulted in the divestiture of AT&T into a long-distance service carrier (and telecommunications equipment manufacturing company) and the BOCs. The era of true longdistance service competition had begun; however, the local service market still remained monopolistic and regulated. Telecommunications Act of 1996. The Telecommunications Act of 1996 is the culmination of the reform started with the breakup of AT&T. It is based on the premise that no sector of the telecommunications market should be immune from competition. Its intent is to make regulatory policy a catalyst for investment and to create a truly competitive environment. The highlights of the act are: • Allow entry into the long-distance service market by regional BOCs (RBOCs) after meeting the criteria of competition in their region. • Set fair tariffs and interconnection rules. • Allow RBOCs to enter into manufacturing partnerships. • Allow public utility companies to enter into the telecommunications business. • Allow cable-TV–telephone company cross-ownership. • Review telecommunications rules every two years. The Telecommunications Act of 1996 is investor-friendly. With the removal of cross-ownership barriers, a number of joint ventures, alliances, and mergers/acquisitions have occurred. In the end, it is anticipated that only the most competitive carriers will prevail—providing both wireline and wireless services. Mobile Communications Regulation Radio broadcasting has been regulated by the US government since the Radio Act of 1912 (prompted by events involving failed radio communications such as the 1912 Titanic disaster). The Radio Act of 1927 created the Federal Radio Commission, which regulated all radio frequency use, including mobile communications. When the FCC was formed in 1934, the Federal Radio Commission was incorporated into it. The management of frequency spectrum in the United States is done jointly by the National Telecommunications and Information Management (NTIA) for federal government users and by the FCC for all other users. The allocation of frequency spectrum for cellular and PCS is discussed below. Spectrum Allocation for Cellular Service. While cellular technology was ready for deployment by the early 1970s, the unwillingness of opposing interest groups to compromise delayed its commercial introduction until 1983. The FCC completed its report on how to allocate frequency spectrum for cellular in 1970. Another dozen years would pass before the actual assignment of spectrum took place and commercial mobile communications systems were up and running. Originally, only a single carrier was to be allowed to provide cellular service in a given area (i.e., a wireline carrier). This policy however would have mirrored the earlier monopolistic practices of the telecommunications industry. Confronted with considerable opposition, the FCC modified its rules and issued a new order in 1981 to create a ‘‘duopoly’’ for cellular.

Uplink

Downlink

A

B

A’ B’

11

10 1.5 2.5

A

B

A’ B’

11

10 1.5 2.5

MHz

824

849

MHz

869

894

f (MHz)

License A, 25 MHz—MSA and RSA License B, 25 MHz—MSA and RSA Figure 22. US cellular spectrum.

The country was divided into 306 metropolitan service areas (MSAs) and 428 regional service areas (RSAs). Two licenses were awarded in each type of area. In 1974, the FCC allocated 20 MHz of bandwidth per carrier in an area (this would be increased by 5 MHz in 1989 due to a shortage of channels). The two 20 MHz groups were identified as block A and block B (also known as the A and B bands). Figure 22 shows the cellular spectrum allocation in the United States. In each area, the wireline carrier was awarded the license in the B band (to be operated by a separate subsidiary), while the A band licenses were initially awarded on a comparative bidding basis to nonwireline carriers. When the process of evaluating essentially similar bids proved to be too slow, the FCC finally resorted to a lottery scheme—creating a bonanza for the winners. As the buildout of cellular service occurred, a large number of individual A-band licenses were consolidated by the larger carriers (e.g., AT&T/McCaw Cellular). The B-band carriers, mostly BOCs, also acquired licenses on the A side outside their territory. Consequently, the cellular licensing process ended up creating a handful of large ventures instead of many smaller, highly competitive carriers as had originally been desired. PCS Spectrum Allocation. Soon after the frequency spectrum in the 2 GHz band was set aside for third generation mobile communications systems by the ITU in 1992, the FCC started looking at ways of allocating this spectrum in United States. The FCC decided upon the final procedure for awarding licenses in the 2 GHz band for PCS on June 9, 1994. This time the country was divided into major trading areas (MTAs) and basic trading areas (BTAs) based upon the Rand McNally definitions. There are 51 MTAs and 493 BTAs. The MTAs consist of a number of BTAs; consequently the licenses for MTAs and BTAs overlap. The 150 MHz available for PCS was divided into six distinct blocks, or bands—A, B, and C, each with 30 MHz (full duplex), and D, E, and F, each with 10 MHz (full duplex)—and an unlicensed band of 20 MHz in the middle (see Fig. 23). Each MTA has an A and a B license assigned; each BTA has a C, a D, an E, and an F license assigned. Consequently, there are a total of approximately 2000 licenses, each market having as many as six PCS service providers in addition to the two cellular service providers—a highly competitive landscape. The FCC also made a rule stating that no carrier can own more than 45 MHz in any market. This allowed existing cellular carriers (with 25 MHz of li-

MOBILE TELECOMMUNICATIONS STANDARDS

403

Paired bands

• License A, B 30 MHz (15 MHz split) • License C 30 MHz (15 MHz split) • License D, E, F 10 MHz ( 5 MHz split) Total: 140 MHz, including 20 for unlicensed 102 licenses in MTA blocks A, B 493 licenses in BTA block C 1479 licenses in BTA blocks D, E, F

Reserve

5 21 0 60 22 00

21

10

Reserve

Broadcast auxiliary

90 19

70

C

19

19

50

BTA MTA MTA A D B E F

30

DATA

20

Licensed

19

19

19

10

C

90 18

70 18

18

50

BTA MTA MTA A D B E F

VOICE

BTA BTA BTA

21

Unlicensed

Licensed

Frequency (MHz)

51 MTAs (major trading areas) 493 MTAs (basic trading areas) 493 MTAs

Approx. 2000 licensees

Figure 23. US PCS 1900 spectrum.

censed cellular spectrum) to purchase two 10 MHz licenses (but not a 30 MHz license) in the same market. The licenses were awarded by auction. The auction process was conducted electronically and was completed in February 1997. The auctions generated considerable revenue for the federal government, and other countries are likely to emulate the US PCS auction process.

In Europe, other than the spectrum for DECT, most of the IMT-2000 spectrum is set aside for UMTS (universal mobile telecommunications systems)—the European version of IMT2000. Similarly in Japan, other than spectrum for PHS, the spectrum for IMT-2000 is reserved. However, in the United States a large part of the IMT-2000 spectrum has already been assigned for PCS. This situation will create challenges for global roaming, which is the objective of IMT-2000.

IMT-2000 Spectrum. In 1992, the ITU identified 230 MHz of global spectrum in the 1885 MHz to 2025 MHz and 2110 MHz to 2200 MHz bands for IMT-2000. This global spectrum also contained a satellite component (1980 MHz to 2010 MHz and 2170 MHz to 2200 MHz bands). Figure 24 shows the IMT-2000 frequency spectrum.

00

MHz

22

70

MSS

21

10 21

25

10

IMT-2000 20

19

18

80

MSS

85

IMT-2000

20

ITU

Type Approval for Wireless Equipment. The FCC rules and regulations require products that emit radio signals to obtain FCC type approval before they can be offered for sale in the United States. The approval applies to products that are designed to emit radio signals (e.g., cellular phones) as well as

DECT 90 22

30 22

00 22

21

70

MSS

10 21

40

10

UMTS 20

19

80

MSS

20

19

00 19

18

80

UMTS

20

Europe

*DCS-1800 1717–1785 MHz paired with 1805–1880 MHz

PCS Unlicensed Reserved

Figure 24. IMT-2000 spectrum.

00 22

21 5 21 0 6 21 0 65

10

MSS 21

25 20

90

MSS

19

30 19

00

Reserved Broadcast auxiliary

PCS licensed 10

PCS licensed 19

18

50

USA/ Region2

MHz

22

00

MSS

22

10 21

25

3rd gen. 20

19

18

80

MSS 10

3rd gen.

8 18 5 95 19 18 .1

PHS

20

Japan

MHz

MHz

404

MODELING AND SIMULATION

to products that emit radio signals as a side effect (e.g., personal computers). The FCC has assessed substantial fines to manufacturers that offer products for sale that do not have FCC approval. Products can be displayed at trade shows before approval, if a notice indicates they will not be offered for sale until approval is obtained. The type approval process for radio products includes UL (Underwriters Laboratories) electrical safety compliance, outof-band signal compliance, limits for equipment connected to the PSTN, power limits, and compliance with new FCC human exposure limits. The human exposure limits mostly affect cellular phone design; the shielding and antenna configuration can make the difference between a product passing or failing the exposure test. BIBLIOGRAPHY J. Bellamy, Digital Telephony, 2nd ed., New York: Wiley, 1991. M. Gallagher and R. Snyder, Mobile Telecommunications Networking with IS-41, New York: McGraw-Hill, 1997. L. Knauer, R. Machtley, and T. Lynch, Telecommunications Act Handbook, Rockville, MD: Government Institutes, 1996. B. Kumar, Broadband Communications, New York: McGraw-Hill, 1994. W. Lee, Mobile Communications Design Fundamentals, 2nd ed., New York: Wiley, 1993. A. Mehrotra, Cellular Performance Engineering, Norwood, MA: Artech House, 1994. A. Mehrotra, Cellular Radio: Analog and Digital Systems, Norwood, MA: Artech House, 1994. A. Mehrotra, GSM System Engineering, Boston: Artech House, 1997. C. W. Mines, Policy development for cellular telephone service in the United States and United Kingdom Program on Information Resources Policy, Harvard University, September 1993. M. Mouly and M.-B. Pautet, The GSM system for mobile communications, M. Mouly et Marie-B. Pautet, 4, rue Elisee Reclus, F-91120 Palaiseau, France. T. Rappaport, Wireless Communications: Principles and Practice, Upper Saddle River, NJ: Prentice-Hall, 1996. S. Redl, M. Weber, and M. Oliphant, An Introduction to GSM, Boston: Artech House, 1995. W. Sapronov, Telecommunications and the Law—An Anthology, Rockville, MD: Computer Science Press, 1988. B. Sklar, Digital Communications, 2nd ed., New York: Van Nostrand Reinhold, 1991. H. Taub and D. Schilling, Principles of Communications Systems, New York: McGraw-Hill, 1971. www.ansi.org/default.htm www.arib.or.jp/ www.atis.org/ www.cdg.org/ www.ctia.org/ www.etsi.fr/ www.iso.ch/ www.itu.ch/ www.pcia.com/ www.tiaonline.org/site.html www.ttc.or.jp/index.html www.uwcc.org/

GIRISH PATEL NORTEL

MOBILE USER LOCATING. See PAGING COMMUNICATION FOR LOCATING MOBILE USERS.

MOBILITY. See ELECTRON AND HOLE MOBILITY IN SEMICONDUCTOR DEVICES.

MODEL CHECKING. See TEMPORAL LOGIC.

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELECTRICAL...NICS%20ENGINEERING/62.%20Vehicular%20Technology/W7703.htm

}{{}}

●

HOME ●

ABOUT US ●

CONTACT US ●

HELP

Home / Engineering / Electrical and Electronics Engineering

Wiley Encyclopedia of Electrical and Electronics Engineering Multiple Access Mobile Communications Standard Article Michael Moher1 1Communications Research Centre, Ottawa, Canada Copyright © 1999 by John Wiley & Sons, Inc. All rights reserved. DOI: 10.1002/047134608X.W7703 Article Online Posting Date: December 27, 1999 Abstract | Full Text: HTML PDF (217K)

●

●

● ●

Recommend to Your Librarian Save title to My Profile Email this page Print this page

Browse this title ●

Search this title Enter words or phrases

Abstract The sections in this article are Basic Techniques Summary Mobile Radio Spectrum Generalized Multiple Access Example Multiple Access Systems Interference Cancellation About Wiley InterScience | About Wiley | Privacy | Terms & Conditions Copyright © 1999-2008John Wiley & Sons, Inc. All Rights Reserved.

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELE...ENGINEERING/62.%20Vehicular%20Technology/W7703.htm15.06.2008 12:45:43

❍

❍ ❍

Advanced Product Search Search All Content Acronym Finder

650

MULTIPLE ACCESS MOBILE COMMUNICATIONS

MULTIPLE ACCESS MOBILE COMMUNICATIONS Multiplexing and multiple access refer to the sharing of a physical medium, often the radio spectrum, among different signals or users. When all signals access the medium through a common access point and can easily be coordinated, this is usually referred to as multiplexing. When the signals access the medium from different physical locations, this is usually referred to as multiple access. The key shared resources are time and radio frequency. There are several possible physical configurations for the multiple access channel or system. Figure 1(a) shows a centralized network where every user communicates with a central access node (or base station) and vice versa. Signals are multiplexed in the forward direction from the access node to the mobile terminal, and a multiple access strategy is used in the return direction. The access node could provide a connection to a wired network such as the public switched telephone network (PSTN) or it could be a private dispatch office. When there are a number of access points in the system, then the access nodes become an important resource that must be shared. This introduces a third dimension to the multiple access problem, in addition to time and frequency. Figure 1(b) shows a decentralized network where each user can communicate directly with the other users. In this case, a multiple access strategy, often the same one, is required for both transmitting and receiving. In the following, we will concentrate mainly on centralized networks, although many of the ideas carry over to decentralized networks. There are a number of similarities between the multiple access issues for fixed and mobile wireless systems. The main difference with wireless mobile systems is the time-varying nature of the communications channel. In mobile communications, multipath fading is the time-varying amplitude of the received signal resulting from constructive and destructive interference that is caused by receiving the same signal from multiple reflected paths. Shadowing is signal attenuation or

Mobile Mobile

Base station

Mobile (a)

(b)

Figure 1. The (a) centralized and (b) decentralized networks are two examples of mobile multiple access channels.

J. Webster (ed.), Wiley Encyclopedia of Electrical and Electronics Engineering. Copyright # 1999 John Wiley & Sons, Inc.

blockage due to terrain, buildings, or vegetation. Both are important propagation effects. These propagation effects make single user communications more difficult and result in significant power differentials between users. Consequently, there is greater potential for harmful interference. A significant part of the design of a successful multiple access strategy is controlling the mutual interference between users to an acceptable level. The major multiple access issues that are unique to mobile applications are 1. Multipath Fading and Shadowing. These problems are usually addressed through the modulation, coding, and antenna strategy. The solutions are usually some form of diversity—sending signals by multiple paths—and are often a combination of time, frequency, and space diversity. Time diversity is spreading information in time, usually through use of a forward error correction code and interleaving. Frequency diversity can be obtained by sending the same signal at multiple frequencies or spreading the signal over a large bandwidth. Space diversity is obtained by sending the same information over multiple radio links. 2. The Near-Far Problem. Received signal strength naturally decreases as a function of the distance between the transmitter and receiver. In a mobile environment, the rate of decrease is faster because of shadowing, and it implies potentially large differences in the received signal strength of signals from transmitters at different distances. The traditional solution is frequency separation (guard bands) between adjacent channel users and spatial separation between cochannel users. Recent systems have resorted to dynamic power control techniques to relax these separation requirements. 3. Synchronization and Tracking. Spectral efficiency, the average number of user transmissions per unit spectrum on a system-wide basis, can often be improved if there is some form of synchronization between users. In a mobile environment, acquiring and maintaining synchronization is a challenging and dynamic problem. 4. Handover and Paging. When there are multiple access points in a system, it is often necessary during a mobile communication session to switch between one access point and another. This handover requires that an access node keep sufficient resources to handle calls that are in progress at other access points, and which could potentially be handed over. Paging refers to the problem of locating a user to receive a call with a minimum of radio resources, when there are multiple access points in a system. These are important secondary issues in a cellular system with mobile users. The object of an efficient multiple access strategy is to maximize the number of users that can simultaneously access the system. In a digital system, this spectral efficiency is often measured in bits/sec/Hz/user. Because the key constraint limiting the number of users is mutual interference, a design goal is to minimize interference between users, or in a sense, to make users as orthogonal as possible.

Frequency

MULTIPLE ACCESS MOBILE COMMUNICATIONS

651

User 3 Guard band User 2 Guard band User 1

Time Figure 2. With FDMA, each user is given a dedicated frequency band in the time–frequency plane.

BASIC TECHNIQUES Frequency Division Multiple Access The traditional multiple access strategy is frequency division multiple access (FDMA). With this approach, the time-frequency plane is divided into a number of channels as illustrated in Fig. 2. Each user is assigned a distinct channel and is thus frequency ‘‘orthogonal’’ to the other users. This orthogonality is not perfect. Implementation limitations mean that the out-of-band transmissions of any user are nonzero. Causes of these out-of-band transmissions can be poor frequency control, poor transmit filtering, and amplifier nonlinearities. There is usually a requirement to limit these out-of-band transmissions to a specified level. This level will depend on the application and is often determined by the potential power differences between adjacent channel users as received at the base station. In mobile radio applications where one user can be significantly closer to the base station than another, these power differences can be as much as 80 dB or more. This is referred to as the near–far problem. This adjacent channel interference determines how closely the channels can be spaced, and ultimately the spectral efficiency. FDMA has the advantage that it is a relatively simple system. If all transmitters meet the out-of-band transmission limits then coordination of the users is simply a matter of assigning frequencies. In the simplest scenario, these frequencies are assigned on a fixed basis, but in many systems they are assigned on a dynamic basis when the user requires service. In the latter case, in addition to the traffic channels shown in Fig. 2, the system will need signaling channels to handle channel requests and channel assignments. Time Division Multiple Access With time division multiple access (TDMA), the time-frequency plane is divided into time slots as shown in Fig. 3. Each user is assigned different slots, often on a periodic basis, during which transmissions are allowed. This approach reduces the spectrum wasted because of the guard bands required with FDMA, since frequency errors are typically a smaller fraction of the transmitted bandwidth. However, it introduces the need for time synchronization between users.

652

MULTIPLE ACCESS MOBILE COMMUNICATIONS

User 1

Time Figure 3. With TDMA, each user is given dedicated, often periodic, time slots in the time–frequency plane.

Frequency

In a centralized network, coarse time synchronization can usually be obtained from the forward link, either explicitly from a transmitted timing reference or implicitly from the multiplexed signal. However, fine time synchronization is often necessary to compensate for the different distances of the users from the access node. This fine synchronization is often done through a feedback loop between the access node and each mobile unit. The use of global timing sources such as the Global Positioning Satellite (GPS) system can sometimes simplify this. Synchronization is never perfect in practice, and some guard times must be left to allow for timing errors between user transmissions. The spectral efficiency of the system depends on the ratio between these guard times and the length of a time slot (transmission burst). In mobile applications, timing errors are compounded because the terminal’s distance to the base station varies as the terminal moves. Thus, a feedback mechanism must be implemented to track the timing and insure synchronization is maintained. A disadvantage of TDMA, relative to FDMA, is that it requires higher peak powers from the transmitter, as the instantaneous data rate is higher to achieve the same average throughput.

Figure 4. With CDMA, users are not dedicated time and frequency slots in the time– frequency plane. (a) With FHSS, users are independently assigned random time and frequency slots and packet collisions are possible. (b) With DSSS, users are assigned modulating waveforms that span the time– frequency plane, but have low cross-correlations with other users.

With code division multiple access (CDMA) users are assigned unique codes to access the time-frequency plane, which produces low correlation with the signals of other users, that is, minimal average interference. There are two major variants: frequency-hopped spread spectrum (FHSS) and direct sequence spread spectrum (DSSS). These techniques were originally developed for military communications because of their low probability of interception, but, they have since found commercial application. With FHSS, the conceptual approach is to divide the time-frequency plane into time and frequency slots as shown in Fig. 4(a). Each user is given a distinct pseudorandom sequence that defines which time and frequency slots to use. This sequence is known by the receiving terminal but not necessarily by other users. Given a large number of frequency slots and short time slots, the probability of users colliding is low, depending upon the number of users, and the effects of collisions can often be compensated by error correction coding. With DSSS, each terminal uses a distinct modulating waveform derived from a pseudorandom sequence of bits. These modulating waveforms are approximately orthogonal, that is, they have low cross correlation with each other. These modulating waveforms or spreading codes span the allocated frequency band, as shown in Fig. 4(b). A conventional DSSS receiver correlates the received signal, which is the sum of all user signals, with the modulating waveform of the desired terminal. The desired signal produces a strong correlation, while the other signals produce weak correlations. The interference caused by other terminals because of their imperfect orthogonality can often be approximated as Gaussian noise. As with FDMA, there is a serious near–far problem with CDMA, and for mobile radio applications power control is often a requirement. This implies a feedback loop between the base station and the mobile to dynamically adjust mobile transmit power to obtain an acceptable level at the base station receiver. In their simplest form, the capacity of FHSS and DSSS systems are typically quite low, compared to FDMA and TDMA, without some coordination between users. In CDMA, this coordination takes the form of power control. While FHSS is relatively insensitive to power level variations, capacity can be significantly improved with some synchroniza-

Frequency

User 3

Guard time

User 2

Guard time

User 1

Guard time

Frequency

Code Division Multiple Access

User 3

User 2

User 1

User 1

User 1

User 2

User 3

User 3

User 3 User 2 User 1

User 2

Time (a)

Time (b)

MULTIPLE ACCESS MOBILE COMMUNICATIONS

90°

180°

0°

Sector 2 Sector 1 Sector 3

270 ° (a)

(b)

Figure 5. SDMA uses directional antennas to subdivide the service area: (a) polar plot of antenna gain versus azimuth angle, and (b) division of service area into sectors with 120⬚ antennas located at center of service area.

tion between users. In practice, both approaches rely heavily on forward error correction (FEC) coding to reduce the effect of multiple access interference. Spatial Division Multiple Access Spatial division multiple access (SDMA) is a technique that can be overlaid on any of the previous time and frequency sharing techniques to allow sharing in space. This is one of the primary techniques for allowing frequency reuse. In its simplest form, systems are allowed to reuse the same frequency by physically separating the service areas where a particular frequency is used, so that the natural attenuation of signal strength with distance insures that interference is reduced to minimal levels. An example of this is commercial AM and FM broadcast radio, where frequencies are only reassigned with separations of hundreds of kilometers. It is possible to more fully exploit the spatial dimension in a wireless communication system by equipping the access points of a wireless network with directional antennas. A directional antenna, with a gain versus azimuthal angle characteristic such as shown in Fig. 5(a), can increase the base station range and improve coverage. Due to the directivity of the antenna, interference may be reduced, resulting in improved performance. The capacity gains that result depend upon the modulation technique. With narrowband modulations such as used with FDMA, the interference is often not reduced enough to allow reuse of the same frequencies in adjacent antenna beams. Thus, with FDMA and a three-sector system, as shown in Fig. 5(b), there would be improved performance but no increase in capacity. With greater sectorization, one can reuse frequencies and increase capacity. With the wideband modulations typical of CDMA, one can achieve nearly full spectral reuse with as few as three sectors. A second key to maximizing frequency reuse is to limit the transmitted power to the minimum required. With mobile transmitters this implies that power control is an important factor in all multiple access strategies, and not just with CDMA. SUMMARY In practice, a combination of techniques is often used. Two common combinations are FDMA/TDMA/SDMA and FDMA/ CDMA/SDMA. The choice of technique is a tradeoff between

653

economics and difficulties associated with the side issues. The economics depend on system capacity and expected fiscal return. Potential side issues include growth options, available spectrum, signaling demands, integration of different services having potentially different data rates, and performance requirements. The capacity or spectral efficiency of a wireless system depends on the combination of the modulation and coding strategy of the individual user and the choice of multiple access strategy. The choice of modulation and coding strategy will depend on the service parameters, such as data rates, tolerable delay, tolerable outages, performance requirements and complexity. Often what is spectrally efficiency from a single-user viewpoint is not spectrally efficient from a system level. MOBILE RADIO SPECTRUM Since the late 1980s, there has been a huge growth in wireless telecommunications, and a significant portion of this is mobile. However, the amount of useable radio spectrum is limited and regulated. Internationally, the radio spectrum is regulated by the International Telecommunications Union (ITU) based in Geneva, Switzerland. The recommendations of this international body are enforced by local authorities. The limited spectral resources has led to significant competition and auctions for spectral licenses in some countries. Some frequency bands allocated for mobile radio and the corresponding user population are listed in Table 1. The majority of current mobile radio systems are in frequency bands lower than 2 GHz, primarily because of limitations in mobile terminal technology. Although these limitations are disappearing, the demand for spectrum is still outstripping the supply, which emphasizes the importance of spectrally-efficient multiple access strategies. The dominant application to this point has been voice, until quite recently, using analog transmission techniques such as frequency modulation and single-sideband amplitude modulation. However, digital techniques (1) permit greater spectral efficiency; (2) allow applications such as fax and data; and (3) promise mobile multimedia for the future.

