Interference analysis and reduction for wireless systems

observation vectors {m i1 , m2 , . . . , m iM , i = 1, . . . , N }. For each state, a set of M prototype vectors model the centroids of the signal space associated with each state. 4. State observation vector probability model. This can be either a discrete model composed of the M prototype vectors and their associated probability mass function (pmf) P = {Pij (?), i = 1, . . . , N, j = 1, . . . , M }, or it may be a continuous (usually Gaussian) PDF model F = { f ij (?), i = 1, . . . , N, j = 1, . . . , M }. 5. Initial state probability vector p = [p1 , p2 , . . . , pN ]. The first step in training the parameters of an HMM is to collect a training database of a sufficiently large number of different examples of the random process to be modeled. The objective is to train the parameters of an HMM to model the statistics of the signals in the training data set. B.3 HMM—Kalman Filter Algorithm In this section we shall show how a recursive HMM estimator and a Kalman filter in conjunction with an EM algorithm can be used to estimate nar- rowband interference, signal, and their parameters. This algorithm is used in Chapter 7 as a narrowband interference suppressor [5, 6]. B.3.1 Problem Formulation We assume that the received spread-spectrum signal s (t ) is sampled at a rate higher than the chip rate of the PN sequence. This yields samples that are correlated in time. Hence, we assume sk is a finite-state discrete-time 382 Interference Analysis and Reduction for Wireless Systems homogeneous first order Markov chain. Consequently, the state sk at time k is one of the finite number of known states Mq = (q1 , q2 , . . . , qM ). The transition probability matrix is: A = (amn ) (B.2) where amn = P Xst +1 = qn | st = qm C; m , n ˛ {1, . . . , M }. Of course, amn ‡ 0, å N n =1 amn = 1 for each m , with p denoting the initial state probability vector: p = (pm ), pm = P (s1 = qm ). We assume that the number