Table 1. Some Mobile Radio Bands and Example Applications Band 118–136 150–174 450–470 825–870

Some Applications MHz MHz MHz MHz

902–928 MHz 890–960 MHz 1452–1492 MHz 1525–1559 MHz 1610–1660 MHz 1930–1980 MHz 1980–2010 MHz 2170–2200 MHz 2400–2480 MHz

Aeronautical safety radio Public safety radio European FM cellular telephony North American FM cellular and first generation digital ISM band for low-power unlicensed spreadspectrum users (North America) GSM cellular telephony Digital audio broadcasting Mobile satellite downlink Mobile satellite uplink PCS cellular telephones Mobile satellite telephony downlink Mobile satellite telephony uplink ISM band for low-power unlicensed spreadspectrum users (North America)

654

MULTIPLE ACCESS MOBILE COMMUNICATIONS

GENERALIZED MULTIPLE ACCESS Proakis (1) shows that a real valued signal s(t) with frequency content concentrated in a band of frequencies near a frequency f k can be expressed as sk (t) = ak (t) cos[2π f k + θk (t)]

(1)

where ak(t) represents the amplitude and ␪k(t) represents the phase. We use the subscript k to indicate that this is the signal of the kth terminal in a multiple access system with K users. The received multiple access signal can then be represented as the sum of Eq. (1) over all users

X

X P

K

r(t) =

transmit during a preassigned time interval. Unlike FDMA, the center frequency is identical for all users f k ⫽ f j all k and j. With mobile users, a feedback mechanism must be included to track delay variations and maintain timing sync. The delay is given by ␶k ⫽ rk /c, where rk is the distance between the transmitter and the receiver. For a CDMA strategy, all users have a common frequency band, and the modulation depends upon whether DSSS or FHSS is used. With DSSS, a linear modulation is often used and Simon et al. (3) show that the pulse shaping in Eq. (3) for a DSSS binary phase-shift keyed (BPSK) signal can be expressed as

wk sk (t − τk ) + n(t)

(2)

pk (t − iT ) =

ak ( j)pc (t − jTc − iT )

(4)

j=1

k=1

where ␶k are the relative delays, and wk are the propagation losses of the different users, respectively. The factor wk can also include a relative phase rotation for each user. The term n(t) represents additive white Gaussian noise. The delays and propagation losses are position dependent, and thus for a mobile user may also be time dependent. Equation (2) represents what is referred to as a flat fading, or frequency-independent, model of a mobile channel (2). Depending upon the bandwidth of the signal and nature of channel, other models may be appropriate (2). Although multiple-access questions apply to the transmission of both analog and digital information, in this article we will focus on digital systems. To emphasize this, we will often represent each user’s signal as sk(bk,t) where bk represents the information bits in the message of the kth user. A message could represent a data packet, for example, or a segment of digitally encoded speech. For digital signals, a linear modulation strategy (1) is often used and the kth user’s signal can then be expressed as

X

where pc(t) corresponds to a rectangular pulse of width Tc, P is the number of these pulses (chips) per symbol, and 兵ak其 is the spreading sequence. The spreading sequence is usually synchronous with the bit sequence 兵bk其, that is, it repeats every symbol period, but it is sometimes overlaid with a further randomizing code that has a much longer period. The bandwidth of the resulting signal is approximately 1/Tc. Considerable research has been invested in determining the optimum spreading or pseudo-noise (PN) sequence. Maximal-length sequences and Gold codes (3) are two common examples. Taking a basic modulation technique and changing the carrier frequency in some pseudorandom manner is the frequency-hopping approach to generating a spread spectrum signal. The modulation technique often used is M-ary frequency-shift keying (M-FSK) (1). When binary frequency-shift keying (2-FSK) with FHSS, Simon et al. (3) show that the transmitted signal can be presented as sk (bk (i), t) = cos[2π ( f c + f k,i + bk (i) f )t + φk ] iT ≤ t < (i + 1)T

(5)

N

sk (bk , t) =

bk (i)pk (t − iT ) cos(2π f kt + φk )

(3)

i=1

where T is the symbol period, N is the number of transmitted symbols, bk ⫽ 兵bk(1), . . . bk(N)其 are the data symbols, and pk(t) is the time domain representation of the pulse shaping or filtering applied to the data. For an FDMA scheme, the modulation schemes of the users are typically identical, for example, with linear modulation the pulse shaping pk(t) is the same for all users, and only the center frequencies, f k ⫽ f min ⫹ (k ⫺ 1)⌬f, differ where ⌬f is the frequency spacing of channels. A critical issue for FDMA is guard bands required to limit adjacent channel interference (ACI). This depends upon the modulation and filtering strategies and nonlinearities present in the transmit chain. It also depends upon the frequency accuracy of the transmitter oscillator and Doppler-induced frequency errors due to terminal motion. The Doppler shift f D is given by f D ⫽ f RFv/c where f RF is the radio frequency, v is the speed of the terminal in the direction of the receiver, and c is the speed of light. This frequency shift, f D, can be greater than a kilohertz for an aircraft terminal operating at 1.5 GHz. For a TDMA strategy, user modulation schemes are also typically identical, except that each user is only allowed to

where f k,i is a sequence of randomly chosen frequencies, and ⌬f is the modulation frequency. Each user uses a different sequence of hop frequencies 兵f k,i其 that range over the allocated bandwidth and are known at the receiver. Frequency-hopping strategies are classified into fast- and slow-hopping, depending upon how long the transmitter dwells on a particular hop frequency. With the former, only one or a fraction of one symbol is transmitted each dwell time. With the latter, multiple symbols are transmitted at each frequency. Slow-hopping strategies may use FSK but are often combined with a TDMA strategy and use M-ary phase-shift keyed (PSK) modulation. This combination provides a degree of frequency diversity that is not normally available with a TDMA strategy. Conventional Detectors Whalen (4) shows that the optimum linear single-user receiver for an additive white Gaussian noise channel is a bank of correlators matched to all possible transmitted sequences for the kth user. That is, for the kth user, one computes the correlation values Lj =

R

r(t)sk (b kj , t) dt

(6)

Z

yk (i) =

∞

−∞

r(t)pk (t − iT − τk ) cos(2π f c t + φk ) dt

(7)

That is, the optimum receiver for an additive white Gaussian noise channel filters the received signal with a filter matched to the transmitted pulse shape. This receiver can be applied to all the multiple access systems described thus far. Implicit in Eq. (7) is the assumption that the modulating waveform of each user is known at the receiver, and that the receiver locks to the signaling interval and phase of each active user. In a mobile multiuser environment, constructing this type of a coherent receiver is a challenging problem that is not dealt with here. For binary signaling, a conventional detector without FEC coding simply takes the sign of the bits at the output of the matched filter, that is, bk(i) ⫽ 1 if yk(i) ⬎ 0 and ⫺1 otherwise. In practice, most current receivers are an approximation of the optimum receiver. This is what we will assume in the remainder. With this type of receiver, the interference due to another user can be represented as I j (i) =

Z

∞

−∞

w j s j pk (t − iT − τk )

cos(2π f ct + φk ) dt

(8)

Multiplexing for the Forward Link In the forward direction, base station to mobile, there is the opportunity to coordinate and synchronize users to minimize the multiple access interference. That is, one can provide access to the channel on a contention-free basis. There are a number of ways to multiplex several users onto a single channel. Time-Division Multiplexing. In packet switched networks, a common method is time-division multiplexing (TDM) of the packets for each user. That is, the bits of each packet are transmitted sequentially over the same channel with no need for guard times between the packets. Depending upon the regularity of the users, each packet may include addressing information, and each terminal monitors all packets to deter-

...

Addr.

Packet 2

Packet K

655

...

Figure 6. Time-division multiplexing of different user packets eliminates the need for guard times.

mine those that are addressed to it, as shown in Fig. 6. If packets are regular or periodic, then individual addressing may be forgone in favor of a look-up table that is broadcast at less frequent intervals. This approach does not require synchronization or guard time overhead to be associated with each packet because it is a continuous data stream. Some regular synchronization information is required, however, to speed initial acquisition and aid reacquisition, should the mobile terminal lose the signal. Frequency-Division Multiplexing. For analog channels, frequency-division multiplexing (FDM) is a common traditional approach, where each mobile is assigned a unique forward frequency for the duration of the call. This has the advantage of simplicity, and that the mobile can use the dedicated channel for just about any application that fits in the defined channel bandwidth. For this reason it has also been used in many digital and analog transmission systems. Orthogonal Frequency-Division Multiplexing. An enhancement to basic FDM is orthogonal frequency-division multiplexing (OFDM). Weinstein (5) shows that digitally modulated carriers can be significantly overlapped without harmful interference, as long as the carrier spacing was equivalent to the symbol period, eliminating the need for guard bands. That is, in a system with K carriers, data is transmitted K symbols at a time, which can be represented as

X K /2

s(t) = or its frequency domain equivalent. This interference is classified as cochannel or adjacent channel interference, depending upon whether most of the spectrum of sj(t) overlaps that of sk(t) or not. The sum of the interference from all the other users is often referred to as the multiple access interference (MAI).

Packet 2

Addr.

for all possible data sequences 兵bkj : j ⫽ 1 . . . 2N其 of the kth user, assuming binary modulation, and chooses the one with the largest Lj value. This optimum receiver has a complexity that is exponential in the sequence length. In practice, when specialized to a particular modulation, the complexity of the receiver can often be reduced to a linear function of the sequence length. Proakis (1) describes specific modulation techniques and the corresponding receivers. There it is shown that, for linear modulation, see Eq. (3), a sufficient statistic for optimum detection of an individual data symbol is given by

Addr.

MULTIPLE ACCESS MOBILE COMMUNICATIONS

bk (i)e−2π j( f c +k f u )t

iTu ≤ t < (i + 1)Tu

(9)

k=−K /2+1

where f u ⫽ 1/Tu. This corresponds to K different users or some other combination of data from fewer users. There are several advantages to this approach. The first is that the modulator and demodulator can be implemented as a Discrete Fourier Transform (DFT) when pulse-shaping is rectangular. With this approach a set of K data, often referred to as frequencydomain symbols, are transformed by the inverse DFT (6) to form a set of time domain symbols Bn(i) to be transmitted over the channel (7),

X K2

Bn (i) =

bk (i)e j2π in/K

n = 0, . . ., K − 1

(10)

k=−k 2 +1

and these samples are transmitted sequentially in the interval iTu ⱕ t ⬍ (i ⫹ 1)Tu on a carrier of frequency f c. This corresponds to a sampled version of Eq. (9) with K samples per symbol period. The demodulator is implemented using a DFT. This has a fast implementation when K is a power of 2, or a product of prime powers, that is known as the Fast Fourier Transform (FFT). From a transmission viewpoint, the multicarrier approach is advantageous for higher data rates in a fading environment where there may be frequency-selective

656

MULTIPLE ACCESS MOBILE COMMUNICATIONS

fading. With the multicarrier rates, the effective symbol rate is reduced by a factor equivalent to the number of carriers. Since any time dispersion—multipath where the relative delays of the different paths are significant relative to the symbol interval—will cause intersymbol interference (ISI) (1), a short guard time is usually added to each symbol period to avoid ISI due to multipath. If the resulting guard time is longer than the expected time dispersion of the channel (2), there is not a need for an equalizer other than for channel gain and phase compensation. A disadvantage of OFDM is that it can be sensitive to frequency errors. A second disadvantage is that the transmitted signal is not a constant envelope and requires a linear transmitter. This is often not a concern for transmissions from a base station. Code-Division Multiplexing. An alternative approach to multiplexing the data from several users is code division multiplexing (CDM). Because one can make the sequences of the different users synchronous at the transmitter, it is possible to choose perfectly orthogonal spreading codes for the different users. When the spreading factor is a power of 2, a common choice for the spreading codes are the rows of the corresponding Hadamard matrix, which is given recursively by


H1 =

1 1






1 H1 H2 = H1 −1






H p−1 H1 . . . Hp = H p−1 −H1



H p−1 −H p−1 (11)

The rows of these matrices are often referred to as Walsh functions (Hp provides 2p Walsh functions), and they would be used as the 兵ak( j), j ⫽ 1 . . . 2p其 in Eq. (5). Similar to OFDM, this approach spreads each user over the available bandwidth and has potential frequency diversity advantages in a frequency selective fading environment. Time dispersion of the signal due to multipath is a potential problem with wideband signals, such as, non-flat fading, and Proakis (1) shows that a RAKE receiver can be employed to recover the energy in timedispersed multipath channels. Multiple Access for the Return Link The number of issues on the return link is greater than on the forward link, and they are difficult to treat in isolation. In the following, we present a number of multiple access protocols and show how they deal with these difficulties. Packet switched networks, cellular systems, and spread spectrum systems have been selected as examples, but the techniques presented can be applied to a much wider variety of systems. Packet Switched Networks. The problem that packet switched networks attempt to solve is the sharing of packets of digital information between a number of different users who share a common channel. The classic sample of this is the ARPANET network (8), which consisted of a number of research institutions that were linked by a common satellite channel. One characteristic that is used advantageously in this system is that, because of the delay when transmitting over a geostationary satellite (approximately 240 ms), one can listen to one’s own transmission if the packet is sufficiently short. This gave rise to a number of packet-switching protocols that require increasing degrees of cooperation between the mobile units.

Pure ALOHA. The first such protocol has come to be known as pure ALOHA, in which users transmit any time they desire. If, after one propagation delay, they hear their successful transmission, then they assume no collision with a packet from another user has occurred. Otherwise, a collision is assumed and the packet must be retransmitted. The pure ALOHA strategy is a form of uncoordinated TDMA. Let the packet transmission period be Tp and let ␩ denote the channel throughput or efficiency (average number of successful transmissions per transmission period Tp). Collisions between packets (cochannel interference) is the main source of degradation here. If the total channel traffic G, the average number of packets (initial plus retransmitted) offered per transmission period Tp, comes from an infinite population of users each with an independent Poisson distribution, then Kleinrock (8) shows η = Ge−2G

(12)

The maximum efficiency is 1/2e 앒 0.184 which occurs at G ⫽ 1/2. Slotted ALOHA. The second technique is known as slotted ALOHA, and is a more coordinated form of TDMA, where time is segmented into slots matching the packet length Tp (plus some guard time) and all users are required to confine their transmissions to slots. This confines a collision between packets to a slot and results in increased efficiency. Then, for K statistically equivalent users with total offered traffic, Kleinrock (8) shows that




η=G 1−

G K

K −1

−−−−→ Ge−G K →∞

(13)

Thus, the slotted protocol has double the maximum efficiency of the pure ALOHA system. To compute the average packet delay, let R represent the delay (in slots) before a user knows whether a transmission was successful. When a collision is detected, the packet is retransmitted, at random, in one of the M subsequent slots. Then, the average packet delay in slots is approximately (8)
 M−1 M−1 T∼ − = eG R + 1 + 2 2

(14)

when M ⬎ 10. The delay increases exponentially as the loading on the channel increases. There is a fundamental tradeoff between throughput and delay for this strategy and it is often operated at levels much below the maximum throughput in order to have reasonable delays. This is also necessary to maintain system stability, for if the system is operated too close to the optimum level, statistical variation of the traffic can lead to excessive collisions with offered traffic and delays approaching infinity. Carrier Sense Multiple Access (CSMA). In terrestrial systems, the propagation delay is often so short that one cannot listen to one’s own transmission. However, short propagation delay does allow one to listen to determine if the channel is occupied before transmitting, which gives rise to carrier sense multiple access (CSMA). With this protocol, if the channel is in use, the terminal postpones its transmission until the channel is sensed to be idle. If, through lack of positive acknowledgment, the mobile determines its transmission was

MULTIPLE ACCESS MOBILE COMMUNICATIONS

657

unsuccessful, then it reschedules the retransmission according to a randomly distributed transmission delay, and repeats the protocol. If all packets are of the same length with packet transmission time P, and the one-way propagation delay d/2 is identical for all source-destination pairs, then the throughput is given by (8) η=

Ge−aG G(1 + 2a) + e−aG

(15)

where a ⫽ d/(2P). There are many variations on this basic approach. A common approach is p-persistent CSMA where a mobile transmits only with probability p, if it has a packet ready and detects the channel as idle. These latter approaches offer the possibility of greater throughputs with lower delays (8). Both persistent and nonpersistent approaches can be used with slotted or unslotted formats. Packet Reservation Multiple Access (PRMA). PRMA is a packet protocol for the return link of a centralized network based on the reservation ALOHA protocol. Reservation Aloha is a slotted ALOHA system with both reserved and unreserved slots, together with a reservation system that assigns slots on a dynamic basis. PRMA adds the cyclical frame structure of TDMA to reservation ALOHA in a manner that allows each TDMA slot to carry either voice or data, where voice is given priority. With PRMA, time is divided into frames of a fixed length, and each frame is divided into a number of slots. The frame length typically equals the frame period of the speech encoding algorithm being used in each terminal. The slot sizes are designed to handle one speech frame (voice packet) at the given transmission rate. It is assumed that the base station can organize the forward traffic on a contention free basis. It is assumed that the network is small physically so that propagation delays are very small, and it is possible to acknowledge a burst in the same time slot that it is transmitted, or at least within one time slot. All terminals are assumed to use voice detection algorithms and to transmit only when voice is present. Each terminal keeps track of those frames that are reserved. The first packet of a voice spurt is transmitted in any unreserved slot, much like the slotted ALOHA protocol. If the voice packet transmission is successful, it results in that slot being reserved for that user in all future frames. A reservation is canceled by not transmitting during a reserved slot. If the voice packet transmission is unsuccessful, retransmission is tried in the next unreserved slot with probability q. Voice packets are only kept for up to one frame period, at which point they are discarded. Nonperiodic data packets are integrated into the system by simply using the slotted ALOHA protocol. If successful, they do not result in a periodic reservation, and if unsuccessful, the data packet is retransmitted with probability p in the next unreserved slot. If the data packets are not delay sensitive, they need not be discarded as they age. Goodman et al. (9) show that one can achieve quite high channel efficiencies that takes advantage of voice activation with acceptable dropped packet rates (voice quality) with PRMA. The one limitation is that PRMA is only applicable to local area systems, since it requires immediate acknowledgments. Cellular Networks. In cellular systems, frequencies are often assigned in a hexagonal pattern as shown in Fig. 7 with users in each cell communicating with a base station at the

i=2 j=1

R D

Figure 7. The one-in-seven hexagonal pattern with a cell radius R and reuse distance D is one frequency reuse scheme for cellular systems.

centre of the cell. The term cellular is usually applied to terrestrial systems, but similar considerations apply to multibeam satellites. For non-geostationary satellite systems, there is an added difficulty in that the cells move with respect to the earth as the satellite moves. A hexagonal cell shape is often used because of its close approximation to the circle and its ease of analysis. With FDMA strategies, frequencies are not reused in each cell due to excessive co-channel. With a hexagon geometry, the reuse pattern can be defined relative to a given reference cell (10): Move i cells along any chain of hexagons, turn counterclockwise 60 degrees; move j cells along the chain that lies on this new heading, as shown in Fig. 6. The jth cell and the reference cell are cochannel cells. With this hexagonal geometry, the cells form natural clusters around the reference cell in the centre and each of its cochannels cells. The number of cells per cluster is given by (10) N = i2 + ij + j 2

(16)

The ratio of D, the distance between the centres of nearest neighbouring cochannel cells, to R, the cell radius, is the normalized reuse distance, and is given by D √ = 3N R

(17)

From Eq. (16) this allows reuse factors of one-in-three, onein-four, one-in-seven, one-in-nine, one-in-twelve, and so on. In terrestrial systems, cochannel isolation is determined by propagation losses, and consequently the reuse distance is a function of the propagation loss. For satellite systems, the reuse distance is a function of spotbeam isolation. In the above expression, the reuse distance D is a function of both the cell radius and N the number of cells per cluster. For terrestrial propagation, the mean propagation loss between a transmitter and a receiver can be approximated by (2) Pr =

Po rn

(18)

658

MULTIPLE ACCESS MOBILE COMMUNICATIONS

where Pr is the received power, Po is the power at a reference distance, r is the transmitter-receiver separation, and n is a parameter that can range from two to five depending upon the propagation environment. The value of n ⫽ 2 corresponds to free space loss, while n ⫽ 5 approximates a dense urban environment. The reference power also depends upon a number of factors such as the height and gain of the transmitting and receiving antennas. For users with similar modulation, Rappaport (11) shows that the mean carrier-to-interference ratio can then be approximated by

C = I

X

−n d

rd

k = d

−n k

rk

(19)

where d corresponds to the desired user, and k ⬆ d corresponds to interfering cochannel users. Assuming that the six closest interferers in a cellular system cause almost all of the interference, and that these interferers are at the centre of their cells while the desired user is at the edge of its cell, then one has R−n ≥ 6D−n


C  I

(21)

The resulting efficiency of an FDMA cellular strategy is given by Rk N(B + Bg )

(22)

where Rk is the information of the kth user, B is the channel bandwidth, Bg is the guard band required between channels to reduce adjacent channel interference to acceptable levels, and 1/N is the frequency reuse factor required to reduce cochannel interference to acceptable levels. Spread-Spectrum Systems. DSSS Systems. In a DSSS system the cochannel interference given by Eq. (8) is treated as equivalent to Gaussian noise of the equivalent power with a flat power spectral density I0. The exception to this rule is when the channels are synchronized or partly synchronized, and are using orthogonal spreading codes. In the latter case, the multiple access noise is zero or close to it, assuming an ideal implementation. In the simplest case, all users within a cell are assumed to be power controlled such that they are received at similar level at the base station and all cells are equally loaded. Under these conditions, Viterbi (12) shows that the intracell in-

Pr W

(23)

where Pr is the received power of each user, and W is the bandwidth over which each user is spread. In a multicell scenario, one must also consider the intercell interference— interference from users in other cells. In terrestrial systems, the intercell interference is determined by the propagation losses; in a satellite system it is determined by the spotbeam rolloff characteristics. Viterbi represents the intercell interference as a factor f times the intracell interference. The total noise that the desired signal must contend with is N0 + I0 = N0 + (1 + f )(K − 1)

E b Rk W

(24)

where N0 is the thermal noise density, Pr ⫽ EbRk, Eb is the received energy per bit, and Rk is the bit rate. The interference constraint faced in this system is that the signal-to-noise ratio at the receiver

(20)


  2/n C 1 6 3 I min

η=

Ii = (K − 1)

Eb ≤ N0 + I0

min

where n is the common propagation loss, and (C/I)min is the minimum tolerable carrier interference ratio for the system. Typical values of the latter are 18 dB for an analog FM signal and 12 dB for a narrowband digital system, but exact values depend on the modulation and coding strategy and quality of service required. It follows from Eqs. (17) and (20) that the frequency reuse factor is lower bounded by N≥

terference—interference from users in the same cell—density can be written as


E  b

N0

(25) min

For this system, the frequency reuse is given by KR/W, so approximating (K ⫺ 1) by K in Eq. (24) and substituting the result in Eq. (25) one obtains the upper bound η=

1 KRk ≤ W 1+ f




1 1 − (Eb /N0 )min Eb /N0

 (26)

This efficiency can be increased by considering voice activation and sector reuse. In a typical telephone conversation each user speaks approximately 40% of the time, so a signal that is transmitted only when voice is active reduces interference by a factor, GV 앒 2.5, on average. In CDMA, unlike FDMA, frequency bands can be reused in adjacent antenna sectors. In practice, sectored antennas covering 120⬚ in azimuth are often used in high-traffic areas so the corresponding gain is spectral reuse is GA 앒 3. In some terrestrial situations, the background noise can be negligible relative to the multiple access noise. Considering all these factors, Viterbi (12) shows that Eq. (26) can be approximated by η≈

1 GA GV 1 + f (Eb /N0 )min

(27)

For a cellular telephony values for ␩ approaching one have been suggested (12). The spreading codes in DSSS can be implemented as a combination of pseudorandom sequences and forward error correction coding. One of the advantages of CDMA is that forward error correction coding can be introduced to reduce the required Eb /(N0 ⫹ I0) and thus increase system capacity with no compromises other than an increase in detector complexity. In practice, to minimize transmit power, extend battery power, and minimize health concerns with handheld transmitters, operation is typically at low Eb /N0 as well. As an example, to achieve a bit error rate of 10⫺3 that is a typical re-

MULTIPLE ACCESS MOBILE COMMUNICATIONS

659

quirement for vocoded speech, in an additive white Gaussian noise channel with a rate 1/2 constraint length 7 convolutional code requires an (Eb /N0)min of approximately 3 dB (1). A potential advantage of DSSS CDMA is high efficiency in cellular systems. From a modulation viewpoint, it has a nearly constant envelope, and with a RAKE receiver, one can take advantage of frequency diversity. The disadvantages are the power control requirements and that large bandwidths are required for even low-traffic areas. In addition, for higher data rates, the available spectrum is often insufficient to allow significant spreading. FHSS Systems. The probability of collisions between hops of different users is what determines the performance in a FHSS system. Let 애 represent the probability that there is a collision between any two or more users. For K independently hopping users using M-FSK, where Nt is the total number of frequency bins,

where R is the service area, and p(x) is the probability distribution of interferers over the service area. With 120⬚ sectors, out-of-sector users are significantly attenuated, but generally not enough to allow frequency reuse in an FDMA system. This improves performance because of reduced interference and provides better coverage because of higher gain antennas, but there is no effective capacity gain. With greater sectorization, one can achieve some frequency reuse with narrowband modulations. With CDMA, one can achieve significant spectral reuse even with three sectors.

K µ= M Nt

A widespread TDMA standard that is used for cellular telephony is the Global System for Mobile communications (GSM) (14). The channel transmission format consists of a superframe that is divided into frames, which are subdivided into slots that are assigned to users. A superframe of 6.12 s is divided into 1326 TDMA frames of 4.615 ms each. Each frame has eight time slots of 0.577 ms for 148 bits, plus a guard time equivalent to 8.25 bits. The resulting TDMA burst rate is 271 kbps, and there are eight full-rate users per channel. There are 125 duplex channels paired between the 890 to 915 MHz return link band and the 935 MHz to 960 MHz forward link band. This system uses constant-envelope partial-response Gaussian Minimum Shift-Keyed (GMSK) modulation with a channel spacing of 200 kHz. The controlled GMSK-induced ISI and the uncontrolled channel-induced ISI are removed by a channel equalizer at the receiver. The traffic channels use a r ⫽ 
, k ⫽ 5 convolutional code, but some control channels have greater error protection through the use of block codes. Since the GSM standard allows frequency hopping, each physical channel corresponds to a sequence of RF channels and time slots. Logical channels, listed in Table 2, are assigned to the physical channels in either a fixed or dynamic manner. The various control channels are defined in Table 2; BCCH, FCH, SCH, and CCH are transmitted on a single RF channel from each base station. These channels allow the terminal to acquire the system [first in frequency (FCH), then in time (SCH)] and then determine the current system configuration (BCCH). The remaining control channel (CCH) is used to notify the user of an incoming call or provide a channel assignment. The traffic channels (TCH) carry the voice/data, and each terminal is assigned one slot (and frequency) in 24

(28)

By itself, this would imply a relatively low spectral efficiency for reasonable performance. However, frequency-hopping systems usually include FEC encoding. Reed–Solomon encoding (13) is a common choice because it can be used to map M-FSK symbols directly into code symbols, and because it can very effectively use side information about whether a hop was jammed or not. A common frequency-hopping detector is an FFT followed by an energy detector, which will often indicate the presence of more than one tone when there is a collision on a hop. If all collided hops are known, a rate 1/2 Reed– Solomon code could effectively reduce a symbol error approaching 50% to negligible levels. In particular, the probability of a code word error with a (n, k) Reed–Solomon code for M-ary symbols, and knowing which symbols are jammed, is given by (13)

X




n

Pe =

j=n−k+1

n µ j (1 − µ)n− j j

(29)

If the jammed symbols are not known then the lower bound on the summation changes to (n ⫽ k)/2 ⫹ 1. Frequency-hopping systems have the advantage that they do not require power control. The disadvantage is that even with FEC encoding, the spectral efficiency is low. Other Multiple Access Techniques Spatial Division Multiple Access. The increasing demand for capacity in wireless systems traditionally translates into a demand for bandwidth. However, limited bandwidth has led to the consideration of other ways of increasing spectral efficiency, such as more efficient use of the spatial dimension by employing antenna arrays at the base stations. The primary benefit is a reduction of the multiple access interference. Let Pr(x) be the received power with an omnidirectional antenna of an interferer at position x, and let G(␾) be the gain of the directional station antenna as a function of azimuth angle ␾. Then the average MAI with a directional antenna is given by MAI =

Z

G(φ)Pr (x)p(x) dx R

(30)

EXAMPLE MULTIPLE ACCESS SYSTEMS There are a number of existing systems that illustrate the principles defined in the previous section. A TDMA Cellular Telephony System

Table 2. The Logical Channels into Which the Physical Frequency-Hopped TDMA Channels Are Divided BCCH FCH SCH CCH TCH SACCH FACCH RACH

broadcast channel frequency channel synchronization channel access grant/paging channels traffic channel slow associated control channel fast associated control channel random access channel

660

MULTIPLE ACCESS MOBILE COMMUNICATIONS

out of every 26 frames (in the half-rate mode each mobile is assigned one slot in 12 out of every 26 frames), while the remaining two frames are used for SACCH. During a call the base station continually monitors the mobile’s timing error and received power levels, and sends any corrections via the SACCH. The FACCH can carry the same information as the SACCH but is only used when there is a need for heavy duty signaling, such as in a cell handover. The FACCH obtains capacity by stealing frames from the TCH when required. In the return direction, one also has the TCH plus the random-access channel (RACH). The format of the RACH differs in that the slots are 235 ms long in order to accommodate initial timing errors of the users, and uses the slotted ALOHA protocol. Frequency reuse is similar to that described for FDMA system, with frequencies reused in cells of sufficient distance. Typical frequency reuse numbers are 21, 12, and 9 (in sectored systems). A CDMA Cellular Telephony Standard. A widespread CDMADSSS standard that is used for cellular telephony applications is IS-95 (15). To aid synchronization to the spreading sequence, each base station emits a unmodulated pilot PN sequence in the forward link to identify itself. The period of the PN sequence is 215 chips with a chip rate of 1.22288 MHz. Different offsets of the same PN sequence are used to identify different base stations. In all, 512 possible offsets (52 애s) are allowed. The terminal first acquires the strongest pilot sequence, which identifies the closest base station. It also allows the terminal to immediately acquire the sync channel that is a synchronized but modified (Walsh spread) version of the pilot sequence. The sync channel provides synchronization information to allow the mobile to listen to the paging channel for that base station and submit an access request (on the return link access channel). This allows call setup and then, knowing the appropriate spreading codes, forward and return traffic channels are initiated. All spreading codes are approximately synchronized to the forward link pilot signal as this reduces the search range in the receiver acquisition process. The pilot signal, sync channel, paging channel, and forward traffic all share the same 1.25 MHz frequency band. Similarly all access request channels and reverse traffic share the same return 1.25 MHz frequency band. The access request channel is an ALOHA channel with capture. That is, the receiver may be able to correctly demodulate two colliding bursts as long as their timing is not identical. Each base station is allowed to use the same pair of forward and return 1.25 MHz frequency bands. This allows the mobile to initiate handovers between base stations based on the received signal strength of their pilot signals. The return link is in the band from 825 MHz to 850 MHz, and the forward link is in the band from 870 MHz to 895 MHz. In the forward direction, the primary data rate for this system is 9.6 kbps with submultiples of this rate also implemented by reducing the transmission duty cycle proportionately. The data is rate 1/2 encoded with a constraint length 9 convolutional code to achieve a coded data rate of 19.2 kbps. The coded data is block interleaved to provide time diversity against fast fading. The data is then triply spread. The initial spreading is by a long PN code that has a period of 242 ⫺ 1 chips and has a chip rate of 1.2288 MHz. This long PN code is specific to the user. It is then spread by a Walsh sequence, also with a chip rate of 1.2288 MHz, and with 64 chips per

coded bit. There is a final spreading by the PN same code as used for the pilot signal for that base station, which is also at a chip rate of 1.2288 MHz. This final spreading is combined with a quaternary PSK (QPSK) modulator, the output of which is filtered before transmitting. The nominal transmit bandwidth is 1.25 MHz. In the return direction, the data rates are similar, but encoding and modulation differ. The data is rate 1/3 encoded with a constraint length 9 convolutional code to achieve a coded data rate of 28.8 kbps. The data is block interleaved to provide time diversity against fast fading. Then, six code symbols are modulated as one of 64 modulation symbols (Walsh functions) with an orthogonal modulator (1). This results in a Walsh chip rate of 307.2 kHz. The resulting signal is then spread by the long PN code for that user, each Walsh chip being spread by four PN chips. The signal is further spread into both the inphase and quadrature channels by the pilot sequence corresponding to the forward link. The resulting signal is offset-QPSK modulated and filtered. It is the multiple access noise from the mobile’s cell and the surrounding cells that limits the capacity of this system. To minimize the multiple access noise, all mobile units include both open-loop and closed-loop power control. This has the secondary benefit of extending battery life for handheld terminals. The system is designed such that ideally about 60% to 65% of the multiple access interference is intracell, approximately 36% comes from the surrounding six cells, and less than 4% comes from more remote cells (7). An OFDM Digital Audio Broadcasting System One digital audio broadcasting standard (16) for mobile, portable, and fixed receivers uses OFDM for broadcasting data and music, and so on, from a number of sources. There are four possible modes of operation in system. All four modes provide a channel rate of 3.072 megabits per second, but are implemented with different numbers of carriers and symbol rates. This allows the system to tradeoff robustness and coverage versus complexity. The nominal channel bandwidth in all three cases is 1.536 MHz. The Mode 1 transmission frame consists of 76 consecutive OFDM symbols, excluding the null period. The length of a mode 1 frame is Tf ⫽ 96 ms. Each OFDM symbol consists of a set of 1536 equally spaced carriers, with carrier spacing equal to 1/Tu, and where Tu ⫽ 1 ms. The data is convolutional encoded with puncturing to allow unequal error protection. The mother code is a rate 1/4 constraint length 7 convolutional code with interleaving to provide time diversity. The data is differentially encoded QPSK modulated before being placed on a carrier with rectangular pulse shaping. Each Mode 1 frame consists of: 1. A synchronization block consisting of the first two OFDM symbols, the null symbol, and the phase reference symbol. The null symbol is a transmission-off period of Tnull ⫽ 1.297 ms followed by a phase reference symbol that constitutes a reference for differential modulation of the next OFDM symbol. The phase reference symbol and all subsequent symbols are of length Ts ⫽ Tu ⫹ ⌬ ⫽ 1.246 ms. A null period of length ⌬ ⫽ 0.246 ms is inserted at the end of each symbol to reduce ISI resulting from delay spread of the channel. As a result

MULTIPLE ACCESS MOBILE COMMUNICATIONS

the receiver can be implemented without the need for an equalizer. 2. A fast information channel (FIC) made of four blocks of 256 bits carrying the information necessary to interpret the configuration of the main service channel (MSC). This information is rate 1/3 code, split into three blocks of 3072 channel bits, and transmitted on the first three OFDM symbols following the synch block. 3. An MSC is made up of a sequence of four common interleaved frames (CIF). Each CIF consists of a data field of 55296 bits including coding, which may be subdivided to form subchannels. The minimum subchannel throughput is 8 kbps and can be allocated to handle either packet or stream data. The four CIFs are divided into 72 blocks of 3072 bits and transmitted on the last 72 OFDM symbols in a frame. INTERFERENCE CANCELLATION The theoretical maximum capacity of a multiple access channel in additive white Gaussian noise is derived in (17). Achieving the capacity requires forward error correction coding and a method for extracting the interference caused by cochannel users from the desired signal. The latter is a serious problem. In mobile applications, the problem is even more challenging because of nonuniform propagation characteristics and cellular reuse strategies. The conventional receiver treats interference as equivalent to thermal noise. The approach of more advanced receivers is to recognize that there is some coherence to the interference and to attempt to process it in a manner that reduces its effect. These interference reduction/cancellation techniques form the leading edge of research in multiple access systems. In the following, we will briefly discuss a number of these techniques. We classify these techniques into three broad categories: (1) minimization techniques, (2) compensation techniques, and (3) multiuser detection techniques. Minimization Techniques Minimization techniques are those that attempt to minimize the interference before it gets into the receiver. They tend to apply more to narrowband than to wideband modulations. Intelligent Antennas. The simplest approach used in both terrestrial and mobile satellite systems is to use directional antennas at base station or satellite. This can reduce both intracell and intercell interference seen at a receiver. The simplest approach to this is to use sectored antennas as was discussed earlier under SDMA. This approach is passive in the sense that the antenna configuration is independent of the mobile terminal. A more advanced approach to this is using so-called intelligent antennas. Intelligent antennas take an active receiver role. With a phased array antenna, one can electronically optimize the antenna for each user, maximizing the gain in the direction of the desired signal and possibly positioning an antenna null in the direction of the strongest interferer (18). Furthermore, in a multipath environment research is continuing into designing the antenna beam to coherently capture as much multipath energy as possible. The latter is much easier to do in the return link than in the forward.

661

A second antenna-related approach that has been suggested for mobile satellite applications is to use signals with orthogonal polarizations. This is possible in satellite applications because there is a strong direct path, and it is suggested that this could possibly double capacity. However, there are some unanswered questions in this area, as reflected paths often reverse their polarization, which may cause problems for mobile applications. Dynamic Channel Assignment. Dynamic channel assignment is an interference avoidance technique. With this approach a base station continually monitors and/or probes channels during a call to determine the local interference characteristics (19). If the interference reaches a level that is degrading to the call, the call is automatic switched to a new frequency channel that has better characteristics. This approach not only provides protection against time-varying interference, but it also provides a form of switched frequency diversity that will provide some alleviation against multipath. For research in this area, see Ref. 19 and the references therein. New Multiple Access Techniques. A considerable amount of research is about future services to be provided to mobile terminals. The general consensus appears to be that the future terminal will be a form of multimedia terminal. Although the multimedia services that will be demanded is a subject of debate, it appears clear that significantly higher data rates than used at present will be required. Thus the problem is providing reliable high data rate service, such as 64 kbps to 2 Mbps, over a mobile fading channel. Considerable research (20) has been done on multicarrier (MC) techniques such as MCOFDM and MC-CDMA because of their potential to provide higher data rates in limited bandwidths and yet, because of the problems of the multicarrier approach, provide some frequency diversity that allows protection against multipath. It also allows implementation without the need for a high-speed equalizer capable of tracking channel variations. Compensation Techniques Compensation techniques attempt to minimize the effect of the interference once it gets into the receiver. These techniques rely only on minimal a priori knowledge of the interfering signal. They tend to apply to wideband desired signals, although the interference can be narrow or wideband. Narrowband Interference. With wideband modulation schemes such as DSSS there is the opportunity to excise narrowband interferers. The motivation of these approaches often comes from military applications where the desired signal is a DSSS signal and it is intentionally being jammed by a CW signal. However, the problem can also occur in commercial environments where a wideband system is being operated close to a narrowband system, or in some cases may be overlaid on a narrowband system. The object of these approaches is to reduce the spectral density of the interference to the level of the desired signal. The simplest technique is a notch filter that simply filters out the interferer. The degradation to the desired signal caused by the notch filter is approximately proportional to the fraction of the bandwidth removed. More advanced approaches recognize that interference-rejection

662

MULTIPLE ACCESS MOBILE COMMUNICATIONS

techniques need to be adaptive because of the dynamic nature of interference and the channel. Hence, adaptive filters or equalizers based on the LMS algorithm or Kalman filtering techniques are often considered. Nonlinear filters have also been shown to be advantageous for impulsive noise. For an extensive survey of these techniques see Laster and Reed (21) and the references therein.

the vector n(i) is a set of zero-mean correlated noise samples with E[n(i)n( j)T] ⫽ ␴2H( j)웃(i ⫺ j), and W is the diagonal matrix with non-zero elements 兵wk其, the propagation losses. With synchronous users H(1) and H(⫺1) are zero and the discrete time model reduces to

Wideband Interference. When one or more DSSS signals interfere with the desired signal, the interference can be viewed as colored noise, and equalization techniques can be considered. The techniques range from symbol-spaced to fractionally spaced to chip-rate based equalizers. An example of these techniques is the minimum mean square error (MMSE) multiuser detector (22), which produces a linear filter c(t), sampled at the chip rate, that minimizes the MSE between the received signal and desired signal. These techniques can be formulated with knowledge of the other user parameters but, in practice, they can be implemented without this knowledge using simple gradient search techniques. The simplest practical approach requires a training sequence to perform the initial adaptation before it proceeds to a decision-directed mode. See Ref. 21 and the references therein.

and the decorrelating detector estimates the data as

Multiuser Detection. Multiuser detection techniques are equivalent to detecting all the users and subtracting their effect from the desired user. Hence, these techniques require as much knowledge about the interference as they do about the desired signal. These techniques originally addressed DSSSCDMA systems (23), but they have been extended to FHSS systems and narrowband systems. Optimum. In Eq. (6) we described the optimum single-user detector as a bank of correlators with a correlator for every possible transmitted sequence bk. In a completely analogous fashion, if we let S(b, t) represent the corresponding K-user signal, then the optimum multiuser detector is a bank of correlators with a correlator for every possible transmitted Kuser sequence b. For binary transmission, this means 2NK correlators (of length N symbols), which is clearly beyond reason. Verdu (24) showed that for asynchronous DSSS systems this complexity could be reduced to a complexity of O (2K) per symbol using a variation of the Viterbi algorithm. In most CDMA applications, this is still not practical and has motivated the search for simpler but possibly suboptimal multiuser detectors. Decorrelator. Lupas and Verdu (25) showed that, with a DSSS system with rectangular pulses has the discrete time equivalent model

y (i) = H(+1)W 1/2b (i − 1) + H(0)W 1/2b (i) + H(−1)W 1/2b (i + 1) + n (i)

(31)

where y(i) is the K-vector of outputs from a bank of single user matched filters, optimally sampled at the symbol rate. The matrices H(i) are the crosscorrelation between the modulating waveforms [H(i)]kl =

Z

∞ −∞

sk (t − iT − τk )sl (t − τl ) dt

(32)

b (i) + n (i) y (i) = W 1/2 H(0)b

bˆ (i) = sign{H(0)−1y (i)}

(33)

(34)

In the asynchronous case, the asymptotic decorrelator is the discrete time matrix filter (25) D(z) = [H(−1)z−1 + H(0) + H(1)z]−1

(35)

that is applied to the vector sequence [y(i)] before detection. The decorrelating approach works well if the inverse operation is well-conditioned, otherwise it can result in noise enhancement that can significantly degrade the performance of some users. Multistage Detector for a Direct Sequence System. Varanasi et al. (26) describe a multistage detector for asynchronous CDMA that uses a conventional detector for the first stage but in the nth stage use the decisions of the (n ⫺ 1)st stage to cancel MAI present in the received signal. The performance of this feedback approach depends on the relative powers of the users. A more reliable multistage detector is obtained if the conventional detector in the first stage is replaced by a decorrelator. Multistage Detection for Frequency-Hopped Systems. There are various ways to address frequencies in a frequencyhopped system. One method is to perform a one-to-one mapping of the frequencies to the elements of a Galois field (13). A hopping sequence corresponds to a sequence of frequencies (addresses) that are represented as elements of a Galois field. Similarly, the data can be represented as elements of the same Galois field. The actual transmit frequency corresponds to the Galois-field sum of the two quantities. For this scheme, a simple iterative interference cancellation has been proposed by Fiebig (27). Effectively, the scheme assumes diversity transmission, with the same symbol transmitted on each of L hops, often referred to as fast frequency hopping. In Ref. 27, L is the log of the alphabet size, but this does not appear to be a requirement. For a particular symbol at the receiver, one despreads (dehops) the received signal for each of users, respectively. One then searches for any unambiguous decisions for individual users. Those unambiguous decisions are ‘‘erased’’ from the dehopped representations of other users, and further unambiguous decisions are searched for. This is continued until no further progress is made. The majority of the gain appears to be made with three iterations. This approach provides large improvements in the capacity over a system that uses a conventional detector and does so with minimal processing and no coding. Decision Feedback Equalizers (DFE). With a DSSS-CDMA system, the MAI can be modeled as equivalent to K-dimensional ISI. Consequently, equalizer approaches used for single

MULTIPLE ACCESS SCHEMES

663

user channels (1) can be extended to multiuser channels. In Ref. 28 Duel-Hallen et al. describe a multiuser decision feedback equalizer (DFE), characterized by two matrix transformations: a feedforward filter and a feedback filter. In addition to equalization, these multiuser decision feedback detectors employ successive cancellation using decisions made in order of decreasing user strength. The performance of the DFE is similar to the decorrelator for the strongest user, and gradually approaches the single-user bound as a user’s power decreases relative to the power of other users.

11. T. S. Rappaport, Wireless Communications, Englewood Cliffs, NJ: Prentice-Hall, 1996.

Multiuser Decoding. A limitation of many of the previously mentioned multiuser detectors is that their performance degrades at low SNR, or if there is high correlation between the users. This implies a degradation in performance if the data is FEC encoded, which is often the case in mobile channels. Encoded systems typically operate at lower SNRs, and consequently initial decisions on channel bits tend to be unreliable and insufficient to bootstrap an iterative scheme. It also implies a degradation if these techniques are used for narrowband modulations. The same behaviour occurs if users are highly correlated, that is, they are not spread as in a CDMA system, but are more narrowband as in an FDMA system. Theory still claims detection should be possible (17), but it is only recently that there have been clues about how to do it practically. In (29) an iterative technique is proposed for multiuser decoding, where each user employs FSK modulation. This can be viewed in some ways as an extension of Fiebig’s work (27) for frequency-hopped systems to the case where FEC encoding is employed with a soft decoding algorithm. In (30) another approach based on iterative decoding is applied to users whose correlations approach 1. This can be applied to users using PSK modulation with very little isolation between frequency reuses, and has the potential of closely approaching the theoretical capacity of the multiple access channel.

16. European Telecommunication Standard, Radio broadcast systems; Digital audio broadcasting (DAB) to mobile, portable and fixed receivers, ETS 300 401, June 1996.

BIBLIOGRAPHY 1. J. G. Proakis, Digital Communications, 2nd ed., New York: McGraw-Hill, 1989. 2. D. Parsons, The Mobile Radio Propagation Channel, New York: Wiley, 1992. 3. M. K. Simon et al., Spread Spectrum Communications, Rockville, MD: Computer Science Press, 1985, vols. 1, 2, 3. 4. A. D. Whalen, Detection of Signals in Noise, New York: Academic Press, 1971. 5. S. B. Weinstein, Data transmission by frequency-division multiplexing using the discrete Fourier transform, IEEE Trans. Commun. Technol., 19: 628–634, 1971. 6. A. V. Oppenheim and R. W. Schafer, Digital Signal Processing, Englewood Cliffs, NJ: Prentice-Hall, 1975. 7. J. D. Gibson (ed.), Handbook on Communications, Boca Raton, FL: CRC Press, 1997. 8. L. Kleinrock, Queueing Systems, New York: Wiley, 1976, vols. 1, 2. 9. D. J. Goodman et al., Packet reservation multiple access for local wireless communications, IEEE Trans. Commun., 37: 885–890, 1989. 10. V. H. MacDonald, The Cellular Concept, BSTJ, 58 (1): 15–41, 1979.

12. A. J. Viterbi, CDMA—Principles of Spread Spectrum Communications, Reading, MA: Addison-Wesley, 1995. 13. R. E. Blahut, Theory and Practice of Error-Control Codes, Reading, MA: Addison-Wesley, 1983. 14. S M. Redl, M. K. Weber, and M. W. Oliphant, An Introduction to GSM, Boston: Artech House, 1995. 15. TIA/EIA Interim Standard, Mobile station—base station compatibility standard for dual-mode wideband spread spectrum cellular system, TIA/EIA/IS-95-A, May 1995.

17. T. M. Cover and J. A. Thomas, Elements of Information Theory, New York: Wiley, 1991. 18. B. Ottersten, Array processing for wireless communications, Proc. 8th IEEE Workshop Stat. Sig. Array Process., Corfu, Greece, 1996, pp. 466–473. 19. I. Katzela and M. Naghshineh, Channel assignment schemes for cellular mobile telecommunication systems: a comprehensive survey, IEEE Personal Commun., 2 (6): 10–31, 1996. 20. K. Fazel and G. P. Fettweis (eds.), Multi-carrier Spread-spectrum, Dordrecht, The Netherlands: Kluwer Academic, 1997. 21. J. D. Laster and J. H. Reed, Interference rejection in digital wireless communications, IEEE Sig. Process., 14 (3): 37–62, 1997. 22. U. Madhow and M. Honig, MMSE interference suppression for direct-sequence spread spectrum CDMA, IEEE Trans. Commun., 42: 3178–3188, 1994. 23. A. Duel-Hallen, J. Holtzman, and Z. Zvonar, Multiuser detection for CDMA systems, IEEE Personal Commun., 1 (4): 46–58, 1995. 24. S. Verdu´, Minimum probability of error for asynchronous Gaussian multiple-access channels, IEEE Trans. Inf. Theory, 32: 85–96, 1986. 25. R. Lupas and S. Verdu, Near-far resistance of multiuser detectors in asynchronous channels, IEEE Trans. Commun., 38: 496–508, 1990. 26. M. K. Varanasi and B. Aazhang, Multistage detection in asynchronous code division multiple-access communications, IEEE Trans. Commun., 38: 509–519, 1990. 27. U.-C. G. Fiebig, Iterative interference cancellation for FFH/ MFSK MA systems, IEE Proc. Commun., 143 (6): 380–388, 1996. 28. A. Duel-Hallen, A family of multiuser decision-feedback detectors for asynchronous code-division multiple-access channels, IEEE Trans. Commun., 43: 421–434, 1995. 29. A. J. Grant and C. Schlegel, Collision-type multiple-user communications, IEEE Trans. Inf. Theory, 43: 1725–1736, 1997. 30. M. Moher, An iterative multiuser decoder for near-capacity communications, IEEE Trans. Commun., 46 (7): 1998.

MICHAEL MOHER Communications Research Centre

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELECTRICAL...NICS%20ENGINEERING/62.%20Vehicular%20Technology/W7717.htm

}{{}}

●

HOME ●

ABOUT US ●

CONTACT US ●

HELP

Home / Engineering / Electrical and Electronics Engineering

Wiley Encyclopedia of Electrical and Electronics Engineering Paging Communication for Locating Mobile Users Standard Article Sumita Mishra1 and Ozan K. Tonguz1 1State University of New York at Buffalo, Buffalo, NY Copyright © 1999 by John Wiley & Sons, Inc. All rights reserved. DOI: 10.1002/047134608X.W7717 Article Online Posting Date: December 27, 1999 Abstract | Full Text: HTML PDF (161K)

●

●

● ●

Recommend to Your Librarian Save title to My Profile Email this page Print this page

Browse this title ●

Search this title Enter words or phrases

Abstract The sections in this article are System Layout Paging Traffic and Normalized Paging Delay Determination of p Location Update and Signaling Traffic Cost Function Results Case Study: Limiting the Number of Paging Areas to Two Discussion About Wiley InterScience | About Wiley | Privacy | Terms & Conditions Copyright © 1999-2008John Wiley & Sons, Inc. All Rights Reserved.

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELE...ENGINEERING/62.%20Vehicular%20Technology/W7717.htm15.06.2008 12:46:04

❍

❍ ❍

Advanced Product Search Search All Content Acronym Finder

572

PAGING COMMUNICATION FOR LOCATING MOBILE USERS

cells. Whenever an MS changes its LA, location information about the MS, stored in certain registers in the fixed network, will be updated. This process is described in detail in the following section. Location updating will generate wireless traffic in the form of access traffic and will sometimes generate wireline traffic in the form of signaling traffic of Signaling System No. 7 (SS7) network (1). The act of seeking an MS when an incoming call has been placed to it is called paging. This reference to paging is specific to a function within cellular telephony, as opposed to paging systems (2,3). When a call arrives at the MS, its current location area is determined by interrogating the registers in the fixed network. Exact locating of mobile stations will reduce the time for call setup and increase the paging efficiency. Hence, the paging process generates wireless traffic in the form of paging traffic and wireline signaling traffic on the SS7 network when the registers are interrogated. Keeping track of the current LAs associated with each MS, as well as locating the user when a call is received, requires frequent interactions between the mobile stations and the system. Reducing such interactions is one of the goals of efficient location management (4–11). Several strategies have been proposed in the literature that deal with the reduction of signaling traffic and database loads imposed by the need to locate and track users (12–15). Some researchers have looked into methods of reducing the access traffic (16–18). For example, in Ref. 16, instead of the LA-based method of updating the current user location, the authors investigate time-based, distance-based, and movement-based methods of updating. However, in this article, we define the overall cost of location management in terms of the three components, namely, paging cost, access cost, and signaling cost, with emphasis on the reduction of the wireless paging cost.

SYSTEM LAYOUT

PAGING COMMUNICATION FOR LOCATING MOBILE USERS The increased demand for wireless personal communication systems (PCS) coupled with limited spectrum has motivated many researchers to look into the techniques that minimize the radio traffic needed to keep track of the mobile stations (MSs) and deliver information (voice, data, video, etc.) to them. The process of keeping track of the MSs by updating the location information and paging for them when they receive a call is known as location management. In order to keep track of the MS, the whole geographical area is divided into location areas (LAs) which are nothing but a logical group of

The main idea behind location management is to locate the MS with minimum overhead, whenever there is an MS-terminated call. Note that the entire process of locating users is carried out between calls when the mobile stations are in stand-by mode. In the European Global System for Mobile Communications (GSMs), SS7 interconnects Mobile Switching Centers (MSCs), Home Location Registers (HLRs) and Vehicle Location Registers (VLRs). An MSC is connected to several base stations, other MSCs and the public switched telephone network (PSTN). The permanent records of the mobile users are stored in the HLR. It also contains the pointers to the MS’s current VLR. The VLR is a database associated with a particular MSC. When an MS crosses into an LA under the MSC, the VLR associated with that MSC is updated. On the other hand, when the MS crosses into an LA under a different MSC, the old and new VLRs as well as the HLR is updated (1, 14). These processes generate traffic on reverse control (access) channels. Signaling System No. 7 (SS7) traffic is generated whenever there is an HLR update. When an MS receives a call, it needs to be paged for in a certain area. Its current position (LA) is found out by interrogating the HLR. This generates SS7 traffic. The size of the area that is paged depends on the paging technique used by that particular system. In any case, there is traffic generated

J. Webster (ed.), Wiley Encyclopedia of Electrical and Electronics Engineering. Copyright # 1999 John Wiley & Sons, Inc.

PAGING COMMUNICATION FOR LOCATING MOBILE USERS

(1)

SS7

HLR (3)

(2) MSC

PSTN Cell

VLR

(4)

573

kind of user information (the most recent interaction area, speed, movement pattern etc.), so that the probability of finding the MS in a certain PA or PAs is more than that for the rest of the LA. The VLR(s) and, in the cases described above, the HLR are updated every time the MS changes LAs and performs location update using the access channel. PAGING TRAFFIC AND NORMALIZED PAGING DELAY

Paging area Location area Figure 1. Cellular communications system configuration with visitor location register (VLR), home location register (HLR), paging area, and location area. The call setup procedure is shown by numbers (1) to (4). (1) Interrogation with HLR, (2) wireline call setup, (3) interrogation with VLR, (4) paging and radio call setup. MSC: mobile switching center, SS7: signaling system no. 7.

on the forward control (paging) channels. The various steps involved in a call setup are depicted in Fig. 1. The above-mentioned traffic components (i.e., access traffic, paging traffic, and SS7 traffic) are the primary factors that govern the cost of location management. Since this is just an overhead cost for the service providers, the lower the cost incurred for locating the MS, the better it is. Hence, there is a need for better techniques that would lower the overall cost. Location area (LA) and paging area (PA) are differentiated in the following discussion (4,5). A location area is a logical group of multiple paging areas, and a paging area is a logical group of cells. To find out the relationship between the size of location area, paging traffic, radio access traffic, signaling traffic, and their dependence on call arrival rate and the boundary crossing rate, as well as to determine the optimal paging techniques, the following cases of location management are considered in this article. Case 1: Simultaneous Paging. One location area is composed of n cells. All the cells inside an LA will be paged simultaneously when there is a call to any one cell in the LA. The VLR(s) and, in cases described above, the HLR are updated every time the MS changes LAs and performs location update using the access channel. It is noted that if the size of LA is changed, the paging and access traffics generated operate in a counteractive fashion (4,5). In other words, small LAs generate less paging traffic and more access traffic compared with large LAs. Therefore, it is not possible to decrease the cost generated by the simultaneous paging technique by varying the size of the LA. Case 2: Sequential Paging. One location area is composed of n cells. Paging will be done in k-cell (k ⫽ n) PAs sequentially until the called MS is found. The PAs are selected in a random fashion for this preliminary analysis. The probability of finding an MS is assumed to be the same in all the PAs of the LA. The VLR(s) and, in cases described above, the HLR are updated every time the MS changes LAs and performs location update using the access channel. Case 3: Intelligent Paging. One location area is composed of n cells. Paging will be done in k-cell (k ⱕ n) PAs sequentially until the called MS is found. The PAs are selected using some

When simultaneous paging is used, for an n-cell LA, the paging traffic generated for each cell is given by tp ⫽ n␭, where ␭ is the call-arrival rate (incoming calls/s) for each cell. Here, we assume that the call-arrival rate is uniform for all the cells in the LA. In the case of sequential paging, LAs are composed of multiple PAs. The PAs are paged one by one until the called MS is located. The paging traffic will depend on the average number of paged cells until the MS is located. When the probability of locating an MS is assumed to be the same in each PA, k/n, the probability of finding a called MS after paging i number of PAs, remains the same; that is, p ⫽ k/n. Therefore, the average number of paged PAs (␰k), which is the same as normalized paging delay (normalized with respect to simultaneous paging), will be as follows (4): ξk = 0.5


n k

+1



(1)

Here we assume that n is an integer multiple of k. The equation for the general case is also derived in Ref. 4. The paging traffic for each cell with k-cells PA and n-cells LA with the same sequential paging technique will be tp = kξk λ = 0.5(n + k)λ

(pages/s)

(2)

As a numerical example, when 10 cells make one LA, the paging traffic compared to one-cell LA increases by a factor of 10 with simultaneous paging, and by a factor of 5.5 with sequential paging with k ⫽ 1. In this case, the paging delay is 5.5 times longer with sequential paging compared to simultaneous paging. However, when k ⫽ 5, the paging traffic for sequential paging is 7.5 times that of a one-cell LA, but its paging delay is reduced to just 1.5 times that of simultaneous paging. Hence, when multiple-cell PAs are used in the case of sequential paging, there is a considerable decrease in the paging delay. The price paid is a slight increase in the paging traffic. Results on paging delay decrease and paging traffic increase normalized to the case of n being 1 with four different values of k (1, 3, 5, and 10) are shown in Figs. 2 and 3, respectively. In Eq. (2), it seems that in order to minimize the paging cost, k can take any value from 1 to n; in other words, the paging delay is completely ignored. This is not true for real systems. There is a finite value of the acceptable average delay (␶max). ␶max is normalized with respect to the delay of simultaneous paging in the remainder of this article. This ␶max can be used to determine the maximum ratio of the size of LA and the size of PA [see Eq. (1)]. Since ␰k ⱕ ␶max, using Eq. (1), the size of PA (k) can be expressed in terms of the acceptable average paging delay (␶max) and the size of LA (n) as follows: k≥

n 2τmax − 1

(3)

574

PAGING COMMUNICATION FOR LOCATING MOBILE USERS

50

200 k=1 k=3 k=5 k = 10

150 Paging traffic

Paging delay (normalized)

40

SIM = SEQ (delay = 1) SEQ (delay = 3) SEQ (1 < delay < 3) SEQ (infinite delay)

30

20

100

50 10

0 10

20

30

40

50

60

70

80

90

100

n

Hence, kmin =



n

(4)

2τmax − 1

where  is the ceiling function, or the ‘‘rounding-up’’ function. Substituting this kmin for k in Eq. (2), we obtain the paging traffic generated for each cell while taking the paging delay into account. This paging traffic is given by
 n  tp = 0.5 n + λ (pages/s) (5) 2τmax − 1 Note that Eq. (4) gives the minimum k that would satisfy the delay requirement. So, k can be larger than kmin and still satisfy the requirement. Of course, as k increases, the paging

60 k=1 k=3 k=5 k = 10

Paging traffic (normalized)

50

40

30

20

10

0 10

20

30

40

50

0

20

40

60

80

100

n

Figure 2. Normalized paging delay (normalized with respect to the delay for simultaneous paging) of the sequential paging scheme for different paging area sizes.



0

60

70

80

90

100

n Figure 3. Normalized paging traffic (normalized with respect to the traffic for simultaneous paging with n ⫽1) of the sequential paging scheme for different paging area sizes.

Figure 4. Normalized paging traffic [normalized with respect to the traffic for sequential paging (ignoring delay) for n ⫽ 1] of all schemes.

traffic increases. Hence, it is desirable to operate with PA size as kmin. In Fig. 4, the normalized paging traffic for simultaneous paging, sequential paging (without considering delay), and sequential paging (with paging delay consideration) are plotted versus the LA size. Observe that the traffic for sequential paging (without considering delay, i.e., infinite delay) has a normalized slope of 1 and the simultaneous paging traffic has a normalized slope of 2. The paging traffic due to sequential paging (with paging delay consideration) has a curve which lies between the other two. In other words, the enforcement of the delay requirement on sequential paging increases the gradient of the paging traffic curve. In fact, as the delay requirement becomes more and more stringent (i.e., as ␶max decreases), the curve of sequential paging traffic tends toward that of simultaneous paging traffic, which is the case when ␶max ⫽ 1. The behavior of paging traffic in Fig. 4 for different paging delay requirements confirms the physically intuitive result that to minimize the delay in sequential paging, the logical distinction between paging area and location area should be abolished (i.e., k ⫽ n), in which case we note that sequential paging converges to simultaneous paging. In practice, the permissible paging delay in a PCS environment will dictate the slope of the paging traffic curve. If, for instance, the normalized acceptable average paging delay is 3, then the paging traffic will be the solid line shown in Fig. 4. During the busy hours, if the network chooses to operate with less delay (suppose 1 ⬍ normalized paging delay ⬍ 3), then one will have a paging traffic curve which will lie between that of simultaneous paging and the curve for delay ⫽ 3, as shown in Fig. 4. Also note that when the required paging delay becomes more stringent (i.e., decreases), the value of paging traffic for a fixed n increases. This implies that the effective weight of paging traffic and, consequently, the cost of location management goes up. In order to decrease this paging traffic for sequential paging, we need some more information (intelligence) about the users in the system. In this article, the forms of intelligence that are studied are the speed information and the most re-

PAGING COMMUNICATION FOR LOCATING MOBILE USERS

where Fped is the fraction of the total number of MSs in a cell that are pedestrians and Fcar is the corresponding fraction of cars.

LA 1 PA 1

PA 2 R

Cell 1 Cell 2

575

DETERMINATION OF p

Cell 16 PA 3

PA 4

Figure 5. Cellular layout showing the cells and the paging areas within a location area.

cent interaction information about the MS. For simplicity, instead of finding the exact speed, the users are classified into two categories, namely, pedestrians and cars. The average speed of pedestrians is considered to be 5 km/h and the average speed of fast-moving MSs is assumed to be 30 km/h. The determination of the category is done by the MS, and this information is passed on to the system in the form of an extra bit when the LA update process is being carried out. This is done so that no additional radio traffic is generated. The information about the MS speed is stored in the VLR along with the other records. The speed is also updated when the MS uses the access channel to answer a paging message or to make a call. Suppose an MS is in PA1 of LA1 when its most recent interaction with the fixed network takes place (see Fig. 5). This interaction could be in the form of a location update or a call origination or termination, and so on. We consider squareshaped cells in this work. If the MS receives a call within a certain time duration after update, then the probability (p) of finding the MS in PA1 is greater than that of the other PAs in the LA, that is, p ⬎ k/n. The expected number of paged PAs in terms of the probability of finding the MS in the first PA paged (p), which is referred to as Probability of Successful First Paging Step (PSFS) in Ref. 11, is derived in Refs. 6 and 7, and is given by




ξk = p + (1 − p) 1 +

n 2k

(6)

Hence, the paging traffic generated per MS, tpu, is given by

tpu = kλu ξk
 n  = kλu p + (1 − p) 1 + 2k

(pages/s)

(7)

where ␭u ⫽ ␭ /N is the average call-arrival rate per user per cell. Here, N is the average number of in-use and standby MSs in a cell. Note that k is to be determined considering the acceptable average delay (␶max) for sequential paging. Since p is dependent on the MS speed, it is different for pedestrians (pped) and cars (pcar) at a certain instance of time. Hence, the total paging traffic generated for each cell by intelligent paging (pages/second) is given by


 n  tp = Fped pped + (1 − pped ) 1 + 2k


n  +Fcar pcar + (1 − pcar ) 1 + kλ (8) 2k

It is obvious that the probability of successfully finding the MS in the first area paged (i.e., p) is inversely proportional to the time lapse between the last update and received call (T). Assume that p ⫽ 1 up to T ⫽ ti. Also, for T ⬎ tm [where tm is a certain time duration that depends on the MS speed, the cell radius, the number of cells in a PA (k), etc.], p ⫽ k/n —that is, the case of nonintelligent sequential paging. Hence, this simple version of intelligent paging is different from the sequential paging technique only when T ⬍ tm. If we assume a linear relationship between p and T for ti ⬍ T ⬍ tm, then   1 for 0 ≤ T < ti      

    t − k t −T 1− k m n i n p= (9) for ti ≤ T < tm   t − t  m i      k   for T ≥ tm n Intuitively, ti and tm is inversely proportional to the MS speed (v). However, they are directly proportional to the cell radius (R) and the number of cells in a PA (k). Of course, the exact relationship between them depends on the trajectory of the MS. For example, for a straight-line motion of the MS across a square-shaped PA, ti ⫽ 2兹kR/v. On the other hand, for diagonal motion across the PA, ti ⫽ 2兹2kR/v. It can be shown that this is a good approximation for non-square-shaped PAs also, provided that the PAs are chosen in a manner that is a good approximation to the regular square shape. If we have a square-shaped LA also, then tm ⫽ (兹(n/k) ⫺ 1)ti for 兹n/k ⬎ 2 in each case. For, 1 ⱕ 兹n/k ⱕ 2, tm ⫽ ti. Again, this can be used as an approximation for non-square-shaped cases also. We consider straight-line motion for square-shaped cells in this article. So, other things being equal, since vped ⬍ vcar, the p versus T curve for pedestrians is different from that for fastmoving users, as shown in Fig. 6. Note that the above discussion is for the scenario when we assume that the MS is always on the move. If it stays in the same location after speed update until it receives its next call, it will always be found in the first paging attempt because we start paging from the most recent interaction area. LOCATION UPDATE AND SIGNALING TRAFFIC The location management schemes considered in this article vary only in terms of the paging technique used. Since access and SS7 traffics are generated exactly in the same manner for all three cases, they are discussed in common in this section. Whenever an MS crosses an LA, access traffic is generated. This traffic is dependent on the boundary crossing rate (BCR) of the MS, which is defined as the number of MSs crossing into a cell-boundary per second. This BCR is denoted by 애, and the average access traffic per cell is ta ⫽ 애/ 兹n accesses/ second. This means that the average access traffic per cell, for location updating, will decrease as n increases. Here, we

576

PAGING COMMUNICATION FOR LOCATING MOBILE USERS

Probability (p)

1

Pedestrian

Car

k/n

ti(car)

tm(car) ti(ped)

tm(ped)

Elapsed time (T)

Figure 6. Plot of the probability of first successful paging step versus the elapsed time from last update for pedestrians and cars.

average out the access traffic over all the cells of an LA to see the overall picture. However, it is clear that the border cells of an LA will encounter updating traffic whereas the center cells have no access traffic due to updating. This implies that more control channels should be assigned to the border cells of an LA compared to the center cells. Note that 애 is dependent on the number of MSs in a cell (N), the cell radius (R), and the speed of the MS (v) and is given by 2Nv/앟R crossings/ second (4,18). SS7 traffic is generated whenever the MS crosses LAs under different MSCs, since this requires the HLR to be updated. Also, when there is an MS-terminated call, we need to interrogate the HLR for the current position of the MS, thus generating SS7 traffic. If nt is the total number of cells controlled by an MSC, then the SS7 traffic generated per cell is ts ⫽ 애/ 兹nt ⫹ ␭ signaling-messages/second.

the cost of location management) will be different. Also, if the SS7 network is constructed using the not-so-expensive fiberoptic technology, then ws may be significantly less than wp and wa. Hence, the location management cost for the three schemes is as follows:   µ µ Costsim = wp nλ + wa √ + ws √ + λ (11) nt n    µ λ µ   w (n + k) + w √ + λ + w √ p a s   2 nt n       (delay ignored)  Costseq =    
   µ n λ µ   wp n+ + wa √ + ws √ + λ    2τmax − 1 nt n  2    (considering delay) (12)


n  Costint = wp Fped pped + (1 − pped ) 1 + 2k


n  (13) + Fcar pcar + (1 − pcar ) 1 + kλ 2k   µ µ + wa √ + ws √ + λ nt n Note that for the intelligent scheme, the delay requirement is taken into account, and so k ⫽ [n/(2␶max ⫺ 1)]. RESULTS Costs of the three cases are plotted versus the location area size n for two values of BCR 애. In the first scenario, we consider the case of an area where the pedestrian population is more than the fast-moving MSs. For example, this could be the situation in a downtown city area. The plots are computed for nt ⫽ 400, ␶max ⫽ 3, N ⫽ 200, Fped ⫽ 0.9, Fcar ⫽ 0.1, R ⫽ 200 m. For this case, ␭ ⫽ 0.75 calls/s and 애 ⫽ 1.33 crossings/s. In Fig. 7, we consider the case when all weights are equal. We find that the intelligent scheme is always the most cost-

20

COST FUNCTION

Cost = wptp + wata + ws ts = f (λ, µ, n, k, nt )

(bytes/s)

15

Cost

As mentioned before, the key components of location management cost are the costs generated by paging traffic (tp), access traffic (ta), and SS7 signaling traffic (ts). These costs depend on the BCR (애), call arrival rate (␭), the number of cells in a location area (n), the number of cells in a paging area (k), and the number of cells controlled by an MSC (nt). Comparison of the three schemes is done via the cost function defined as follows:

Intelligent Sequential (delay ignored) Sequential (delay = 3) Simultaneous

10

5

(10)

where wp (bytes/page), wa (bytes/access-message), and ws (bytes/signaling-message) are weights assigned to each traffic component. These weights may be different depending on the access method used, the availability of channels, and the cost of constructing the signaling network. Depending on the access method, the number of bytes required per second (i.e.,

0

0

5

10

15

20

25

n Figure 7. Overall cost of the four cases versus the size of location area (n) for equal weights assigned to paging, access, and signaling costs and dominant pedestrian population.

PAGING COMMUNICATION FOR LOCATING MOBILE USERS

577

60

4.5

Intelligent Sequential (delay ignored) Sequential (delay = 3) Simultaneous

4 3.5

Simultaneous Sequential Intelligent 40 Cost

Cost

3 2.5 2

20

1.5 1 0.5

0

0

10

20

30

40

50

n Figure 8. Overall cost of the four cases versus the size of location area (n) when access cost is assigned more weight than the paging and the signaling costs.

effective scheme for this situation. Also, there is an optimal value of n (nopt) in each case, for which the overall cost is minimum. We find that nopt(sim) ⫽ 1, nopt(seq) ⫽ 2 and nopt(int) ⫽ 5. Hence, larger location areas are possible in the case of the intelligent scheme. The PA size k is 1 for both sequential and intelligent schemes, since it is dictated by the acceptable average delay with respect to simultaneous paging, i.e., ␶max. In Fig. 8, we consider the case when the weight of access traffic is 10 times more than each of the other components. It is worth noting that the general trends of all the schemes remain the same, although a much larger size of optimal LA is obtained when the access traffic is the dominating factor. For dominant paging traffic and dominant SS7 traffic scenarios, it is found that the trends are the same as the ‘‘all weights equal’’ case. In Fig. 9, we consider the situation when the fast-moving population is dominant. This could be a situation in the sub20

Intelligent Simultaneous Sequential (delay = 3) Sequential (delay ignored)

15

0

50

100

150

200 T (s)

250

300

350

400

Figure 10. Overall cost of the three schemes versus the elapsed time between the last interaction of the mobile with the system and the call received. 애 ⫽ 1.33.

urban area. Here N ⫽ 200, Fped ⫽ 0.2, Fcar ⫽ 0.8 so that 애 ⫽ 4.42 crossings/s. All the other parameters are the same as the previous case. Again the intelligent scheme is a winner. However, the margin of difference between the costs of the sequential and intelligent paging techniques becomes smaller. This shows that the intelligent scheme is more beneficial in situations where the pedestrian population is dominant. This is because the probability of finding a slow-moving MS during the first paging attempt is higher than that of finding a car at a given instance of time. Figure 10 is plotted to show that the intelligent scheme fares better than the other schemes only up to a certain value of T for a certain scenario. Beyond this value of T, it is the same as the sequential paging scheme. For 애 ⫽ 1.33 crossings/s, n ⫽ 5, and k ⫽ 1 , this particular value of T is 288 s. Note that this analysis is for the worst-case scenario when it is assumed that the MS is constantly on the move. For a dominant slow-moving population in office-buildings in a downtown area, intelligent paging fares better than sequential paging up to a much larger T. The costs of the simultaneous and the sequential paging schemes change with T because of changing ␭. Also, since k ⫽ 1, sequential paging has just one curve.

Cost

CASE STUDY: LIMITING THE NUMBER OF PAGING AREAS TO TWO 10

5

0

0

5

10

15

20

25

n Figure 9. Overall cost of the four cases versus the size of location area (n) for equal weights assigned to paging, access, and signaling costs and dominant fast-moving population.

The above results are obtained on the basis of the theoretical framework provided in the article. However, in practice, it may not be desirable to have more than two paging areas (8). In other words, if the MS is not located in one attempt, it should be located in the second try (9). If that is the case, then n/k ⫽ 2 so that k ⫽ n/2 and Eq. 5 becomes   n  tp = 0.5 n + λ (14) 2 For intelligent paging, the expected number of paged PAs becomes 2 ⫺ p so that the paging traffic given in Eq. (8) becomes tp = [Fped (2 − pped ) + Fcar (2 − pcar )] kλ

(15)

578

PAGING COMMUNICATION FOR LOCATING MOBILE USERS

In this case the expression for p becomes

0.5

100 90

for 0 ≤ T < ti

(16)

for T ≥ ti

The costs of sequential and intelligent paging schemes are now given by Costseq = wp

 n  λ
µ n+ + wa √ + ws 2 2 n



 µ √ +λ nt

(17)

Costint = w p [Fped (2 − pped ) + Fcar (2 − pcar )]kλ   µ µ + wa √ + ws √ + λ nt n

(18)

Percentage cost reduction (%)

p=

 1

80 70 60 50 40

All-step (straight line) All-step (biased Markovian) All-step (random walk)

30 20 10

Figure 11 shows the cost curves under this constraint for a dominant pedestrian population. The parameter values are the same as those for Fig. 7. It is found that the optimum cost of intelligent paging is 82% of the optimum cost of sequential paging when the SS7 traffic is negligible. On the other hand, when the SS7 traffic is considered to have the same weight as the other two components, there is a 12% reduction in the cost of intelligent paging, compared to sequential paging. DISCUSSION In this article, location management in PCS is studied via three different paging schemes. It is also shown that if a simple form of intelligence (i.e., additional information about MS speed and recent interaction) is added to the sequential paging scheme, the overall cost does come down under certain circumstances. The performance of this intelligent scheme is highly dependent on the time lapse between the speed update and call-arrival. In any case, the performance of this intelligent scheme is never worse than the sequential paging scheme. This performance (overall cost of location management) can be further improved if more information about the MSs is incorporated (11).

10 9

Simultaneous Sequential Intelligent

8

Cost

7 6 5 4

0

0

250

500

750 Time (s)

1000

1250

1500

Figure 12. Percentage cost reduction with respect to One-Step Intelligent paging in All-Step Intelligent paging strategies (for straightline, biased Markovian and Random Walk motion models).

To assess the total cost of location management, a simple cost function is defined and used whereby the weights assigned to each of these traffic components would also depend on other factors, such as (1) the specific realization of the PCS environment and (2) the architecture of the signaling network. This analytical framework can be used to analyze the performance of simultaneous, sequential, and ‘‘intelligent’’ paging schemes in a unified manner. While this framework leads to similar results with those obtained via the more rigorous approaches (19,20), it also provides valuable physical insight into the impact of main system parameters such as maximum paging delay, call-arrival rate, and boundary crossing rate on the overall cost of location management. In this article, we have provided results for simplistic cases such as deterministic call-arrival rate and boundary crossing rate, straight-line motion for the mobile users, and so forth. However, our preliminary investigations show that the analytical framework provided in this chapter holds for timevarying call-arrival and boundary crossing rates and for random user-movement patterns (Markovian, Random Walk, etc.) as well (21). Also, the concept of ‘‘intelligence’’ has so far been investigated only for the first paging step. It has been verified that this can be extended to the second and subsequent steps under the same framework (21). Figure 12 compares the All-Step strategy with One-Step Intelligent paging, for straight-line, Markovian and Random Walk movement models. Our goal is to come up with a generalized theory for analyzing any location management scheme for the future PCS.

3 2

1

2

3

4

5

6

7

8

9

10

Figure 11. Overall cost of the three schemes versus the size of location area (n) for two paging areas.

BIBLIOGRAPHY 1. CCITT Blue Book Recommendations Q.1000, ‘‘Public Land Mobile Network Interworking With ISDN and PSTN,’’ The Interna-

PAPER INDUSTRY, SYSTEM IDENTIFICATION AND MODELING tional Telegraph and Telephone Consultative Committee, Nov. 1988. 2. D. Baker, A Comprehensive Guide to Paging, BIA Publications Inc., 1992. 3. K. Siwiak, Radiowave Propagation and Antennas for Personal Communications, 2nd ed., Norwood, MA: ArtechHouse, 1998. 4. O. K. Tonguz, S. Mishra, and H. Jung, A simple analytical framework for location management in personal communication systems, IEEE Trans. Veh. Technol., 42 (2): 428–439, 1998. 5. D. Plassmann, Location management strategies for mobile cellular networks of 3rd generation, Proc. 44th IEEE Vehicular Technology Conference, 1994, pp. 649–653. 6. S. Mishra and O. K. Tonguz, Most recent interaction area and speed-based intelligent paging in personal communication systems, Proc. 47th IEEE Vehicular Technology Conference, May 1997, pp. 505–509. 7. S. Mishra and O. K. Tonguz, Analysis of intelligent paging in personal communication systems, Electronics Letters, 34 (1): 12– 13, Jan. 1998. 8. S. Madhavapeddy, K. Basu, and A. Roberts, Adaptive paging algorithms for cellular systems, Proc. 45th IEEE Vehicular Technology Conference, 1995, pp. 976–980. 9. S. Mishra and O. K. Tonguz, Location management with intelligent paging in personal communication systems, IEEE Proc. Communications, submitted for publication. 10. H. Xie, S. Tabbane, and D. J. Goodman, Dynamic location area management and performance analysis, Proc. 43rd IEEE Vehicular Technology Conference, 1993, pp. 536–539. 11. G. L. Lyberopoulos et al., Intelligent paging strategies for third generation mobile telecommunication systems, IEEE Trans. Veh. Technol. 44 (3): 543–554, August 1995. 12. R. Jain et al., A caching strategy to reduce network impacts of PCS, IEEE J. Selected Areas in Communications, 12 (8): 1434– 1445, 1994. 13. Y.-B. Lin and A. Noerpel, Implicit deregistration in a PCS network, IEEE Trans. Veh. Technol., 43 (4): 1006–1009, Nov. 1994. 14. K. S. Meier-Hellstern and E. Alonso, The use of SS7 and GSM to support high density personal communications, Proc. ICC ’92, 1992, pp. 1698–1702. 15. G. P. Pollini, The intelligent network signalling and switching costs of an alternate location strategy using memory, Proc. 43rd IEEE Vehicular Technology Conference, 1993, pp. 931–934. 16. A. Bar-Noy, I. Kessler, and M. Sidi, Mobile users: to update or not to update?, Proc. IEEE INFOCOM ’94, 1994, pp. 570–576. 17. S. Okasaka et al., A new location updating method for digital cellular systems, Proc. 41st IEEE Vehicular Technology Conference, 345–350, 1991. 18. I. Seskar et al., Rate of location area updates in cellular systems, Proc. 42nd IEEE Vehicular Technology Conference, 1992, pp. 694–697. 19. C. Rose and R. Yates, Minimizing the average cost of paging cost under delay constraints, Wireless Networks, I, 211–220, 1995. 20. J. S. M. Ho and I. Akyldiz, Mobile user location update and paging under delay constraints, Wireless Networks, I, 413–425, 1995. 21. O. K. Tonguz et al., All-step intelligent paging: a location strategy for mobile networks, submitted for presentation at IEEE INFOCOM ’99.

SUMITA MISHRA OZAN K. TONGUZ State University of New York at Buffalo

579

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELECTRICAL...NICS%20ENGINEERING/62.%20Vehicular%20Technology/W7716.htm

}{{}}

●

HOME ●

ABOUT US ●

CONTACT US ●

HELP

Home / Engineering / Electrical and Electronics Engineering

Wiley Encyclopedia of Electrical and Electronics Engineering Vehicle Navigation and Information Systems Standard Article Yilin Zhao1 1Motorola, Inc., Schaumburg IL Copyright © 1999 by John Wiley & Sons, Inc. All rights reserved. DOI: 10.1002/047134608X.W7716 Article Online Posting Date: December 27, 1999 Abstract | Full Text: HTML PDF (178K)

●

●

● ●

Recommend to Your Librarian Save title to My Profile Email this page Print this page

Browse this title ●

Search this title Enter words or phrases ❍

❍ ❍

Advanced Product Search Search All Content Acronym Finder

Abstract The sections in this article are Subsystem Technologies System Technologies Conclusion Keywords: location tracking; electronics in navigation; positioning; digital map database; map matching; route planning; route guidance; human-machine interface; wireless communications; autonomous navigation; centralized navigation; automatic vehicle location; fleet management; mayday systems; dynamic route guidance; telematics About Wiley InterScience | About Wiley | Privacy | Terms & Conditions Copyright © 1999-2008John Wiley & Sons, Inc. All Rights Reserved.

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELE...ENGINEERING/62.%20Vehicular%20Technology/W7716.htm15.06.2008 12:46:24

106

VEHICLE NAVIGATION AND INFORMATION SYSTEMS

VEHICLE NAVIGATION AND INFORMATION SYSTEMS Travel in a vehicle has traditionally involved following the directions given by a person, posted along the road, or found from a paper map. With the rapid advance of computer, control, communication, and information technologies, these travel methods will be gradually augmented by much more powerful and convenient ones. Systems built using these advanced technologies are called vehicle navigation and information systems. The goal of these systems is to guide vehicle occupants to their destinations safely and efficiently, with less congestion, pollution, and environmental impact. This is a goal shared with intelligent transportation systems (ITS), which are transportation systems that apply advanced technologies to help their operations. In working toward this goal, ITS can take different forms. As an important portion of ITS, vehicle navigation and information systems provide the foundation necessary to understand, design, and implement advanced ITS.

A variety of navigation and information systems have been developed to assist vehicle operators, such as traffic display boards, variable message signs, rollover advisory systems, parking guidance, and ramp metering (breaking up the platoons on the ramp). In this article more emphasis is placed on the technologies used in the vehicle, in particular those relevant to guide the surface vehicle (navigation) and to advise the user about it (information) (1). The earliest vehicle navigation and information system can be traced according to legend, back to around 2600 B.C. in ancient China. It was called ‘‘south-pointing carriage.’’ This carriage had a two-wheeled cart on which was mounted a wooden human figure. No matter which way the cart was moving, the figure was kept continually pointed toward the south (2,3). Another interesting invention was the ‘‘li-recording (distance-measuring) drum carriage.’’ Two wooden human figures sat on the carriage along each side of the drum. As the carriage moved, the wheels turned to drive the arms of these figures via a set of gears. An arm of one figure would strike one side of the drum once every li (about half of a kilometer); the other figure would strike the other side once every 10 li (2,3). The working mechanisms of these carriages are similar to the modern positioning technologies. Basic positioning and navigation technologies have been incorporated gradually into modern automobiles (1,4). In the early twentieth century, mechanical route-guidance devices were introduced into modern automobiles. During World War II, an electronic vehicle navigation system was developed in the United States for military vehicles. In the late 1960s, an electronic route guidance system (ERGS) was proposed in the United States to control and distribute the traffic flow with wireless route guidance capability (centralized dynamic navigation). Similar projects were later developed and tested in Japan and Germany in the 1970s. During the same period, autonomous navigation systems were developed in the United States and United Kingdom. In the 1980s, improved autonomous navigation and information systems began to appear in Japanese, European, and American markets. These systems have been further improved in the 1990s to include many modern technologies discussed below. There is a range of vehicle navigation and information systems from the very low end to the very high end. Most of them consist of some or all of the basic modules depicted in Fig. 1. To facilitate the discussion, we first present subsystems of vehicle navigation and information systems as separate modules and then discuss how they work together. The modules shown in Fig. 1 can be implemented by different hardware and software components. A positioning module automatically determines the position of the vehicle. It can employ either integrated sensor data or radio signals to identify the coordinates of the vehicle (position) or the placement of the vehicle relative to landmarks or other terrain features (location). A variety of sensor fusion methods and radio-signal-based methods have been developed for positioning. A digital map database contains digitized map information. It can be processed by a computer for map-related functions, which are very similar to those provided by conventional paper maps and travel guides. Many expanded features may also be facilitated by the computerized map. Map matching is to use a position (or route) on a map to match the position (or trajectory) determined by a positioning module. This method can improve the accuracy of the positioning module, provided

J. Webster (ed.), Wiley Encyclopedia of Electrical and Electronics Engineering. Copyright # 1999 John Wiley & Sons, Inc.

VEHICLE NAVIGATION AND INFORMATION SYSTEMS

Route guidance

Map matching

Human-machine interface

Digital map database

Route planning

Positioning

107

For a large and complex system, implementation typically goes through certain well-defined phases. A top-down approach is often used, which typically includes identifying system requirements, determining functions and system architecture, specifying appropriate modules, selecting hardware components and software tools, and designing, implementing, integrating, and testing the system. Each of these individual phases needs to be iterated a few times to come up with solutions agreed upon by all the project teams and customers involved. SUBSYSTEM TECHNOLOGIES

Wireless communications

Figure 1. Basic modules for a vehicle navigation and information system. The wireless communications module interacts with the system over the air while the rest of the modules interact with each other on board. This modular approach makes system study, design, and implementation easy to accomplish. The same approach can be applied to the vehicle side and the service (traffic management/dispatch) center side of the system which is connected by a wireless communications network.

the map database has sufficient precision, normally within 15 m of ‘‘ground truth.’’ Route planning helps vehicle drivers plan a route prior to or during their journey. Different criteria can be used to find a minimum-travel-cost route. These include time, distance, and complexity. Route guidance directs the driver along the route calculated by the route-planning module. It often generates instructions either via audio or video devices to guide drivers to their destinations. A human–machine interface permits users to interact with the navigation and information computer and devices. Many key navigation activities must rely on this interface to control the system and to inform the user. A wireless communications subsystem connects the vehicle and its users with a service or a traffic management center that can further enhance the performance and increase the functionalities of the system. Certain technologies of this subsystem can also be used for vehicle-to-vehicle communications. As society continues its rapid advance to an information age, more and more people and their vehicles will depend on wireless technologies to keep them connected with others and to facilitate safe and efficient travel. Recently, Europeans coined a new term for this exciting field: telematics, that is, the use of computers to receive, store, and distribute information over a telecommunication system. Based on whether remote hosts (centralized computing facilities) and wireless communications networks are involved, one can broadly divide the navigation system into either the autonomous system or centralized system. Note that different criteria may result in other classifications of these systems. In autonomous navigation systems, all the navigation capabilities are located solely on the vehicle. The system is responsible for a single-vehicle navigation. In centralized navigation systems, communications networks, host facilities, and other infrastructures work together to navigate. The system is, in general, responsible for multivehicle navigation operations.

In this section subsystem technologies that can be used to construct a vehicle navigation and information system will be discussed. As in the last section, we first present these as separate modules, and then discuss how they are integrated into a working system. Positioning The positioning module is a vital component of any location and navigation system. In order to provide accurate navigation information to the vehicle occupants, the system must determine the coordinates of the vehicle on the surface of the earth (position) or the placement of the vehicle relative to landmarks or other terrain features (location). There are three most commonly used technologies: standalone, satellite-based, and terrestrial radio-based. A typical stand-alone technology is dead reckoning. A typical satellitebased technology is global positioning system (GPS). A typical terrestrial radio-based technology is the ‘‘C’’ configuration of the LOng RAnge Navigation (LORAN-C) system. The principles behind these technologies are discussed below. Stand-alone technology is differentiated from the others in that it does not require a communications receiver to determine vehicle position or location. A very primitive stand-alone technology is dead reckoning, which determines the vehicle location (or coordinates) relative to a reference point. Dead reckoning depends on the system deriving an initial vehicle position. It uses distance traveled and directional (heading) information provided by vehicular sensors to calculate a relative coordinate in two-dimensional planar space (Fig. 2). Relative distance measurements are commonly derived from the vehicle odometer, and directional information is usually provided by a gyroscope or magnetic compass.

(x2, y2) (x0, y0) d0

(x1, y1)

d1

θ1

θ0 Figure 2. Dead-reckoned positioning. It is a method to continuously integrate successive displacement vectors. This method must know the starting position and all previous displacements to calculate the current position. For a vehicle, sensors are used to measure the distance traveled (dn) and direction of travel (␪n) or angular velocity at time tn to a known position (xn, yn). Next, vehicle position (xn⫹1, yn⫹1) is then calculated from these measured data.

108

VEHICLE NAVIGATION AND INFORMATION SYSTEMS

For dead reckoning, the vehicle position (xn, yn) and orientation (␪n) at time tn can be calculated from the equation

xn = x0 +

n−1


di cos θi

i=0

yn = y0 +

n−1


di sin θi

i=0

θn =

n−1


ωi

i=0

where (x0, y0) is the initial vehicle position at time t0, di is the distance traveled or the magnitude of the displacement between time tn⫺1 and time tn, ␪i is the direction (heading) of the displacement vector, and 웆i is the angular velocity for the same time period. In practice, a constant is assumed for the sampling period (positioning data processing cycle of the embedded microcontroller). Due to sensor inaccuracy and the assumption that the heading remains constant over the sampling period, dead reckoning generally accumulates errors as the vehicle travels. This will make the derived position of the vehicle less and less accurate. Many techniques are available to eliminate the accumulative errors, such as map-matching algorithms, short-range beacon networks (signpost), and complementary or redundant sensor compensation. Some of them are discussed in later sections. There are a variety of sensors available to detect the distance traveled and direction of the vehicle (1,5–7). In general, surface vehicles are much less expensive than space vehicles or marine vehicles. Therefore, designers of the dead-reckoning-based system for the surface vehicle usually select lowcost sensors. There are relative sensors and absolute sensors. Relative sensors measure the distance or directional change based on predetermined or previous measurement. Absolute sensors provide the distance and directional data of the vehicle with respect to a coordinate system affixed to the earth. For automotive location and navigation, transmission pickup sensors and wheel sensors are used for relative distance measurement. Low-cost gyroscopes and electronic magnetic compasses are utilized for direction measurement. As their names indicate, transmission pickups and wheel sensors obtain their measurements from different places in the vehicle, one from the transmission shaft and the other from the wheel shaft. Their operational principles are very similar; they convert mechanical motion into electronic signals by measuring the angular position of the shaft. Variable reluctance, Hall effect, magnetoresistive, and optical technologies are among many distance sensors that can be used for the measurement. People refer to such distance-measuring sensors as odometers. Obtaining the number of pulse counts per revolution and a proper conversion scale factor, the output of the sensor can be converted into distance traveled. Commonly used sensors are variable reluctance and Hall effect sensors. They are both electromagnetic pulse pickups, which consist of toothed wheels (exciter rings) mounted directly on the rotating component. As the toothed wheels rotate, they produce voltage waves in a sensing circuit as the teeth pass a magnet. The variable reluctance sensor uses a permanent magnet with a wire coil wound around it, which typically produces a sine wave. The Hall effect sensor uses a probe with a biasing

magnet and a circuit board, which typically produces a square wave. These waveforms can be easily converted to the number of pulse counts per revolution. Due to its low cost and relative reliability, the variable reluctant sensor is still the most popular wheel sensor for the antilock braking system (ABS). Combining the outputs from two distance sensors (odometers), one each for a pair of front or rear wheels, a differential odometer is formed. It can obtain both relative distance traveled and heading change information. Knowing the initial vehicle position, the traveled distance in the current sampling period is determined by averaging the left and right wheel rotation counts during the period and multiplying by a proper scale factor. The travel direction change in the current sampling period is determined by the difference between the counts for the left and right wheels multiplied by the same scale factor and divided by the axle length. Gyroscopes are relative direction sensors. A popular lowcost gyroscope is the vibration gyroscope. A vibration gyroscope measures the Coriolis acceleration that is generated by the angular rotation of a vibrating bar or fork. The acceleration can be represented by the Coriolis force. This force is detected and converted to a voltage by the detector element attached on the vibrating bar and a simple circuit. Since the mass of the bar and the vibrational velocity of the bar are known quantities, the angular velocity can be easily calculated and provided to the user. Integrating this angular velocity over time, the change in the vehicle direction is obtained. Compasses are absolute direction sensors that measure the earth’s magnetic field. For on-board navigation applications, an electronic compass is better than a conventional one because of its quick response, portability, and high-vibration durability. The fluxgate compass is commonly used. Its measurement is obtained from the gating action imposed by an alternating current (ac)-driven excitation coil, which induces a time-varying permeability in the sensor core. Permeability is the property of a magnetizable substance (often made by a core) to modify the magnetic flux of the surrounding magnetic field. These varying flux lines induce positive and negative electrical current spikes in a sense coil. Two orthogonal sensing coils can be configured in a symmetrical fashion around a common core. The integrated direct current (dc) output voltages (obtained from the spikes) of these two orthogonal coils are then converted to an angle by taking the arctangent of the quotient. Since the compass itself cannot distinguish between the earth’s magnetic field and other magnetic fields present, algorithms must be applied to extract the earth’s magnetic field from the measurement. These other magnetic fields include those of the vehicle and of nearby objects, such as power lines, big trucks, steel structures, and reinforced concrete buildings and bridges. Nearby operational devices, such as the rear-window defroster or automatic car wash brushes, can also change the compass measurement temporarily. Satellite-based technology uses satellites emitting radio signals to the receiver to determine the position of the receiver, often on the surface of the earth. A satellite system typically consists of a space segment (satellites), a user segment (receivers), and a control segment (monitor and control stations) as shown in Fig. 3. Since the early 1990s, the satellite-based GPS receiver has become a dominant device for vehicle navigation and information systems. It provides an affordable means to determine

VEHICLE NAVIGATION AND INFORMATION SYSTEMS

Uplinks

Downlinks

Monitor and control stations Transmitter

Receiver on board

Figure 3. Satellite-based positioning. This is a typical GPS-based positioning system. The four satellites shown here are a part of a space segment, which emit radio signals from space. A GPS receiver on the vehicle is a part of a user segment, which receives the radio signals to calculate its position. The monitor and control stations are a part of a control segment, which control and monitor all the satellites in the system. These three segments make up a satellite system.

109

SA on). Combining GLONASS with GPS as a dual receiver should provide position fixes most of the time, even in urban canyons where tall buildings are crowded in dense areas. The dual receiver can take advantage of 48 satellites instead of only 24 for a receiver based on GPS satellites alone. Other existing and planned systems include many low-earth-orbit (LEO), medium-earth-orbit (MEO, or ICO in Europe), geosynchronous (GEO) satellite systems, and DGPS-based systems, such as INMARSAT, OmniTRACS, Orbcomm, Iridium, Globalstar, ICO, wide area augmentation system (WAAS, US), European geostationary navigation overlay system (EGNOS, Europe), and multifunction transport satellite (MTSAT, Japan). Terrestrial radio-based technology typically uses base stations or devices emitting radio signals to the mobile receiver to determine the position of its user. Signals can also be emitted from the mobile device to the base stations. Commonly used techniques are short-range beacon positioning, polar positioning, angle of arrival (AOA) positioning, time of arrival (TOA) positioning, and time difference of arrival (TDOA) positioning. The latter four are shown in Fig. 4. All these methods

Possible location

α

d Site (a)

position, velocity, and time around the globe (8,9). Its distance and direction measurements are absolute (with respect to the earth). GPS was developed and is maintained by the US Department of Defense. Its constellation consists of 24 satellites orbiting at an altitude of 20,183.61 km above the earth’s surface (equatorial radius of 6378.137 km). The positioning measurements of the GPS receiver are based on the time of arrival (TOA) principle. When four or more satellites are in the line of sight of the receiver, the latitude, longitude, and altitude of the receiver are determined. Two GPS service levels are provided. Standard positioning service (SPS) is for civilian users and precise positioning service (PPS) is for military users. The SPS is deliberately degraded by selective availability (SA). As documented, SPS provides horizontal position accuracy within a circle of 100 m radius 95% of the time (10). On the other hand, PPS provides horizontal position accuracy within a circle of 21 m radius 95% of the time. In 1996, the US government decided conditionally to phase out SA placed on SPS starting in the year 2000, subject to annual review by the US president. Even with SA, a much better accuracy can be obtained by using differential correction techniques. Differential GPS (DGPS) can reduce the position error to under 15 m, while SA is in effect. It uses a master receiver at known (surveyed) coordinates to send correcting information to a mobile receiver over a communications link for deriving a more accurate position. Typically, the receiver separation of DGPS is up to 50 km. In addition to GPS, there are other satellite-based systems. The most notable is the global navigation satellite system (GLONASS). GLONASS was developed by the former Soviet Union and is now maintained by Russian Military Space Forces (11,12). The system does not have any measures to intentionally degrade system accuracy. Therefore, it provides a better position fix to civilian users than does GPS (SPS with

Possible location

(b)

Possible location Site 1

α1 α2

Site 2

(c)

Possible location d1 d3

d2 Site 2

Site 1

h2 Site 3

h1 (d)

Figure 4. Terrestrial radio-based positioning: (a) Polar positioning, (b) time of arrival (TOA) positioning, (c) angle of arrival (AOA) positioning, and (d) time difference of arrival (TDOA) positioning. Multiple radio transmitters and receivers are used to determine the location of a mobile device, which can be attached to a human, a vehicle, or other objects. One or more of these four positioning technologies may be utilized.

110

VEHICLE NAVIGATION AND INFORMATION SYSTEMS

require communications transmitters, receivers, or transceivers. In other words, they depend on emitting and receiving radio signals to determine the location of an object to which a simple reflective element, a receiver, or a transceiver is attached. To make the position determination, applications generally require that one end of the positioning system is fixed and the other end is moveable such as a mobile device attached on the vehicle. For performance improvement, hybrid methods (various combinations of the techniques discussed or with additional techniques) have also been used. The short-range beacon system determines the position of the mobile device by sensing the device close to a reference point. Hence, the name of proximity beacon system is used in some of the literature. Since the beacon head of this system is sometimes installed on a signpost along the road, some people refer to it as the signpost system. Beacon systems do not always restrict themselves to signposts. One example is the loop detector, which is embedded in the road surface. Popular short-range beacon systems tend to use microwave or infrared frequency as their communication medium. Despite its robustness and low-cost mobile device, the beacon system has limited communications zones, discontinuous communication, as well as high installation and maintenance costs. The polar system determines the position from a single base station that measures the distance to and the direction of the mobile device. The distance can be derived from the round-trip time to the device, which defines a circle around the station. The direction measurement forms a radiation line. If these two measurements are error free, their intersection defines the position. Despite its simplicity, for a long radius situation, even a relatively small error in a direction measurement can produce a very large error in geometric accuracy. The time of arrival (TOA) system determines the position based on the intersection of the distance (or range) circles. The range is calculated from the signal transmission time, that is, multiplying the time by the speed of the signal. Two range measurements provide an ambiguous fix and three measurements determine a unique position. Within the triangle formed by the centers of the three circles, geometric accuracy is the highest. Away from the triangle, the accuracy gradually decreases. The same principle is used by GPS, where the circle becomes the sphere in space and the fourth measurement is required to solve the unsynchronized-receiver-clock bias. Because of this unsynchronized satellite and receiver clock issue, the signal transmission time determined by the GPS receiver is not accurate so the actual measurement is a pseudorange measurement. The angle of arrival (AOA) system determines the position based on triangulation. The intersection of two direction measurements defines a unique position. When the two directional lines cross at right angles, geometric accuracy is the best. When they cross at 180⬚ angle as a connected straight line, the accuracy is the worst. This technique requires only two stations to determine a position, but is susceptible to signal blockage and multipath reflection. Furthermore, a phased array of antennas is needed, which adds additional cost to the system without the phased array. The time difference of arrival (TDOA) system determines the position based on trilateration. This system is commonly referred to as the hyperbolic system, where a time difference is often converted to a constant difference in distance to two

base stations (as foci) to define a hyperbolic curve. The intersection of two such curves defines the position. Therefore, two pairs of base stations [at least three, as in Fig. 4(d)] are needed for positioning. Geometric accuracy is a function of the relative base station locations. Use of the TDOA technique for real-time location calculations requires a synchronized time among all base stations. On the other hand, it requires fewer antennas and is less susceptible to signal blockage and multipath reflection than using AOA. As discussed above, many sensors can be used for positioning. However, no single sensor is adequate to provide position data to the accuracy often required by a navigation and information system. The common solution is to fuse data from a number of different sensors. These sensors usually have different capabilities and independent failure modes. People often refer to this technique as sensor fusion, which can provide the system with complementary, sometimes redundant data for its navigation and information task. Typical methods include various Kalman filters and other filters, such as lowpass, high-pass, and complementary filters (1). Fuzzy-logic, neural network, statistical decision, probability reasoning, and many other inference methods are also candidates for sensor fusion, as long as they are efficient enough for actual implementation in an embedded real-time system. Digital Map Database A digital map database is a module that provides map-related functions. In addition to presenting rich spatial information, as does a paper map, a digital map can be manipulated to support many navigation activities, such as locating an address or destination, calculating a travel route, guiding along a precalculated route, matching a vehicle trajectory with a known road, and providing travel information. Two types of digital maps are used for computers: rasterencoded maps and vector-encoded maps. These maps are digitized from various paper maps, aerial photographs, census bureau data, field data, and many other sources. A raster encoded map stores in each pixel the value of each parameter of interest in a matrix over space and is digitized using a scanner. The map is displayed on the video screen systematically; the electron beam repeatedly paints a series of thin and horizontal stripes while moving from the top to the bottom of the screen. A vector-encoded map stores in a data structure the value of road network features using Cartesian geometry and it is digitized using a digitizer. The map is displayed on the screen unsystematically; the electron beam traces the outlines of the map directly, one line segment at a time. Because it requires less storage space and is easier to manipulate, the vector-encoded map is very popular in navigation. The road network features are typically represented by one or more primitives: points, lines, and polygons. These primitives are encoded in a computer data structure as node (point), segment (line), and area (polygon) records, together with their respective attributes. Points, lines, and areas are graphic information. They are stored as coordinates, symbols, and rules. Attributes are nongraphic information. They are stored as alphanumeric characters to characterize, qualify, and link the graphic map features with their appropriate spatial locations. Examples of attributes are speed limit, street name, address range, road type, driveability, area name, city name, city range, state name, and zip code. The digital map

VEHICLE NAVIGATION AND INFORMATION SYSTEMS

used by vehicular navigation and information systems can also be viewed as a special subset of a geographic information system (GIS) (13), but with additional navigation attributes. GIS is a computer system used to support the capture, management, manipulation, analysis, and display of spatially referenced data for solving complex planning and management problems. The map used in turn-by-turn navigation applications, also referred to as a navigable map database, has great accuracy and includes attributes typically not included (or not as good) in GIS, such as nearly flawless road connectivity, one-way streets, and turn restrictions. A map represents the geometry of the surface of the earth. The earth is an oblate ellipsoid. To represent this complex spatial information on a two-dimensional computer screen as a digital map, one must know the datum, projection, and production techniques used to make it happen. Otherwise, one may encounter some unexplainable problems later on in the design and development of the navigation and information system. A datum is a set of parameters to define the location and orientation of the reference ellipsoid used to model the earth. There are global and local (regional) datums. Local datums are developed to match, as closely as possible, within the regions under consideration. Gobal datums are triggered by the earth-orbiting satellite technology. Unlike most local datums, the origin for global datum is at the center of the earth. In practice, sources for digitized maps are typically based on local datums. Popular satellite-based positioning systems rely on global datums. For instance, many North American maps use local datum North American Datum (NAD) 83 or NAD 27. GPS uses global datum World Geodetic System (WGS) 84. Blindly mixing the coordinates derived from these different datums will result in a displacement of up to 1,500 m or even more for a point on the earth’s surface. More specifically, latitudes and longitudes referenced to different ellipsoids define different coordinate systems and cannot be mixed. Transformation equations between different ellipsoidal coordinate systems or between ellipsoidal and Cartesian coordinates systems are available (1,14). They are also included in some application software. A projection is a technique to transform spatial data into a planar representation. For centuries, people have had difficulty representing all the features of the earth in their true relationship to each other. Projection of three-dimensional geographical information onto a two-dimensional flat plane inevitably introduces distortions to the map. Various projection techniques have been developed over the years but none of them can exactly model the reality. One can only preserve the angles (true directions) or the areas (a constant scale for distances). A valuable projection is the Mercator projection that preserves the angles. Its popular modification is the universal transverse Mercator (UTM) projection. UTM is often used for map projections involving large countries or continents. For instance, many paper maps used in the United States for local surveying and other mapping operations are based on this projection. There are many production techniques used to generate the original source materials used to digitize the map. These techniques are beyond the scope of this article and will not be discussed here. For producing a vector-encoded map, after the map source collection, these data need to be digitized, validated, and updated. The whole process is a tedious and labor-

111

intensive effort. A complete digital map database can be stored on a computer hard disk, a large Flash ROM, a CDROM disk, or a PCMCIA (PC) card. When employing map data sources, one must convert them into one unified coordinate system based on one preselected datum or reference ellipsoid. From the discussion in the beginning of this section, it is already known how to produce a digital map based on these sources. There are no widely accepted standards available to define the format of the results of either the scanning or digitizing process. For the vector-encoded maps used in navigation, quite a few standards have been proposed or are being proposed. They include the geographic data files (GDF), Japan Digital Road Map Association (JDRMA; also as an organization), spatial data transfer standard (SDTS), physical storage format (PSF) and application program interface (API). It is expected that a universal standard will be available in the near future for navigation software to interact with map databases stored on different media. For the moment, a common practice is for a map database vendor and its customers to agree on a (proprietary) software interface. Map Matching Map matching is a computer algorithm that utilizes a digital map to make the position determination more reliable and accurate. Unlike navigation in air and sea transportation, land transportation vehicles are basically constrained to a finite network of roads with only occasionally excursions into parking lots, driveways, or other off-road conditions. This makes it possible to correlate the trajectory of the vehicle with the road on which the vehicle travels. For fast processing and easy analysis, a vector-encoded map is commonly used. Furthermore, since the map-matching-assisted positioning systems depend heavily on the map, the accuracy of the vehicle position derived should be similar to that of the digital map database. A typical map precision requirement for urban areas is within 15 m of ‘‘ground truth.’’ All three positioning techniques discussed above can be assisted by map-matching algorithms. As an example, dead reckoning tends to accumulate errors with additional distance traversed. As time proceeds, the actual vehicle position will not agree with the dead-reckoned vehicle position that may already have drifted away from the road the vehicle is on. By comparing the vehicle trajectory with road segments stored in the digital map database, the error can be corrected by snapping the dead-reckoned position back to the matched position on the road of the map. The basic principle of the map-matching algorithm is to identify the road segment on which the vehicle is traveling and determine the position of the vehicle on this segment to update the position derived by a positioning subsystem. In particular, a possibility is assigned to each candidate road segment. Low-possibility candidates are removed and the highest-possibility candidate is retained. This candidate is presented to the system to determine the vehicles position. Therefore, any inference method can be used to evaluate the possibility, as long as it is fast enough for real-time execution. Two methods have been used and proposed: probabilistic and fuzzy-logic-based. The probabilistic method evolved from the semideterministic method (15). This conventional algorithm requires that the positioning errors be statistically propagated into the posi-

112

VEHICLE NAVIGATION AND INFORMATION SYSTEMS

tioning determination. These errors and error models are used to define confidence regions. It is assumed that these regions may contain the actual vehicle locations. If only one road is contained in the region, a matching segment is found. If more than one road is within the region, road segment heading, connectivity, and closeness are compared against similar data received from the positioning subsystem and previous processing cycles. Finally, one most probable road segment is identified and presented to the system, along with the most likely vehicle position on that segment. Since none of the unconventional algorithms, such as the fuzzy-logic-based algorithm, have been deployed in commercial products, they will not be discussed further here. For more information about these algorithms, readers may refer to Refs. 1 and 16. Besides matching, the digital map database can also be used to calibrate positioning sensors. For instance, the road segment length can be compared against the distance detected by the distance sensor. The road segment direction can be compared against the direction detected by the direction sensor. Once the error is above a certain threshold, corresponding correction and calibration can be done. Because many sensors could be affected by their working environments, recalibration becomes necessary. With a map, this can be done dynamically during the journey. The map and deadreckoning sensors can also be used to detect GPS blunders caused by erratic satellite signals and fill the voids caused by signal blockage. There are many other techniques to utilize maps or other means to improve the positioning subsystem. In summary, a digital map is a good complement to other positioning technologies and has been demonstrated to produce much more robust position performance. Route Planning Route planning encompasses planning a route prior to or during a journey. It can be classified as single-vehicle route planning or multivehicle route planning (1). The former plans one route for a single vehicle based on the current location and a single destination (or multiple destinations). The latter plans multidestination routes for all vehicles on a particular road network. This classification is analogous to the single-source shortest path and all-pairs shortest path problems discussed in the computer science literature (17). Travelers often prefer different route optimization criteria. Some may want to follow a shortest-distance route. Others may opt to minimize the travel time. Some may prefer to avoid expressways, minimize toll charges, or to impose a limit on number of turns and traffic lights during the trip, and so on. All these factors are referred to as the travel cost. The selection of optimization criteria can be done either by the system itself during execution or by the user prior to the planning. Actual planning needs to utilize a digital map. The travel cost must be obtained from the attributes in the digital map. For instance, distance can be derived from the road segment length. Travel time can be derived from the division of the segment length by the speed limit. As defined in the map, a street intersection or a dead-end is defined by a node, and a piece of roadway between two nodes is defined by a segment. For single-vehicle route planning, the most popular algorithm for a single-destination solution is the heuristic search algorithm. For multiple destinations (vehicle routing), Dijkstra’s algorithm, traveling-salesman algorithms, genetic algo-

rithms, tabu searches, robust algorithms, and other algorithms can be used. Heuristic search has information concerning the number of steps or the cost from the initial state and current state to the destination state. One very popular heuristic search is the A* algorithm. Unlike Dijkstra’s algorithm, which finds the optimal route from the origin node to every other node in the road network, the A* algorithm finds the optimal route from the origin node to the destination node while saving computation time and memory space. It determines which is the ‘‘most promising’’ node, which successors to generate, and which irrelevant search branches to prune. The decisions are made based on the heuristic information, which provides an estimate of how far a node is from the destination node. From this information, the algorithm decides the likelihood of a particular node on the best solution route and which nodes to search. It has been demonstrated that if the heuristic evaluation never overestimates the actual cost of reaching the destination node, the A* algorithm will find the optimal solution route, provided that one exists. Generalization of the search algorithms leads to the bidirectional search algorithm. The A* algorithm usually provides a solution route from a given origin node to a destination node, which is a forward search. Reversing the search origin and destination, it becomes a backward search. If both forward and backward searches are conducted at the same time, the algorithm becomes bidirectional. When implemented on a sequential machine (single processor), the bidirectional search must switch repeatedly between the forward and backward searches. Therefore, it needs to determine the alternating period and stopping criterion to force the two searches to meet somewhere in the middle of the prepared route. If these two criteria are chosen properly, a bidirectional search may require less search time and space than a unidirectional search. Because of the hierarchical nature of roads, a popular technique is to switch among different layers of a map that is constructed on the basis of the road hierarchy to make the search more efficient. The algorithm is named as the hierarchical search algorithm. It has been demonstrated that this algorithm can reduce the exponential complexity of the search to linear complexity. A requirement for this algorithm is that the digital map itself must be constructed hierarchically. For instance, one could make a four-layer map as follows: Layer 1 includes all the roads; Layer 2 includes collector roads, arterials, and highways; Layer 3 includes arterials and highways; Layer 4 includes highways only. In this way, one can search major roads first, starting on Layer 4, and then fills in the details for any necessary portion of the route. Note that before reaching Layer 4, the search algorithm may need to determine the nearest starting and ending nodes on that layer from given origin and destination pair in Layer 1. For route planning in a very large map database, the bidirectional heuristic-based hierarchical search is the most popular algorithm. For multivehicle route planning, there are many good candidates among the algorithms proposed in all-pairs shortest path problems, dynamic programming, and operations research. This requires a central service (control) center to provide such service to the vehicle on the road via a communications network. The algorithm often computes optimal or nearoptimal routes for every possible (or selected) origin-and-destination (O-D) pair or designated-zone pair in the road net-

VEHICLE NAVIGATION AND INFORMATION SYSTEMS

work. For on-demand calculation, the route computation is done for each driver on request based on the O-D pair submitted. For periodic calculation, the route computation is done periodically for all (or selected) O-D pairs or for designated zone pairs. Route Guidance Route guidance directs the driver along the route generated by the route-planning subsystem. There are pretrip guidance and en route guidance. En route guidance can provide turnby-turn driving instructions in real time and is much more useful than pretrip guidance (1). To complete an en route guidance, the system requires a navigable map database, an accurate positioning, and the planned route. For real-time route guidance, the system needs to monitor the current vehicle position and heading, and compare them with the best route generated by the route-planning subsystem. As a turn or maneuver approaches, the guidance system informs the driver with visual signals, audible signals, or driving instructions. Visual signals shown on a screen-based system could be in one of the two display formats: turn arrow or route map. The turn-arrow format displays a limited amount of information such as a turn arrow, the shape of the next intersection, and the distance to the next intersection along with a countdown bar. The route-map format displays a detailed map overlaid with the planned route. Audible signals are used to alert the driver that the system is ready to announce an instruction. These instructions are often generated from a speech synthesizer or a prerecorded set of voice messages. A typical approach is to announce an ‘‘early’’ instruction such as ‘‘Drive 5 kilometers to Main Street’’ after finishing a previous maneuver. When moving close to the next maneuver, announce a ‘‘preparing’’ instruction such as ‘‘Right turn half a kilometer ahead, bear right,’’ to tell the driver to move to the appropriate lane and to begin the maneuver preparation. Once the vehicle is very close to the maneuver, inform the driver with an ‘‘approaching’’ instruction to tell the driver to perform a maneuver such as ‘‘Turn right at the traffic light onto Main Street.’’ The last instruction is the most critical one. Per common sense, the guidance system must avoid announcing upcoming maneuvers either too early or too late. Otherwise, it would be difficult to convince drivers that this system could increase safety and improve performance. There are static route guidance (or navigation) systems and dynamic route guidance (or navigation) systems. The static navigation system calculates the route based on historical traffic information and precollected map data. The historical traffic information may include average daily travel time for each road, turn delay data for each intersection, and other collected data. All these data can be stored in the digital map as attributes. The static system cannot respond to unpredictable road conditions. These conditions may include traffic incidents, road congestion, and road closure. Conversely, a dynamic navigation system calculates the route based on the real-time traffic data. It can respond properly to real-time traffic conditions. To receive real-time traffic data, there must be a wireless communications network to connect the vehicle with a service center (or traffic management center). The realtime data processing could be either in the vehicle or in the center. The network can take different forms, as will be discussed later. Real-time traffic data collection can also take

113

different forms, such as from loop detectors, video cameras, short-range beacons, traffic reports, weather data, and equipped vehicles (as probes). Human–Machine Interface The human–machine interface enables the user to interact with the navigation and information computer and devices. It is a special challenge to design and implement a human– machine interface for vehicular applications (1,18,19,20). Surface vehicles, especially automobiles, have a very diverse user community. Crowded instrument panels (or dashboards) create a highly distracting environment. Any on-board interface equipment must operate under environmental constraints such as varying light conditions, varying temperature, vibration, traffic noise, and contamination by dirt or grease. In addition, the interface presents the vital first impression to the potential vehicle owner. Safety and ease of use are the most important design considerations for the human–machine interface. Two important technologies for the in-vehicle human– machine interface are visual display and voice (input/output). For a visual display, there are many control and display technologies available. Control devices include buttons, joysticks, switches, knobs, touch screens, and voice. Voice-based interfaces are discussed shortly. One interesting control device is the touch screen. It works as a multifunction transparent switch to reduce the number of input and output devices. The touch screen is operated by the touch of a finger or stylus. It can be overlaid on any flat panel device discussed below. Display technologies include the cathode-ray tube (CRT), electroluminescent display (ELD), heads-up display (HUD), lightemitting diode (LED), plasma display panel (PDP), and vacuum fluorescent display (VFD). Recently, the field-emitter display (FED) has appeared on the market. It looks sharp from any angle, has better or equivalent brightness, contrast, and resolution, and is thinner and lighter compared with popular LCDs. In addition to the selection of control and visual display technologies, there are many other design considerations. One of them, discussed earlier, is the route map versus the turn arrow. Studies have shown that the route map makes the driver glance more often at the display screen than the turn arrow, suggesting that it is too much of a distraction while driving. The turn-arrow display, augmented with voice guidance, could address this problem. Other design considerations are the use of an interlocking mechanism, a dimmer mechanism, and a rotational mechanism. The interlocking mechanism automatically freezes most of the control functions while the vehicle is moving. The dimmer mechanism allows the driver to control the screen brightness. The rotational mechanism is for adjusting the limited-viewing-angle display device. Additional design considerations include, but are not limited to, reducing the glance number and duration on the display screen, when, where, and how the user should be informed of an upcoming turn, what minimum text size should be used, and how the intersection should be portrayed. In the future, reconfigurable steering wheels and reconfigurable displays may be used to reduce the number of control and display devices required. Techniques to monitor the driver’s state will be deployed. These may include monitoring the driver’s eye and foot movements or the driving behaviors in order to pre-

114

VEHICLE NAVIGATION AND INFORMATION SYSTEMS

vent accidents. A warning can then be fed back to the driver, other vehicles, or traffic personnel in the surrounding area. For voice-based interfaces, the basic technologies used are speech synthesis and speech recognition. Speech synthesis and speech recognition are suitable for output and input, respectively. Speech synthesis is a computer-aided technology that translates text into understandable human language. It is a stable technology, with error rates as low as 3.25% for individual word perception. However, the available synthesizers still do not sound natural, although progress in synthesizers appears to have reached an asymptote. Some navigation systems have chosen to prerecord human-voice driving instructions and replay them during the guidance. However, this method is not flexible enough to adapt to varying environments and may take more memory space to store voice instructions. Speech recognition is a computer-aided technique that understands human language. It is a useful technique for hands-free control while driving. The available technologies are speaker-dependent and speaker-independent with isolated-word, connected-word, and continuous-speech capabilities. Speech recognition is still not suitable for critical inputs. In general, speech recognition products are not as robust as speech synthesizer products of comparable complexity and cost. In particular, speech recognition needs further improvement in order to recognize voice commands in noisy environments. Once the word error rate has been reduced below 5% in its intended environment, a speech-recognition product is ready to be used. Clearly, future drivers will be much better informed. Safety and ease of use will be even more important for the future human–machine interface. Wireless Communications Reliable communication has gradually become an integral part of vehicle navigation and information systems. Wireless

communications permit many quality services to be provided to drivers (1,21,22). Vehicle users can contact each other, traffic management centers, or service centers, and receive services not available to them in the past. The vehicle has been evolving from a transportation tool to a multipurpose tool that combines transportation, information, and mobile computing. Future vehicles may be connected with the Internet, electronic-mail (e-mail), fax, video conferencing, and remote diagnosis and provide games and movies for rear-seat passengers. Vehicle communications require an infrastructure for voice and data that can reliably and efficiently deliver real-time traffic reports and other information. These information exchanges could be between an infrastructure (base stations) and mobile devices (portable radios) or between mobile devices themselves. The mobile device can be attached to the vehicle or to the person as a wearable item. For an ideal deployment, certain key communications network attributes are necessary. These include excellent coverage, high capacity, low cost, full connectivity, and secured access. There are many different communications technologies available and a variety of applications to apply them to, such as traffic management, emergency management, intermodal travel, public transportation operations, commercial vehicle operations, electronic payment, and advanced (or intelligent) vehicle control. Each of these applications has specific needs that may likely be satisfied by a single communications technology. On the other hand, a regional or statewide communications network is likely to require a hybrid implementation of communications technologies. As an initial guide, a highlevel overview of existing wireless technologies is presented in Table 1 (1). Some of communications techniques are now briefly discussed. Paging was originally developed as a one-way commu-

Table 1. Communications Technology Matrix

Technology

Reliability

Coverage

Avg. Data Transfer Rate

Equipment and Airtime Costs

Security

Simplex or Duplex

Paging

E

Metro, some rural

2.4–3.6 kbps

E

E

Simplex/duplex

Cellular PCS Private land mobile radio systems Radio data networks Broadcast subcarriers Short-range beacons Satellites Cordless telephony Radio LAN networks Infrared LAN networks Meteor burst

F-E E VG-E

Metro, some rural Metro Metro, rural

9.6–64 kbps ⱖ8 kbps 1.2–64 kbps

F-E E F-E

P-E E P-E

VG-E

Metro

2.4–256 kbps

F-E

E

Metro

1.2–19 kbps

E

1–100 m

E E VG-E

Worldwide Near the base Indoors/outdoors; 40–11,263 m Indoors/outdoors; 9–6,436 m Worldwide Metro, rural; 40,000 m hops

Microwave relays

E F VG-E

Real Time or Store-andForward

Access Public

Duplex Duplex Simplex/duplex

Store-andforward/ Real time Real time Real time Real time

E

Duplex

Real time

Public

E

E

Simplex

Real time

Private/public

64–1,024 kbps

F

VG

Simplex/duplex

Real time

Private

2.4–64 kbps ⱖ32 kbps 64–5,700 kbps

P-E VG VG

E E E

Duplex Duplex CSMA/CD

Private Public Private

1,000–10,000 kbps 2–32 kbps

F-E

E

CSMA/CD

E

E

Duplex

8,448–250 ⫻ 1,544 kbps

VG-E

VG-E

Duplex

Real time Real time Near real time Near real time Store-andforward Real time

Public Public Private/public

Private Private Private

P ⫽ Poor; F ⫽ Fair; VG ⫽ Very Good; E ⫽ Excellent; Metro ⫽ Metropolitan Areas; CSMA/CD ⫽ Carrier Sense Multiple Access with Collisions Detection.

VEHICLE NAVIGATION AND INFORMATION SYSTEMS

nications service. Recently, two-way paging has been introduced to the market, which is better for vehicular communications. Paging is mainly an urban service and traditionally a one-way service of short data messages. One interesting application of paging is to superimpose paging signals on AM/FM radio or TV signals (broadcast subcarrier) to transmit short traffic messages. Because of possible delay of queued messages, paging networks may not be suitable for vital services requiring immediate response. Cellular technology has been rapidly moving to the digital domain with the channel access methods of GSM, CDMA, and TDMA leading the way. For analog networks, modems are used to modulate circuit data over cellular links. Cellular was originally designed for one-to-one voice communications. Recent developments have enabled cellular networks to carry more data. This will enhance their role in vehicular applications. A network overlaying the existing analog cellular network (CDPD; see below) has been developed to address packet data applications. Personal communications services (PCS) are newcomers. They are proposed as a multienvironment, multioperator, and multiservice infrastructure. The basic concept is to provide highly competitive wireless communications services without any restrictions on providers’ capabilities. Like cellular and other handset-based technologies discussed below, PCS handsets can be used by drivers to obtain traffic information and other services. Private land-mobile radio systems include conventional, trunked, and specialized mobile radio (SMR). Distinction between public and private systems has become increasingly blurred. Some recent systems intend to serve both the private and public sectors, such as the integrated Dispatch Enhanced Network (iDEN). iDEN is a digital communications system that integrates voice dispatch, wireless phone, paging, and data-transmission capabilities into one handset with cellularlike coverage. Traditionally, private land-mobile radio systems have been applied to many transportation, public safety, and industrial applications. Popular radio data networks (RDN) include ARDIS, Mobitex (or RAM Mobile Data in the United States), cellular digital packet data (CDPD), and Ricochet. Both ARDIS and Mobitex were developed for short wireless message communications. Ricochet has less coverage than both ARDIS and Mobitex. It uses an unlicensed band, which may not be suitable for mission-critical applications because it must yield to licensed users when an interference occurs. As mentioned above, CDPD is a system that overlays existing analog cellular networks. The National ITS Architecture Team of the US Department of Transportation selected CDPD as the preferred network for providing traffic and route guidance information to motorists, out of those available now or near-term in the US. This technology is suitable for applications requiring wireless short messages, burst messages, or lengthy sessions, which can afford sporadic communications. Broadcast subcarriers provide communications from a traditional broadcast station (AM, FM, or TV) without a specially allocated frequency spectrum. Unlike receiving the traffic data directly from AM/FM radio such as the highway advisory radio (HAR), subcarriers multiplex traffic data over regularly broadcast signals on a sideband frequency for textual display on specially designed radios. The most popular system is the radio data system (RDS, or RBDS in the United

115

States). A special channel, the traffic message channel (RDSTMC), has been assigned to handle traffic data. Transmitted traffic messages can be converted into a selected language for display in textual format on the radio. RDS can also handle location and navigation messages such as DGPS correction information. Short-range beacons provide a communications link between vehicles and road infrastructures. Applications include proximity positioning, travel information, dynamic navigation, electronic toll collection (ETC), electronic road pricing (automatic fee collection), parking management, fleet management, and many others. Beacons can also be used as complementary devices to correct dead-reckoning positioning errors. Beacon systems typically depend on dedicated shortrange communication (DSRC) protocols. Like the digital map, there is no universal agreed upon standard available for DSRC although international standard organizations are actively working in that area. Beacon communications provide high transmission rates, effective position calibration, location-oriented traffic information, and the ability to detect the vehicle parameters such as types. However, beacon communication coverage zones are very limited and communication is not continuous over time. System installation and maintenance are very costly. SYSTEM TECHNOLOGIES This section covers how the modern technologies discussed earlier can be used to construct a vehicle navigation and information system. There are very primitive systems that only show the map of the road network, or a moving dot representing the vehicle or the mobile device on the map. There are advanced systems that can automatically generate an optimal route to guide the vehicle or mobile device user. To facilitate this discussion, navigation systems are divided into two categories: autonomous and centralized. An autonomous system provides all the navigation activities on board the vehicle. A centralized system needs assistance from a centralized facility to conduct its navigation activities. To design and construct a sophisticated navigation and information system, a good system architecture is very important. The system architecture determines the allocation of functions to specific subsystems and provides foundation for interface standards. It is made up based on its intended working environments. Different subsystem hardware and software components must be harmonized to work in these environments. Besides meeting the current system requirements and specifications, a good architecture can provide a stable basis for the future evolution of the system. Knowing the principles of each individual subsystem (or module), one can integrate them together into a working system under the overall system architecture. Autonomous Navigation Systems An autonomous navigation system provides guidance and information to the user, based solely on its on-board processing capabilities and data resources. To provide navigation, accurate vehicle location is the essential component. As discussed earlier, GPS is the dominant technology for vehicle location determination. However, in addition to the effect of SA and receiver noise, GPS signals are subject to

116

VEHICLE NAVIGATION AND INFORMATION SYSTEMS

multipath and blockage problems. To compensate for these problems, dead-reckoning sensors are commonly used. Integration of GPS and dead-reckoning sensors requires sensor fusion technologies such as simple low-pass and high-pass filters, complex Kalman filters, and many other fusion techniques mentioned in subsystem technologies. In addition, map matching can be used to provide additional location accuracy if a digital map is available. Table 2 (1) presents a comparison of various location technologies. Some of them have also been used in centralized systems, discussed in the next section. To provide visual aids, a human–machine interface can be integrated into the vehicle location system. It can be as simple as a black-and-white text-based display or as complex as a color pictorial display that interacts with a digital map database. The interfaces can use various input and output devices. Popular ones are switches, buttons, knobs, voice, and touch screens. Vehicle navigation systems consist of some or all the subsystems (or modules) depicted in Fig. 1. A complete autonomous system diagram is shown in Fig. 5. Besides algorithm design, software implementation must deal with the real-time and embedded system issues. These include time, memory, and device-driver management, system synchronization, interprocess communications, mutual exclusion, error handling, multitasking execution, and so on. In addition, hardware designs post a great challenge for the in-vehicle components. They must meet the rigid specifications of the automotive environments such as vibration tests, shock limits, humidity limits, and temperature limits, normally ⫺40⬚ to ⫹85⬚C (⫺40⬚ to 185⬚F). As the progress of vehicle navigation systems continues, trends indicate that many navigation and entertainment components will be shared. All the components in the shaded blocks of Fig. 5 could be used both in navigation and entertainment systems. For instance, one radio could have the functions of a AM/FM radio, a RDS radio, a digital audio broadcasting (DAB) radio, a mobile transceiver, a GPS receiver, and receivers for other guidance and information media. A CD-ROM player could be used to contain the navigation software and map on one CD-ROM, or to listen to the music stored on another CD-ROM. A memory card reader could be used for map storage or for other mobile office devices. A display monitor could be used to display a map or to watch TV programs. A computer or microcontroller could be shared with other electronic devices. Many input and output devices could also be shared by both navigation and entertainment systems.

Table 2. Comparison of Different Location Technologies Technology Dead reckoning (DR) Terrestrial radio GPS DR ⫹ map matching (MM) DGPS GPS ⫹ DR ⫹ MM DGPS ⫹ DR ⫹ MM

Performance Poor (poor longer term, but good short term) Fair (150–2,000 m, will improve) Fair (100–300 m with SA and without blockage) Fair (20–50 m without loss) Good (10–20 m without blockage) Better (15–50 m continuous) Best (10—15 m continuous)

Integrated antenna

Receiver

Device control

Visual display On-board computer

Auxiliary memory

Audio device

Position sensors

Safety sensors

Figure 5. Simplified block diagram of an autonomous navigation system. An autonomous system can guide the user based on an onboard map to a pre-selected destination without any intervention. The guidance instructions can be displayed as a visual route, announced as turn-by-turn directions, and so on. The on-board computer is the ‘‘brain’’ of the system. Important actions are finalized there. The visual display, audio device, and device control are a part of a human– machine interface. The auxiliary memory can be used to store digital map data, software, or as an information and entertainment medium. The receiver may be integrated to receive both GPS positioning data and entertainment radio signals. Position sensors are used to measure vehicle distance traveled and direction of travel. Safety sensors are those that provide safety features to the vehicle user, such as the air bag deployment sensor. All the components in the shaded blocks could be used in both navigation and entertainment.

Research and development are under way to further improve driving conditions. Collision avoidance can steer the vehicle out of potential collision situations or inform the driver to take proper actions. Lane holding can keep the vehicle in its current lane. Blind-spot elimination can eliminate areas that cannot be observed by the driver. Vision enhancement can improve the driver’s vision of traffic. Automated highway system (AHS) can automatically control vehicles on the road without intervention from drivers or AHS operators (23). Other existing devices, such as cruise control and air bags, will have intelligence built in to operate more adaptively to their intended environments. Centralized Navigation Systems A centralized navigation system provides guidance and information to the user with the assistance of a remote host or a centralized computing facility. It relies on the communications network to present additional services that autonomous systems cannot provide. In addition to information transmission, communications systems can also provide location capabilities. The terrestrial radio-based technique is an example. It often involves the participation of one or more communications infrastructures in the position determination (see Fig. 4). In 1996, The US Federal Communications Commission (FCC) mandated that, by the year 2001, wireless communications systems must be able to locate the caller within a radius of 125 m in 67% of

VEHICLE NAVIGATION AND INFORMATION SYSTEMS

all cases. Many telecommunications companies are actively looking for technical solutions that can satisfy this regulation. This will certainly boost the positioning performance of terrestrial radio-based technologies. For centralized systems, location and navigation intelligence can be placed either on the mobile unit or on the host. For instance, the mobile unit can play an active role in location determination with an on-board GPS receiver. It can also play a passive role by simply receiving a location determined by the remote infrastructure. Automatic vehicle location (AVL) systems have been widely used in fleet-management systems. An AVL system tracks the position of each vehicle in a fleet and reports the information to a host via a wireless communications network. It can be used for private fleets, secure-delivery fleets, emergency vehicles, and public transportation. One popular approach is to integrate a simple positioning subsystem, a human–machine interface subsystem, and a voice and data communications subsystem in the vehicle. On the host end, a map database subsystem, a human–machine interface subsystem, and a voice and data communications subsystem are integrated into a dispatch or management center. Because of the availability of the communications network, the positioning device used on board is usually a GPS receiver that periodically receives broadcast differential corrections (DGPS) from the center. The driver has a mobile display terminal, while the dispatcher in the center has a graphical display showing the vehicle location on the map. In this way, the whole fleet can be easily controlled by the dispatch center either by voice, data (such as text commands), or both. An AVL system can be as simple as a stolen-vehicle-tracking system or as complex as a dispatching system. For a public-transit system, additional functions could be added such as a route-by-route transit schedule, enroute information (on board and at the bus station), transfer management, fare-collection registration, passenger counting, vehicle diagnosis, emergency alert, paratransit management, and on-board video surveillance. For other fleets, route planning and guidance, vehicle diagnosis, and emergency alert capabilities can be provided by the center. Dynamic route guidance (navigation) is a typical centralized application. As was earlier discussed, a dynamic navigation system can guide the vehicle or its users based on realtime traffic data. When the navigation intelligence is placed on the host and infrastructure, the vehicle passively receives maneuvering instructions and guidance icons. The guidance data can be transmitted to the vehicle via various wireless media, such as short-range-beacons and mobile radios. When the navigation intelligence is placed on the vehicle or mobile device, all the navigation related activities are conducted on board. The vehicle then receives real-time traffic data to assist the on-board guidance. One example of such a system (Chicago ADVANCE system) with on-board components is shown in Fig. 6. In either case, vehicles equipped with a dynamic navigation unit can act as a traffic probe to collect realtime data for the traffic or service center. Integration of all the system components is a very significant challenge. If the project involves multiple organizations, the challenge is even bigger. Meanwhile, additional systemwide impacts must be addressed. For example, if the dynamic navigation intelligence is placed in the vehicle and all the vehicles in the area are equipped with the same such sys-

GPS receiver GPS antenna

117

Compass

RF antenna

Visual display Memory card

Gyroscope

Speaker RF modem

Cable harness Transmission pickup Navigation computer with CD-ROM drive

Figure 6. On-board components of a dynamic navigation system. This particular system communicates with a service (or traffic management) center via its RF antenna and modem. Real-time traffic information is used to provide for a safe and efficient journey. A GPS receiver, a transmission pickup, a gyroscope, and a compass are used for vehicle positioning. A speaker is used to announce maneuver instructions. Visual display and its touch-screen-based control are used to display information such as a map and to control the navigation computer and devices. A CD-ROM drive and a memory card are used to store and retrieve the digital map data and traffic data, respectively.

tem, the Braess paradox could occur. This phenomenon will cause all of these vehicles to be dispatched to a previously uncongested road, which may make the traffic situation in the area even worse. Despite the complicated integration tasks, the advantages of the dynamic navigation system thoroughly justify the pain of implementation. One recent development is the mayday system, whose onboard system integrates a GPS receiver with a cellular transceiver. The system is activated either by the user or by an emergency event to seek assistance from a service center. Therefore, a mayday system does not need to keep the communications channel open on the regular basis as most AVL systems do. Only when it is activated is the communications channel established to transfer GPS-detected location data for the verification on the center’s map. It then keeps the user in voice contact with a human operator for emergency roadside assistance and other help. Some people refer to this system as the emergency call or distress call system. The mayday system can be further improved by including other services such as remote door unlocking, theft detection, stolen-vehicle tracking, vehicle diagnosis, route guidance, and travel information. As the installed base keeps increasing and the hardware unit price keeps shrinking, even more services can be added. CONCLUSION Great mobility has long been a dream of civilization. Modern vehicle navigation and information systems not only help to achieve this dream, but also make it more safe and efficient. More and more people have embraced this new technology. Modern autonomous vehicle navigation systems began to appear in the Japanese consumer market in the late 1980s and in the European and American markets in the mid-1990s. Dy-

118

VELOCIMETERS

namic navigation systems have been marketed in Japan and mayday systems have been sold in the United States since 1996. All of these systems have shown great user acceptance. In the future, vehicles will be equipped with additional devices to enhance performance, which may include collision avoidance, lane holding, blind-spot elimination, vision enhancement, vehicle-to-vehicle communications, and so on. As the explosive advance of communication, computer, control, and information technologies continues, the day when people in the vehicle can be connected with the Internet, e-mail, fax, video conferencing, remote diagnosis, and other office and entertainment devices will not be far away. Increasingly, people will benefit from intermodal travel, mobile offices, and other conveniences that modern technologies will bring. Vehicle navigation and information systems will be in every corner of the world in the twenty-first century. BIBLIOGRAPHY 1. Y. Zhao, Vehicle Location and Navigation Systems, Norwood, MA: Artech House, 1997. 2. T. Tuo and T. Alu (eds.), Song Shi (History of The Song Dynasty, in Chinese), Yuan Dynasty, China, 1345. 3. J. Needham, Science and Civilization in China, Vol. 4, Part II, Cambridge, U.K.: Cambridge Univ. Press, 1965. 4. R. L. French, From Chinese chariots to smart cars: 2,000 years of vehicular navigation, Navigation, 42 (1): 235–257, 1995. 5. R. Jurgen (ed.), Automotive Electronics Handbook, New York: McGraw-Hill, 1995. 6. C. O. Nwagboso (ed.), Automotive Sensory Systems, London: Chapman & Hall, 1993. 7. W. B. Ribbens, Understanding Automotive Electronics, 4th ed., Carmel, IN: SAMS, 1992. 8. B. W. Parkinson et al. (eds.), Global Positioning System: Theory and Applications, Washington, DC: American Institute of Aeronautics and Astronautics, 1996. 9. E. Kaplan (ed.), Understanding GPS: Principles and Application, Norwood, MA: Artech House, 1996. 10. DoD/DoT, 1994 Federal Radionavigation Plan, Springfield, VA: National Technical Information Service, Report No. DOTVNTSC-RSPA-95-1/DOD-4650.5, May 1995. 11. R. B. Langley, GLONASS: Review and update, GPS World, 46– 51, July 1997. 12. Russian Military Space Forces, GLONASS [Online]. Available at http://www.rssi.ru/SFCSIC/glonass. html, 1998. 13. M. N. DeMers, Fundamentals of Geographic Information Systems, New York: Wiley, 1997. 14. C. Drane and C. Rizons, Positioning Systems in Intelligent Transportation Systems, Norwood, MA: Artech House, 1998. 15. R. L. French, Map Matching Origins, Approaches and Applications, Proc. 2nd Int. Symp. on Land Vehicle Navigation, 1989, pp. 91–116. 16. C. Nwagboso (ed.), Advanced Vehicle and Infrastructure Systems: Computer Applications, Control and Automation, Chichester, UK: Wiley, 1997. 17. T. H. Cormen, C. E. Leiserson, and R. L. Rivest, Introduction to Algorithms, Cambridge, MA: MIT Press, New York: McGrawHill, 1990. 18. B. Peacock and W. Karwowski, Automotive Ergonomics, London: Taylor & Francis, 1993. 19. Y. I. Noy (ed.), Ergonomics and Safety of Intelligent Driver Interfaces, Mahwah, N.J.: Lawrence Erlbaum Associates, 1997.

20. W. Barfield and T. A. Dingus (eds.), Human Factors in Intelligent Transportation Systems, Mahwah, N.J.: Lawrence Erlbaum Associates, 1998. 21. S. D. Elliott and D. J. Dailey, Wireless Communications for Intelligent Transportation Systems, Norwood, MA: Artech House, 1995. 22. J. D. Gibson (ed.), The Mobile Communications Handbook, Boca Raton, FL: CRC Press, 1996. 23. P. A. Ioannou (ed.), Automated Highway Systems, New York: Plenum Press, 1997.

YILIN ZHAO Motorola, Inc.

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELECTRICAL...NICS%20ENGINEERING/62.%20Vehicular%20Technology/W7713.htm

}{{}}

●

HOME ●

ABOUT US ●

CONTACT US ●

HELP

Home / Engineering / Electrical and Electronics Engineering

Wiley Encyclopedia of Electrical and Electronics Engineering Vehicular Electronics Standard Article Halit Eren1 and Frank Longbottom1 1Curtin University of Technology, Bentley, Western Australia,, Australia Copyright © 1999 by John Wiley & Sons, Inc. All rights reserved. DOI: 10.1002/047134608X.W7713 Article Online Posting Date: December 27, 1999 Abstract | Full Text: HTML PDF (270K)

●

●

● ●

Recommend to Your Librarian Save title to My Profile Email this page Print this page

Browse this title ●

Search this title Enter words or phrases

Abstract The sections in this article are Power and Signal Management on Vehicles Engine Management and Control Transmission Control Chassis Control Systems Body Electronics Recent Developments in Vehicle Electronics; Position and Route Determination About Wiley InterScience | About Wiley | Privacy | Terms & Conditions Copyright © 1999-2008John Wiley & Sons, Inc. All Rights Reserved.

file:///N|/000000/0WILEY%20ENCYCLOPEDIA%20OF%20ELE...ENGINEERING/62.%20Vehicular%20Technology/W7713.htm15.06.2008 12:46:41

❍

❍ ❍

Advanced Product Search Search All Content Acronym Finder

J. Webster (ed.), Wiley Encyclopedia of Electrical and Electronics Engineering c 1999 John Wiley & Sons, Inc. Copyright


VEHICULAR ELECTRONICS A vehicle can be described as a device on wheels for conveyance. Modern vehicles, used for transport of people and goods, include automobiles, trucks, buses, trains, and streetcars. As a result of recent developments in vehicular technology and improvements in electronics, modern vehicles are improving in safety, convenience, entertainment, and comfort as well as in their primary function. In 1887, the German engineer Gottlieb W. Daimler first applied the gasoline engine to road vehicles. The automobile in the United States began to be developed from 1895 onwards. Motor vehicles for freight and passenger transportation were well established before 1910 in the form of private cars, buses, and trucks. At the turn of the twentieth century, however, there were more battery-operated vehicles in use than either steam- or gasoline-driven vehicles. The period from 1890 to 1910 is generally regarded as the golden age of electric automobiles. Then, an explosive development of the gasoline engine began, and electric vehicles made no further progress on account of the severe range limitations of storage batteries. However, battery vehicles remained in use as fork lifts, for distribution of dairy products, etc. Interest in battery vehicles and alternatives to gasoline-driven engines was rekindled after the 1976 oil crises. Today, although many alternatives exist, the gasoline driven engine is still widely used, because of the cheap and convenient supply of gasoline. The modern automobile is a highly complex technical system consisting of thousands of components and employing many subsystems that perform specific functions. Passenger cars have emerged as the primary means of family transportation. It is estimated that there are over half a billion private cars in the world. Every year, many different vehicles are introduced into the marketplace. The manufacturers capitalize on their particular technological advances to capture a bigger market share. Therefore, considerable research and development efforts are concentrated on the application of new technology, which may be considered to be the key to successful competition. The manufacturers and suppliers continually improve the body, chassis, engine, drive train, control systems, occupant safety, and environmental emission of their vehicles. An important aspect of these developments is the extensive use of new electronic components. The use of electronics in modern vehicles is for the purposes of addressing environmental regulations, increasing reliability of vehicles, maximizing safety, reducing both manufacturing and running costs, and making servicing easy and effective. In parallel to this progress the customer’s expectations increase, therefore the development of vehicle electronics becomes relentless. Today, the electronics are an essential part of the vehicle, particularly in the reduction of exhaust gas emissions, improved diagnostics, maximized safety, and improved vehicle efficiency. Extensive research is taking place worldwide to determine ways of improving the applications of electronics in vehicles’ performance, safety, and comfort without adding excessive cost. Standalone electronic components such as diodes and transistors were introduced into vehicles early in the 1960s. Solid-state (transistorized) ignition systems were well in use by 1973. However, the use of electronics recorded its most impressive rise and public acceptance after the mid 1970s. The main source of the new capability was the microprocessor. The microprocessor shifted the use of electronics from standalone components to increasingly sophisticated systems that linked many components together. 1

2

VEHICULAR ELECTRONICS

Fig. 1. Electronics is an integral part of modern vehicles. Some of the most common electronic systems are engine control, power train, cruise control, antitheft devices, automatic braking systems, driver display, suspension control systems, and safety systems such as airbags. Others electronic systems, such as Global Positioning System (GPS) access and route management, are well developed but not so common, mainly due to cost.

Today’s vehicle electronics spans a wide range of hardware and software that contains many integrated circuit (IC) chips, microprocessors, and smart sensors backed up with highly dedicated programs for real-time operation. Some of the purposes of these electronic systems are engine control, driver information, entertainment, and ignition control. Vehicle electronics today include a total vehicle network that interconnects many sensors and actuators to work together in an efficient and reliable manner. Many vehicles are totally integrated with electronic systems, which are primarily aimed at optimizing their performance. The total system has flexibility and adaptability with extensive software control of multifunctional features. Many components are available in add-on format that can be customized by the buyer to suit his/her requirements and budget. Automobile manufacturers worldwide are the main agents of advancement, research, and development of automobile electronics. There are, however, many other establishments and companies assisting in the research and development of vehicle electronics, offering discrete components or complete systems to manufacturers and consumers. As an information-based system, the onboard electronics of vehicles uses extensive computing capacity: multiplexed networks and extensive memory. Some systems include speed control integrated together with engine and transmission control. Other features include torque demand power-train control, vehicle dynamics such as braking, steering, and suspension, electric power management, and displays for driver information, climate control, and entertainment functions. A block diagram of a typical vehicle control system is shown in Fig. 1. Electronics is also finding increasing applications in chassis systems, consisting of steering, brakes, and suspensions. This leads to synergic integration of mechanical systems through integrated electronic networks. The use of electronics allows adaptive control of the suspension springs, the shock absorbers, and the suspension geometry. The electronic system senses displacement and acceleration in order to control the ride height, aerodynamic angle of attack, and dynamic response of the vehicle’s body. Traction control systems also improve vehicle stability and avoid collisions. In some applications the space all around the vehicle is monitored for the presence of collision risks, using a combination of sensing technologies, such as radar, laser, visual, infrared, or ultrasonic. The output of the sensors are analyzed by artificial intelligence (AI) software that directs the controllers to reduce acceleration, apply brakes, or tighten seat belts. In modern vehicles, human factors and the design for driver information displays and controls are also highly advanced. Essential information such as vehicle speed is displayed on a reformattable multifunction

VEHICULAR ELECTRONICS

3

display panel. It uses several display technologies, such as liquid crystals, vacuum fluorescent devices, and light-emitting diodes (LEDs). The system displays performance data whenever it senses something unusual. In addition, in some vehicles, the driver is able to select particular arrays of information and the style in which it is presented. For example, the driver is able to request a complete display of all engine operating parameters, such as engine speed, oil pressure, coolant temperature, and fuel pressure. The display may also include maintenance information, such as the need for lubrication, brake checking, and so on. Notifications of emergency and alarm conditions are in the form of audible or visual signals. In some cases, voice recognition is used for functions such as entertainment system control, driver display mode selection, and telephone dialing. Some systems include detection of impairment or loss of alertness on the part of the driver. These focus on the actions that are fundamental to the driver’s safe operation of the vehicle, such as appropriate steering and braking behavior. Climate control is another advanced feature of modern vehicles that is electronically controlled and electrically powered. The air distribution is designed to permit different temperature variations in different zones of the passenger compartment. Supplemental electric heaters are capable of fast response to peak heating demands and allow for the reduction of overall heater capacity. Cellular telephones find extensive use in vehicles. Modern navigation aids also find increasing applications in passenger, commercial delivery, service, and rental vehicles. These systems operate on radio navigation systems such as loran C or the GPS. They also use map-matching techniques with digital maps that are stored on compact disks onboard the vehicle. The GPS is used to track vehicles’ geographic locations, particularly in fleet operations. Some of these systems utilize an external data link that will enhance the on-vehicle navigation system by providing the driver with current traffic information. Future development in vehicular electronics is dependent on the progress of the electronic industry, including smart sensors, smart actuators, and advances in communication technology. The use of electronics in modern vehicles can broadly be categorized as follows: • • • • • • • •

Electric power and electronic signal management, such as power management control systems Electronic engine management systems Transmission control, traction assistance Chassis control systems, antilock–antiskid brake systems, control of vehicle dynamics such as stability control and steering systems Security systems, such as antitheft devices, immobilizers, keyless entry, remote locking, and alarms Body electronics for safety, such as crash detection, airbags and intelligent seats, cruise control, human– computer interfaces, dashboard displays, and voice recognition Automatic vehicle monitoring, position and path finding, and root determination. Electric or hybrid technology, such as electric vehicles, hybrid vehicles, and fuel-cell hybrid technology.

In this article, the details of onboard vehicle electronics will be given. However, due to lack of space, this article will not consider off-board vehicle electronics and other supporting systems used in the vehicle industry, such as wheel balancers, engine-tuning units, automatic testing, and other electronic design and manufacturing aids. Vehicle electronics may vary considerably from one application to another, such as common automobiles, trucks, buses, and trains. Also, products may differ from one manufacturer to another or even from one type of vehicle to another from the same manufacturer. There is a diverse range of vehicle electronic products available in the marketplace, and new ones are introduced almost daily. Here, a conceptual approach will be adopted and general explanations will be given without distinguishing the form of application or the type of vehicle. Some examples will be introduced, and the application areas will be highlighted as the need occurs.

4

VEHICULAR ELECTRONICS

Fig. 2. Electric system in vehicles. An alternator coupled to the engine supplies power to the batteries and to the rest of the vehicle when the engine is running. The battery supplies power when the engine is off and during starting. The electric power must be available at all times for continuous load, prolonged loads, and periodic loads.

Power and Signal Management on Vehicles Historically, electric arc lamps preceded both Kettering ignition and incandescent lighting on automobiles. They appeared for the first time on electric automobiles in the 1880s. In the early stages of development of gasoline vehicles, the electric systems were limited to ignition systems and electric starters. From 1912 onwards, electric lights began to replace the kerosene and acetylene lights, and electric horns replaced the mechanical bulb horns. By 1930, the electrification of vehicles could be considered to be complete, 6 V dc was used in most vehicles, with a lead-acid battery as a reservoir to store excess output of the generator. Today, 12 V dc is standard. The electric systems in vehicles supply the power for various operations, as illustrated in Fig. 2. The system comprises a storage battery and charging system, a generator, a motor starting system, lighting, an ignition system, and body electric systems such as heaters, air conditioners, washer–wipers, power seats, windows, and mirrors. The electric power in vehicles is required to be available at all times for the starter, for the ignition and fuel injection system, for the electronic controller unit, for lighting, for safety and convenience electronics, for driver information, and so on. The electric power is supplied by 14 V (28 V for commercial vehicles) alternators. Alternators are designed to generate dc (with brushes) or ac. When three-phase ac alternators are used, the voltage and current must be rectified, as exemplified in Fig. 3. Other electronics such as transistor or integrated circuit (IC) hybrid regulators are also used in power conditioning, voltage regulation, overvoltage protection, and the like. The modern vehicle has complex wiring systems. The wiring system can be divided into two groups: (1) power distribution network, and (2) information distribution network. The power distribution network delivers the necessary power to all electric equipment. The information distribution network connects sensors,

VEHICULAR ELECTRONICS

5

Fig. 3. Rectification of the ac voltage. Many vehicles are equipped with three-phase alternators supplying ac. The ac voltage is rectified by solid-state electronics using diode bridges. Voltage regulators are used to maintain the alternator voltage constant, irrespective of the load and rotation speed of the engine.

actuators, and other electronic components. The information distribution network uses both low-speed and high-speed buses. Currently, among many alternatives, the controller area network (CAN) of Bosch is the accepted standard for information distribution in vehicles. The Controller Area Network. The conventional method of using dedicated wires to interconnecting sensors, actuators, and controllers has led to a rapid increase in the complexity of vehicle wiring harnesses due to the push by manufactures to add more functionality. Wiring harnesses have become excessively bulky. As the electronic devices shrink in physical size, due to technology advancement, the size of connectors has increased to accommodate more connection points. The CAN, shown in Fig. 4, has been developed to address these problems. A CAN is a shared serial broadcast bus currently with a maximum speed of 1 million bits per second. Messages of various lengths, zero to eight bytes, are placed on the bus. Each message includes an identifier that specifies the data within the message and also defines the priority of the message. Two versions of CAN protocols have been defined: CAN 1.0 and CAN 2.0, as indicated in Table 1. The difference is in the length of the identifier field: the standard 11 bits, and the extended 29 bits, respectively. The CAN uses the multimaster principle whereby the devices on the bus do not rely on a central master for accessing it. This approach provides protection against any local malfunctioning in the bus, enabling other components to work normally. Each message is assigned a unique identifier, which also defines its access priority to the bus. The devices are categorized to have either high priority or low priority. A high-priority device can interrupt the transmission of a low-priority device. This mechanism is needed to settle bus contention where two or more devices attempt to access the bus simultaneously. Nevertheless, all devices are required to sense the state of the bus before attempting to transmit. If interrupted, devices transmitting low-priority messages are able to notice that they have been interrupted. Then the low-priority devices discontinue transmission and switch to the receive mode. They can attempt retransmission once the bus becomes idle again. Electromagnetic Compatibility. Electromagnetic compatibility (EMC) of a subsystem means by its ability to coexist with other subsystems, neither being adversely affected by the others nor adversely affecting them by the generation of electromagnetic interference. In electric systems, electromagnetic interference can take place through the leads and connectors or through airborne radiation. A vehicle is a harsh environment where sensitive electronic systems must coexist with high-powered electrical systems in a confined space. Since the vehicle contains a single electric power supply to which all electric systems are connected, any disturbance produced by one subsystem can easily affect all other subsystems. In vehicles, the DIN 40389 (Section 1) standard has been developed to classify types of disturbances in terms of shape, duration, current, and voltage. Five types of pulses have been defined, as shown in Table 2. This standard defines four classes of increasing interference pulse amplitude of each disturbance type. This allows vehicle systems to be designed to specific DIN 40389 classes, for example, where all interference

6

VEHICULAR ELECTRONICS

Fig. 4. A controller area network (CAN) bus connects the microprocessors and microcontrollers to sensors and actuators, forming a complex network. Many devices can be connected to a CAN bus (a). Some vehicles contain more than one bus. The bus has a physical layer and a data link layer (b). Various protocols are used to enable smooth information flow.

sources must not exceed class II and all susceptible devices must be capable of withstanding class III pulses, and so on. Electromagnetic interference can propagate from the source to the receiver by a number of mechanisms: (1) directly, through electric connections between the source and the receiver, (2) indirectly through inductive or capacitive coupling, and (3) by radiation. In general, it is easiest to suppress interference at the source with electronic devices. In vehicles, this is the technique that is commonly applied, using suppression resistors, capacitors, coils, filters, shielding, and grounding:

VEHICULAR ELECTRONICS

7

Suppression Resistors. These are used in high-voltage circuits such as the ignition system. Installed as close to the source (spark plug) as possible they provide an increase in impedance that helps dampen the interference energy. Suppression Capacitors. The impedance of a capacitor decreases as the frequency increases. Therefore, suppression capacitors shunt high-frequency interference energy to ground before it enters the device being protected. Suppression Inductors (Coils). An inductor’s impedance increases as the frequency increases. Inductors are used to provide increased resistance to interference energy. Suppression Filters. A filter is a combination of capacitors and inductors. The correct combination of values and configuration provides greater attenuation of interference energy than a capacitor or an inductor alone. Shielding. Completely enclosing the interference source with a conductive shield prevents electromagnetic energy radiating into the environment. A metallic braid can cover cables, while modules may be encased in metal or metal-coated enclosures (housings). Grounding. Ground connections are extremely important. They are designed to minimize the impedance to the common ground point, and care is taken to ensure that the interference currents do not couple into other systems or cables that the ground is protecting. Improper grounding can be an important source of noise.

Engine Management and Control Today’s vehicles are designed and manufactured to have low exhaust emission. Modern electronics is used extensively to achieve this objective through microprocessor-based engine management systems. However, in order to gain a good understanding of the full functionality of electronics in engine management systems, some of the major processes in internal combustion engines must first be explained. Internal Combustion Engine. The internal combustion engine (ICE) produces power by the combustion of a mixture of hydrocarbon fuel and air. Combustion is achieved either by producing a spark in the presence of the combustible mixture (gasoline engine) or by compressing the mixture above a certain temperature to achieve ignition (diesel engine). The ratio of fuel to air, the density of the air, the temperature of the engine, and the timing of ignition, together with many other factors, affect the power and by-products (emissions) produced. The theoretical ideal air/fuel ratio is 14.7 : 1 and is termed the stoichiometric ratio. The

8

VEHICULAR ELECTRONICS

ratio lambda (λ) is a common parameter and is defined as follows:

where the induction air mass is the mass of air inducted into the combustion chamber. Maximum power is produced for a λ of 0.95 to 0.85, and minimum fuel consumption is achieved for a λ of 1.1 to 1.2. Similarly, no single λ value can achieve minimum emission values for NOx , hydrocarbons (HC), and CO simultaneously. In practice a λ between 0.9 and 1.1 is used in conjunction with a three-way catalytic converter, which reduces exhaust emission by further chemical reactions. The optimal λ value is shifted by several operating conditions, such as: • • • • • • •

Cold Start The low temperature of the combustion chamber decreases the amount of fuel in the mixture due to reduced fuel vaporization and condensation on the sides of the chamber. To compensate, more fuel must be added to achieve reliable starting. Poststart Increased fuel is required to achieve a smooth idle until the chamber reaches normal temperature. Part Throttle During this stage λ is adjusted to achieve minimum fuel consumption. Full Throttle During this stage λ is adjusted to achieve maximum torque output by enriching the mixture. Acceleration The rapid opening of the throttle results in an immediate leaning out of the air/fuel ratio. To compensate, more fuel will need to be added to provide the required power for acceleration. Overrun When the throttle is suddenly closed or during descents and braking, the fuel supply is stopped. High Altitude The air density reduces for increases in altitude. The fuel delivery system must compensate for the reduced air density by reducing fuel supply to ensure the correct λ is maintained.

Electronic Fuel Control Systems. The fuel system must supply the required quantity of fuel for the current operating conditions, depending on the load on the engine. There are two major methods of fuel delivery to the engine: carburetor and fuel injection. For many years the carburetor provided adequate performance at lower cost than fuel injection systems. It is a mechanical device based upon the Venturi effect and is still used in many vehicles. Various mechanical fuel injection systems have existed since 1898, although, it was not until 1951 that a mechanical injection system was first installed as a standard feature. The electronic fuel injection (EFI) system dates to 1957, when a limited-production Bendix EFI system was first offered on the Chrysler 300 sedan. Analog EFI first appeared on the mass market in 1967, in the Bosch D Jetronic system installed by Volkswagen on air-cooled four-cylinder engines. Bosh, Bendix and various licensees worldwide have offered EFI systems continuously since then. The year 1979 saw the introduction of the first digital engine control unit, which was a digital map control for ignition timing. Also, 1979 is a significant date because of the introduction of oxygen-sensing feedback control of air-to-fuel ratio, which enabled the use of three-way catalysts for simultaneous HC, CO, and NOx reduction. The first vehicle to use the λ oxygen-sensing system was a 1979 Volvo sedan. However, O2 feedback did not require digital feedback per se, though it had been in use since 1976, following the introduction of mass-market microprocessors. In the following year the Bosch Mono-Jetronic system appeared on the market; it provided a cost-effective single-point fuel injection system for small vehicles. Since mid-1980s fuel injection systems have become an essential automotive component. A typical electronic fuel control, as shown in Fig. 5, controls the amount of fuel injected into each cylinder, cycle by cycle, in response to information obtained from the sensors. Timed current pulses supplied to the solenoid valve injectors control the amount of fuel delivered into each cylinder. The injector actuation process is performed either by group injection or sequential injection. The engine control unit (ECU) controls the

VEHICULAR ELECTRONICS

9

Fig. 5. A typical fuel injection system. Demand for efficient and smooth running of the engine necessitates precise mixture formation for every cycle. This is achieved by dedicated microprocessor systems, such as the Motorola MC68HC11, monitoring many parameters and issuing correct signals for the controllers.

fuel injectors according to the information received from the condition monitoring sensors such as the airflow sensors. Electronic Ignition Systems. An ignition system consists of a triggering mechanism, an energy storage device, a high-voltage distribution system, and a spark gap. Although, over the years, many ignition systems have been developed, they all consist of those basic elements with slight variations in the design. One of the oldest ignition systems is induction ignition, which consists of a distributor that contains both the triggering and the high-voltage distribution mechanisms. The distributor is a mechanical device geared to the engine to provide the required synchronization. The triggering mechanism consists of a set of contact points, which are mechanically opened and closed at the correct ignition point during the engine cycle. Breaker points are fitted to the distributor and are controlled by a cam that turns in synchronization with the engine. In addition, centrifugal and vacuum advance mechanisms are required to vary the triggering to compensate for variations in engine speed and load. The coil ignition system is mechanically and electrically simple. Although it provides reliable service, it has a number of drawbacks. The first is the deterioration of the break points from both mechanical wear and pitting caused by high-voltage arcing. The second drawback is that the adjustment of ignition timing for speed and load variations can only be approximated by the centrifugal and vacuum advance mechanisms. Thirdly, the high-voltage distribution system is mechanical and prone to wear. The first step in addressing these problems has been the introduction of the transistor-assisted contact breaker system. In this system

10

VEHICULAR ELECTRONICS

Fig. 6. A Hall sensor. When a current-carrying sensor is subjected to a magnetic field, a voltage is produced in the sensor that is proportional to current and intensity of the magnetic field. This is a typical example of many sensors used in vehicle electronics. Smart and miniature sensors are also finding increasing application.

the conventional distributor assembly is retained but the contact breaker points are used to drive high-power transistors, leading to following major advantages: • •

Increasing the primary current, resulting in greater spark energy Increasing the service life of the breaker points, as they only interrupt the transistor control current

In modern vehicles, there are many different ignition systems: breakerless transistorized ignition, constant-energy electronic ignition, electronic spark advance ignition, capacitive discharge ignition, twin-spark ignition, and so on. The breakerless electronic ignition system is widely used, since it eliminates the mechanical wear of contact-type systems, which required regular adjustments to correct timing errors. To eliminate mechanical problems, a number of noncontact triggering methods have been developed, such as Hall-effect and inductive. Hall-Effect Switching. When a magnetic field is applied at right angles to the flow of a supply current, a Hall voltage is produced perpendicular to both the magnetic field and the supply current, as shown in Fig. 6. The Hall voltage is produced by the force acting on the charged carriers moving under the influence of an electric field in a perpendicular magnetic field. When the magnetic field is removed, the Hall voltage vanishes. The output from this device is used to trigger a transistor to either pass or interrupt the primary current to the energy storage device. The Hall sensors are fitted to the distributor in place of the mechanical breaker points, and a permanent magnet is placed in close proximity such that the Hall voltage can be produced. A magnetic barrier is designed to pass between the Hall sensor and the permanent magnet, interrupting the magnetic field and consequently reducing the Hall voltage. The magnetic barrier is a circular vane mounted to the distributor shaft. The number of slots in the vane equals the number of engine cylinders. Inductive Switching. The inductive sensor provides greater phase displacement between the trigger point and just-off-trigger points at high engine speeds. This characteristic improves ignition timing stability. The inductive sensor unit is constructed in two parts, a stator and a rotor. The stator consists of a permanent magnet, core, and inductive winding. The rotor and stator cores are both formed from soft magnetic steel and have a number of teeth that usually equals the number of cylinders of the engine. Both the stator and the rotor are fitted in the distributor in place of the mechanical breaker points. The stator is fixed to the housing and remains stationary, while the rotor is connected to the distributor shaft and rotates in synchronization with the engine. As the rotor turns, the gap between the stator and the rotor varies, which causes a variation in the magnetic flux. This induces a voltage in the inductive winding. A

VEHICULAR ELECTRONICS

11

Fig. 7. A typical waveform of an inductive sensor. These sensors provide good ignition timing stability. The amplitude of the waveform ranges from 0.5 V to 100 V. Alternative switching mechanisms such as Hall-effect switching are widely used in the modern vehicles.

typical voltage waveform is shown in Fig. 7. The amplitude of the waveform ranges from approximately 0.5 V at low engine speed to 100 V at high speed. The advent of the transistorized ignition system led the way to variable dwell-angle control. The dwell angle is defined as the number of degrees the breaker cam rotates from the closed position to the open position of the breaker points. It is also called the dwell period. A large dwell angle is needed if a large time interval is required for the energy storage device to reach full charge. The supply voltage and the impedance of the energy storage device determine the required dwell angle. At high engine speeds less time is available for the energy storage device to reach full charge. The spark produced from a partially charged energy storage device may not be sufficient to ignite the air–fuel mixture, and a misfire then results. At low engine speeds excessive current is dissipated in the energy storage device with no additional gain in spark energy. The inductive ignition system had to compromise on the setting of the dwell angle, as it was a function of the cam shape and fixed for all engine speeds. The transistorized ignition system allows the dwell angle to be varied electronically, thereby maintaining full charge of the energy storage device over the full engine speed range. The next major improvement to take place was the replacement of the centrifugal and vacuum advance mechanisms with a semiconductor ignition system for more accurate and finer-resolution ignition timing control. This system provided the following advantages: • • •

Ignition timing can be matched to each operating condition of the engine. Other engine parameters, such as engine temperature, may be taken into account to further improve engine performance. Ignition timing can be optimized with the incorporation of knock control.

These improvements resulted in better starting, improved idle control, and lower fuel consumption. The system uses several sensors to calculate the current operating condition of the engine. A map of optimum ignition timing is stored within the controller’s memory. Using the engine’s current speed and load, the controller selects the optimum ignition timing from the stored map. The resolution and accuracy of this map are many times greater than those of the mechanical centrifugal and vacuum mechanisms. Unlike the mechanical system, irregularities in the surface of the map can easily be accommodated. The following information can be used to calculate the current operating condition of the engine: • •

Engine speed and crankshaft position Load

12 • • •

VEHICULAR ELECTRONICS Throttle-valve position Engine temperature Battery voltage

The ignition timing is directly affected by the engine speed, which is determined with inductive position sensors together with the crankshaft position. The crankshaft position is required to synchronize the ignition timing with the engine cycle. The load is calculated from the intake-manifold pressure or air-mass-flow devices; it also directly affects the ignition timing. The throttle-valve position is determined with either a switch or a variable-resistor sensor. It is used during the idle and full load to modify the control algorithm to provide optimal performance. The engine temperature is measured with a thermistor. The engine temperature also modifies the control algorithm to ensure good cold starts and idle. The value of the battery voltage is used to calculate the correct dwell angle to ensure the energy storage device is fully charged at all engine speeds. The final improvement has been obtained by replacing the mechanical high-voltage distribution mechanism completely with an electronic system, using a separate energy storage device for each engine cylinder. In the electronic system, the disadvantage of increased weight is outweighed by the following advantages: • • •

Reduced electromagnetic radiation level, as the sparks between the distributor arm and the termination points are eliminated Elimination of moving parts and the accompanying friction, wear, and noise Reduced number of high-voltage connections

Emission Control System. By-products of the operation of the gasoline engine include carbon monoxide, oxides of nitrogen, and hydrocarbons (unburned fuel compounds), all of which are pollutants. To control air pollution, governments establish quality standards and perform inspections to ensure that they are met. Over the years, the standards have become progressively more and more stringent, and the equipment necessary to meet them has become increasingly complex. Over the years, various mechanical modifications to engines and the use of electronic devices that alter emission characteristics have been successfully introduced. These include adjustable carburetor air–fuel ratios, lowered compression ratios, retarded spark timings, reduced combustion chamber surface-to-volume ratios, and tighter production tolerances. Exhaust-gas recirculation is a technique to control oxides of nitrogen, which are formed by the chemical reaction of nitrogen and oxygen at high temperatures during combustion. Reducing the concentrations of these elements or lowering the peak cycle temperatures reduces the amount of nitrogen oxides produced. To achieve this, exhaust gas is usually piped from the exhaust manifold to the intake manifold. This dilutes the incoming fuel–air mixture and effectively lowers the combustion temperature. The amount of recirculation is a function of throttle position but averages about 2%. Fuel injection, as a replacement for carburetion, is widely employed to reduce exhaust emissions. The precise metering of fuel for each cylinder provides a means of ensuring that the chemically correct air-tofuel ratio is being injected into the engine. This eliminates cylinder-to-cylinder variations and the tendency of cylinders that are most remote from the carburetor to receive less fuel than is desired. For this purpose, a variety of metering and control systems have become commercially made available. For example, timed injection, in which a small quantity of gasoline is squirted into each cylinder or intake-valve port during the intake stroke of the piston, is employed on a number of vehicles. Another approach to pollution control is the stratified-charge engine, which is a variation on conventional cylinder combustion. Fuel is injected into a combustion-chamber pocket, and the nonhomogeneous, stratified charge is spark-ignited. Operation of the engine is realized at very lean air-to-fuel ratios, thus permitting high thermal efficiency at light engine loads. This provides good reductions in exhaust hydrocarbons, carbon

VEHICULAR ELECTRONICS

13

Fig. 8. The engine control unit (ECU). In some vehicles the ECU is the main electronic control mechanism, whereas in others it is a dedicated controller connected to the main computer. It contains microprocessors, supporting chips such as analog-to-digital and digital-to-analog converters, memory, and communication buses for networking.

monoxide, and oxides of nitrogen. The primary problem with the system is to make it function over a wide range of speeds and loads with good transient response. Renewed interest in two-stroke cycle gasoline engines led several firms in the early 1990s to develop designs related to patents of the Orbital Engine Company of Australia. Air-assisted direct injection of fuel permits very lean-burning stratified combustion. A variable exhaust port confines exhaust gas within the cylinders. Electronic controls provide for proper actuation under varying speeds and loads to produce lower emissions and higher fuel economy with improved power-to-weight ratio. The Hardware and the Software of the Engine Control Unit. The ECU is composed of a metal housing encasing a printed circuit board (PCB). The PCB holds the electronic components and provides the interconnection between components, as illustrated in Fig. 8. This is one of the most important and largest components in the vehicle electronic system. The PCB houses many electronic components and ICs of various sizes. The ECU is capable of directly driving high-power actuators and switches. The environment within vehicles may be extremely harsh with temperatures ranging between −30◦ C and ◦ +60 C. In addition; the battery voltage can also vary from 6 V during cranking to 14 V during charging. The ECU must be able to operate satisfactorily under all these conditions. Modern ECUs encompass algorithms for both ignition and fuel injection systems within a single central controller. In addition, the ECU allows new systems, such as air conditioning, to be integrated with existing systems in an efficient manner. Stored within the unit are many algorithms developed to control such functions as ignition timing, dwell period, and fuel injection. These algorithms require information regarding the current operating condition of the engine, which the ECU acquires from a multitude of sensors that provide such information as the speed of the engine, airflow into the engine, and battery voltage. The output from these algorithms controls actuators and indicators, as shown in Fig. 8. The algorithms are implemented by a microprocessor within the ECU; they are stored in read-only memory (ROM), which can only be programmed once, or in electrically erasable programmable read-only memory (EEPROM), which can be programmed many times. These are both forms of nonvolatile memory, which retain their contents even when power is removed from the ECU. Also stored in ROM or EEPROM are the performance curves and program maps required for engine control. As each engine and vehicle configuration is different, the engine manufacturers can modify the curves and maps to suit the particular requirements of an engine.

14

VEHICULAR ELECTRONICS

Random access memory (RAM), a volatile form of memory whose contents are lost when power is removed, is used to store calculated values, adaptation factors, and system errors. If the vehicle’s battery is removed, then the ECU will have to recalculate the adaptation factors. To overcome this problem, some ECUs use EEPROMs instead of RAMs. The outputs of the ECU are capable of directly driving actuators such as fuel injectors or ignition coils. Each output is protected against shorts to ground and overloads. The protection circuitry provides information to the microprocessor, enabling the fault to be logged and the defective output to be switched off. The error log of the ECU is used to locate faulty components efficiently. There are three types of sensors used in conjunction with the ECU. The first provides simple on–off indication, such as the throttle idle switch position. This type of signal can simply be processed as a digital input. The second provides either a voltage or a current signal that is proportional to the parameter it is measuring, such as airflow or temperature. This type of signal requires an analog-to-digital converter (ADC) to convert the varying signal to a digital equivalent. The third type uses voltage or current pulses with varying timing to convey information such as the engine speed. This type of signal requires counters and timers to measure the time between pulses and the number of pulses within time intervals. These sensors are placed at different positions on the engine and are connected to the ECU via the wiring harness. Although the ECU is protected from electromagnetic interference with the aid of shielding and by other means as discussed above, the connections to the wiring harness can allow harmful interference and unwanted noise that adversely affect the performance from time to time.

Transmission Control Modern vehicles have automatic, semiautomatic, or manual transmission systems. The automatic transmission provides three or four gear shift positions controlled by hydraulic pressure produced by the engine. A conventional automatic transmission system consists of a number of components such as a torque converter, a gear train, friction elements, an oil pump, a hydraulic control unit, and transmission housing. The crankshaft rotation is transferred to the automatic transmission via a torque converter. The point of gear shift is set by mechanical adjustments that, once set, cannot be changed without readjustment. Compromises used to be required to achieve adequate performance over the widest range of operating conditions. Commands issued from the ECU now control the electric, vacuum, or hydraulic actuators to engage or disengage the lockup mechanisms by directing fluid into the torque-converting chamber using solenoid valves. The use of a microcomputer-controlled system improves the automatic transmission by controlling the hydraulic system, thus offering smooth gear shifting and eliminating hunting shifts and the like. With the availability of microprocessor-based control units, greater flexibility can be built into the control of automatic and semiautomatic transmissions. An example of an electronic transmission control system is shown in Fig. 9. In addition to engine load and speed, other factors such as engine temperature and the driver’s characteristics can be taken into account in determining the optimal gear-shift scheme. A communication channel between the ECU and the transmission controller enables the engine to reduce power output during gear shift, resulting in smoother gear changes.

Chassis Control Systems Chassis systems are associated with controlling the motion of the vehicle such as acceleration, braking, turning, and vibrations. Most chassis controllers are microcomputer-based devices, ranging from 8-bit microcontrollers (e.g., Motorola MC68HC11) to 16-bit or more (e.g., Motorola MC68HC16). There are over 50 versions of the

VEHICULAR ELECTRONICS

15

Fig. 9. A typical transmission control system. Electronic transmission control systems are used mainly in automatic and semiautomatic transmissions. The controller senses all the necessary parameters to actuate controllers for appropriate mechanical variations.

Fig. 10. The adhesion of the tire to the road surface is an important factor in the implementation of an electronic braking system. The acceleration and braking forces at different road conditions vary considerably; therefore, adaptability of the braking system is absolutely necessary.

HC11 mictrocontrollers, some of which are specifically designed for automobile applications. Some important and widely used chassis control systems are discussed below. Antilock–Antiskid Brake Systems. Antilock braking systems (ABSs) improve the steerability and stability of vehicles. They also prevent the lockup of a vehicle’s wheel under braking conditions, with the use of a closed-loop feedback control system. The rotational speed of a wheel being braked is monitored, and in the event of a sharp rise in deceleration the braking effort is reduced. Once the wheel’s rotational speed increases, the braking effort is reapplied. The level of braking effort required to lock up a wheel is dependent upon the adhesion of the tire to the road. The condition of the tire affects the level of adhesion. Figure 10 shows the relationship between the coefficient of adhesion of the tire and its slippage under different road conditions. The main components of an ABS are: • • •

Solenoid valve unit Master cylinder Wheel brake cylinder

16

VEHICULAR ELECTRONICS

Fig. 11. A typical response of vehicles to brake pressure. Under braking conditions pressure activates the brake. The rotational speed of the wheel is monitored continuously, and the appropriate force is applied until the vehicle comes to a halt.

• •

Wheel speed sensor Controller

Under braking conditions, pressure is applied to the master cylinder, which applies hydraulic pressure to the wheel brake cylinder, which in turn activates the brake. The controller monitors the rotational speed of the wheel using the wheel speed sensor. If the deceleration of the wheel exceeds a preset limit, then the controller either stops or decreases the hydraulic pressure applied to the wheel cylinder by controlling the solenoid valve unit. Once the wheel is no longer on the verge of lockup, the hydraulic pressure to the wheel brake cylinders must be increased to prevent the wheel from becoming underbraked. Figure 11 illustrates ABS control. The brake is applied at time zero. Graph (a) shows the wheel speed, equal to the vehicle speed before and shortly after the brake is applied. The brake pressure [graph (c)] at the wheel brake cylinder increases linearly, and the wheel decelerates. Once the deceleration exceeds a preset limit, the brake pressure is held constant by controlling the solenoid valve. If the wheel speed continues to drop below the slip threshold, the controller reduces the brake pressure by again controlling the solenoid valve. As the wheel begins to accelerate and crosses the preset limit, the solenoid valve is switched to hold the current pressure. The wheel’s acceleration increases until it exceeds a preset limit, when the controller controls the solenoid valve to commence increasing brake pressure to the wheel brake cylinder. The wheel’s velocity increases until it crosses the slip threshold velocity, at which time the controller switches the solenoid valve to hold the brake pressure constant. The cycle then continues until the vehicle comes to a stop. An ABS is designed to be capable of adapting to changes such as: • • •

Change in the adhesion between the tire and road surface, such as driving onto gravel or ice Change in pressure from the master cylinder as a result of the driver changing the pressure on the brake pedal Periodic or erratic braking effort due to uneven brake disks/drums

Versions of ABSs have been produced that are characterized by their number of channels and sensor, each with its advantages and disadvantages. Some of the versions are: (1) Four-Channel, Four-Sensor Systems All four wheels have speed sensors and solenoid valves fitted. This version, although the most costly, provides the greatest control. When braking on split-adhesion-coefficient

VEHICULAR ELECTRONICS

17

surfaces the control system ensures that the yaw moment (torque around the vertical axis) does not increase to a point that would adversely affect the stability of the vehicle. This is achieved by applying to both rear wheels the minimum braking pressures. (2) Three-Channel, Three-Sensor Systems In this configuration the two front wheels have individual sensors and solenoid valves, while the two rear wheels share one set of sensors and solenoids. (3) Two-Channel, Three-Sensor Systems In this configuration, speed sensors are positioned on the two front wheels and a common sensor is used for the two rear wheels. One solenoid valve controls the front wheels, and second valve controls the rear wheels. This version offers reduced manufacturing cost but has some performance drawbacks. If the front wheel with the higher coefficient of adhesion is used to set the braking pressure, then the other wheel will lock, causing excessive tire wear. Alternatively, if the front wheel with the lower adhesion coefficient is used to set the braking pressure, then it is possible that the braking distance will be greater than if ABS were not employed. (4) Two-Channel, Two-Sensors Systems In this configuration only two wheels may be sensed, leading to similar limitations to those discussed for the two-channel, three-sensor configuration. When a vehicle fitted with an ABS is braked on a split-adhesion-coefficient surface (e.g., left wheels on gravel and right wheels on asphalt), the vehicle experiences a yaw moment, which causes a turning motion. On larger-wheelbase vehicles, with greater moments of inertia, the driver can compensate for this effect by turning the steering wheel accordingly. For the smaller-wheelbase, lighter vehicles fitted with a ABSs, the turning force is more sudden and can result in vehicle instability. To reduce this effect, a delay is introduced in the application of brake pressure to the front wheel with the greater coefficient of adhesion. This has the effect of giving the driver time to compensate for the turning force. As this method reduces the pressure that can be applied to the wheel with the greatest adhesion coefficient, it can result in an increase in the stopping distance if not correctly tuned to the vehicle in question. ABS principles are applied to commercial as well as passenger vehicles. The major difference is that commercial vehicles use pneumatic rather than hydraulic brake systems. An ABS, for commercial vehicles may be operated in one of three modes: • • •

Individual Control In this mode the maximum braking force is applied to each wheel. This results in the shortest braking distance on consistent road surfaces, but produces high yaw moments on split-adhesioncoefficient surfaces. Select Low Control This mode eliminates yaw moments on split-adhesion-coefficient roads, but can greatly increase the stopping distance. Individual Control Modified This mode provides a compromise between the preceding two modes. A solenoid valve is fitted to each wheel, and the braking pressure on the wheel with the higher coefficient of adhesion is reduced only so far as needed to reduce yaw moments. This results in slightly longer braking distances but with greatly improved stability and steerability.

Traction Control Systems. The aim of a traction control system (TCS) is to control the traction between the tires of the vehicle and the road. It determines the maximum torque that can be applied to the tire during standing starts and moving accelerations. As the wheel is accelerated, the resultant slip produces a corresponding increase in adhesion, thereby reducing further slip. Once a point is reached where any further increase in acceleration results in slip that reduces the adhesion, the wheel will begin to spin. A TCS controls the wheel slip to maximize the transfer of force from the tire to the road, which results in improved traction and enhanced stability. There are two methods used to control the force on the wheels. The first reduces the power from the vehicle’s engine, and the second applies a braking force to the wheel. Where all wheels are on the same

18

VEHICULAR ELECTRONICS

Fig. 12. A vehicle dynamics control system. This electronic system monitors many parameters and is particularly advantageous during acceleration of the vehicle. It operates by adjusting the power of the engine through the throttle or spark timing or fuel intake.

road surface (consistent coefficient of adhesion), wheel spin will be experienced when excessive torque is applied during acceleration. Under these circumstances all drive wheels will spin, and the TCS will respond by reducing the torque at the wheels. This is accomplished by reducing the power produced by the engine, either by throttling or by retarding the spark timing or by reducing the fuel. Vehicles fitted with ECUs provide an integrated approach where any or all of the above methods are used to reduce the power produced by the engine while maintaining other constraints such as reduced emissions. Vehicles without ECUs are limited to mechanical control systems such as closing of the throttle by a servomechanism. When accelerating on a split-adhesion-coefficient road surface (e.g., left wheels on gravel and right wheels on asphalt), the wheel with the lower coefficient of adhesion will experience wheel slip. The TCS responds by applying pressure to the brake cylinder of the slipping wheel. This results in a reduction of wheel slip and in the transfer of power to the wheel with the higher coefficient of adhesion. For heavily loaded commercial vehicles large braking forces are normally required to provide optimal control. Prolonged operation under these conditions results in thermal overload of the braking system. To prevent this effect the following modes of operation have been developed: • •

Disabling traction control for speeds above some limit (e.g., 30 km/h) Monitoring the thermal load at the brake system and disabling traction controls when it exceeds some limit

Vehicle Dynamics Control. A vehicle dynamics control (VDC) system (Fig. 12) provides a further degree of sophistication to antilock brake and traction control systems by encompassing their functionality and providing additional benefits. It is a closed-loop control system that prevents lateral instability. Whereas an ABS is employed under braking conditions and a TCS is employed during accelerations, the VDC system responds under such conditions as full braking, partial braking, coasting, accelerating, load shifting, and engine drag to provide improved lateral stability and steerability. A VDC system utilizes the wheel slip properties to implement servo control for vehicle handling. Consider a vehicle entering a right-hand turn. Part way into the maneuver the lateral force on the rear wheels exceeds the adhesion forces between the tire and the road surface. The vehicle begins to rotate about the front wheels, which have not yet slipped without VDC, the vehicle has already become unstable. For the same situation a

VEHICULAR ELECTRONICS

19

Fig. 13. A suspension control system. The system senses parameters such as acceleration and steering angle to control the air or hydraulic damping through various valves.

VDC system would control the yaw rate and slip angle to ensure that vehicle does not become unstable. The inputs to the VDC system are the wheel speed, the brake pressure, the steering-wheel position, the yaw rate, and the lateral acceleration. The system calculates the vehicle speed and coefficient of friction between the tires and the road surface. The outputs of the system are control parameters for brake pressure modulation and engine management. The VDC system uses all available information to calculate the maximum performance of the vehicle under the current road conditions and the driver’s requests (steering, braking, and acceleration). It also measures the current behavior of the vehicle. The VDC system attempts to minimize the difference between the maximum possible performance and the current performance of the vehicle by manipulating the forces at the interface between the wheels and the road surface. This is achieved by varying the torque and/or braking effort applied to each wheel. Electronic Suspension Control. The suspension system minimizes the transmission of the road surface irregularities to the vehicle body. Suspension systems consist of springs, dampers (shock absorbers), and locating arms to align components, as illustrated in Fig. 13. By the use of electronic damping control different amounts of damping can be engaged according to road surface conditions. The major components of the electronic suspension systems are (1) the sensors (acceleration, steering angle, etc.), (2) the electronic control unit, and (3) the dampers controlled by solenoid valves. Electronically controlled air suspensions are also commonly used.

Body Electronics Many vehicles are fitted with a wide variety of electronic devices within the passenger compartment to enhance comfort and safety of the occupants. Common systems are the dashboard instrumentation, central door locking, antitheft systems, cruise control, air conditioning, and air-bag systems. Also, collision avoidance and navigation aids are finding increasing application. Dashboard Instrumentation. All vehicles are fitted with mechanical or analog or digital devices indicating road speed, engine speed, fuel level, and coolant temperature, and with a series of warning systems such as those for oil pressure, battery charging, open doors, and high beams. Although analog devices are still used, since the mid-1980s the use of digital displays has been increasing considerably. Some of these digital displays include quasianalog functions such as pictorial symbols. In digital instruments, light-emitting diodes (LEDs), liquid crystal displays (LCDs), and vacuum fluorescent displays (VFDs) or cathode ray tubes (CRTs) display the information. Some vehicles are equipped with onboard computers (OBCs) that provide the user with journey information. The computer consists of a keypad

20

VEHICULAR ELECTRONICS

Fig. 14. An-air bag control system. Sensors generate signals on impact, and the air bag is activated and inflated to protect the driver and the passengers. The bag absorbs impact energy from occupants and collapses after a short time. There are also air-bag systems to protect against side crashing.

and a display. When required by the driver, the OBC displays information on time and date, average speed of the journey, fuel consumption, estimated arrival time at destination, outside and inside temperature, and so on, as well as some diagnostic information such as brake pad wear and windshield wash water level. In some cases, a drowsiness warning is given by monitoring the driving pattern of the driver. Safety and Supplementary Restraint Systems. Systems for protecting occupants in the event of an accident fall into four major classes: maintenance of passenger-compartment integrity, occupant restraints, interior-impact energy-absorber systems, and exterior-impact energy absorbers. A recent line of research has centered on passive restraints that do not require any action by the occupant. Supplementary restraint systems (SRSs) such as the air bag, as shown in Fig. 14, are a good example of this concept. The air-bag system consists of deceleration sensors (D sensors) mounted on the front of the vehicle, S sensors mounted in the passenger compartment, and an inflatable cushion that is concealed in the steering column or in areas of the car that are directly in front of passengers. The sensors provide information to the ECU. In case of a crash, the system provides passive crash protection by inflating to a position between occupants and the car structure in less than one-tenth of a second. The bag absorbs impact energy from occupants as they are thrown forward during a frontal crash. The bag collapses in approximately one second. Energy is absorbed by forcing gas out of the cushion through a series of ports or orifices in the fabric. Generally, the crash sensor sends an electric signal directly to an igniter, which triggers an explosion that generates nitrogen gas to inflate the air bag. A complementary device to enhance safety of the occupants in a vehicle is the seatbelt tightening system (STS), which uses similar technology to the air-bag system. Once the signal for activation is generated, the seatbelt is tightened by a wire that wraps round the inertial seatbelt drum. Security Systems and Alarms. Electronic vehicle security systems and alarms are well established. They can be factory-fitted by the manufacturer or can be purchased for add-on. There are two basic types: the antitheft devices (immobilizers) and alarms, which can be combined. Antitheft Devices. These are among the most widely used and essential parts of electronics systems of modern vehicles, since motor vehicle theft is an increasing problem (costing billions of dollars) to owners, insurers, and manufacturers. The number of thefts increases almost every year. However, the problem is not new. Long ago, drastic actions were taken to prevent theft; for example, the 1900 Leach automobile was manufactured with a removable steering wheel. In modern times, antitheft devices range from electronic alarms to radio beacons. An immobilizer is a device that, once activated, breaks one or more connections to several electric systems, making it difficult to start the vehicle. For example, the electric fuel pump is disconnected, a solenoid

VEHICULAR ELECTRONICS

21

Fig. 15. A cruise control system. This system maintains the vehicle at a set speed. The speed of the vehicle is sensed and the throttle position is adjusted. Intelligent cruise control systems are also available to adapt to the conditions of the traffic.

valve is used to block the fuel supply, and the ignition system is disabled. The vehicle’s engine is effectively incapacitated, preventing it from being started without deactivating the immobilizer. A small flashing light located, either on the vehicle’s dash or near the front windshield, is used to indicate that immobilizer is activated. Several methods are used to activate and deactivate the immobilizer, including a radio-frequency (RF) transmitter and receiver, inductive proximity, and electric contact. Keyless Entry. In recent years, the use of RF-based keyless entry systems for locking and unlocking vehicle doors, and for opening the trunk have become common. The systems make use of surface acoustic wave (SAW) resonators, which provide a high degree of frequency stabilization in the transmitters and receivers. The signals from the transmitter are commonly received by a loop antenna located in the car. The information received from the transmitter is processed and directed to appropriate mechanisms to lock or unlock the doors, open the trunk, etc. In many cases, an algorithm incorporated in the key produces a rolling code with which to electronically authenticate the user. Most low-power RF systems operate in the 260 MHz to 470 MHz band, where licensing is not required for transmitters producing less than 1 mW of power. Vehicle Alarm System. An alarm system uses sensors to detect unauthorized admission to the vehicle and then sound an alarm—either the vehicle’s horn or a siren. A wide variety of systems are in production, with different levels of protection, depending on the price. Vehicle security systems offer protection on a number of levels: (1) perimetric protection to monitor the position of the opening panels of the vehicle, (2) volumetric protection, which detects movements (by infrared, ultrasonic, or microwave means) in the passenger compartment of the vehicle, and (3) engine immobilization by either software or hardware. Many different sensors have been used, including door, trunk, and hood switches, ultrasonic sensors to detect air movement within the vehicle and microphones placed on door pillars to detect the opening of a door. Other security systems consist of radio and glove-compartment sensing, glass breakage sensing, tilt sensing, backward or forward rolling sensing, and so on. These systems attempt to scare away a thief by drawing attention to a vehicle being stolen, but because there are many false alarms, they have become less effective. Cruise Control System. The cruise control system, shown in Fig. 15, allows the driver to maintain a fixed speed without using the accelerator pedal. The system has three main assemblies: (1) a switch pack to set and resume the required speed, (2) a throttle actuator unit to control the throttle butterfly unit, and (3) a cruise control unit to determine the vehicle’s operating conditions. Intelligent cruise controls also exist that allow the vehicle to adjust to the traffic flow conditions by maintaining a safe distance. The system senses the range and speed of the vehicles in the vicinity by using a microwave radar system that operates on either the Doppler shift of the frequencies and/or time delays of reflected signals.

22

VEHICULAR ELECTRONICS

Recent Developments in Vehicle Electronics; Position and Route Determination Due to recent advances in automotive technology and the progress in the computer-aided design techniques, manufacturers are delivering products much more quickly than before to fulfil consumer demands. These days many manufacturers sponsor focus groups, where potential customers are asked for new ideas for the future car. Engineers and designers have been teamed up to produce vehicles that meet the expectations of at least a niche market. Some examples of the recent developments are as follows: The automatic temperature control in the Jeep Grand Cherokee uses infrared sensors to scan the surface body temperature of front seat passengers, and the on-board computer controls the mixture of hot and cold air for better comfort. Perhaps the most important example in vehicle electronics is the integration of personal computers and laptops with the onboard electronic systems. This enables the linking of mobile telephones with the GPS, wireless data modem, and voice recognition and text-to-speech software, all running through the PC. The integration of PCs is providing important safety and convenience features. More than a dozen cities in the United States already report traffic flow over the Internet. In addition, rudimentary voice-command systems are already available to adjust the climate control and audio systems in vehicles like the Jaguar S-Type and Mercedes-Benz S-Class. In the United States, General Motors offers head-up display on the Chevrolet Corvette. Speed, oil pressure, and other important information are projected directly onto the windshield. Cadillac has a night-vision system, which uses heat sensors to project onto the windshield an infrared image of objects that may lie beyond the range of headlights. Mitsubishi has demonstrated a small dashboard video camera that is trained on the drivers to keep them awake and alert on long trips. Cruise-control systems are being upgraded considerably. Mercedes-Benz has developed a system that generates radar beams to detect the vehicle in front and adjust the speed accordingly. Collision avoidance systems that sense the closing speed and apply breaks are gaining wider acceptance. In the recent years, vehicle position determination has attracted considerable attention for at least two reasons: •

•

The driver wants to know the distance and direction to the final destination. Radio and dead reckoning navigation aids can satisfy these requirements partially. However, the determination of the optimal route is based on the current traffic conditions, and the driver must receive that information from outside. In the United States and European countries cellular radio and VHF radio systems are used. Dispatch centers want to know their position and status of their fleet vehicles.

In addition to various radio navigation systems (e.g., loran-C), low-cost navigators such as odometer– magnetic-compass ones are used for the position determination of land vehicles such as automobiles, trucks, emergency vehicles, and rental cars. Nowadays, navigation aids using GPS systems are used extensively. Crosscountry trucks, for example, transmit their position to a dispatch center via Geostar or Starfix satellite systems using HF radios. Many modern vehicles contain digital road maps onboard. The distance and the direction of destinations are displayed on the screen.

BIBLIOGRAPHY 1. E. Chowanietz Automobile Electronics, Oxford: Newness, 1990. 2. U. Kiencke L. Guzzella Advances in Automotive Control, Oxford: Pergamon, 1997. 3. M. Shiozaki N. Hobo I. Akahori Development of a fully capable electronic control system for diesel engines, SAE 85-0172, 1985.

VEHICULAR ELECTRONICS 4. 5. 6. 7. 8. 9. 10. 11.

23

H. Bauer (ed.) Automotive Electrical/Electronic Systems, 2nd ed., Stuttgart: Bosch, 1995. H. Bauer (ed.) Automotive Handbook, 4th ed., Stuttgart: Bosch, 1992. Vehicle electronics meeting society’s needs: energy, environment, safety, SAE 92-80986, 1992. Vehicle dynamics and electronic controlled suspensions, SAE SP-861, 1991. Electronic controls of diesel engines, SAE SP-673, 1996. Eighth International Conference on Automotive Electronics, Conference Publication No. 346, London: IEE, 1996. Advances in multiplexing in automobiles, SAE SP-806, 1990. Sensors and actuators, SAE SP-1066, 1995.

HALIT EREN FRANK LONGBOTTOM Curtin University of Technology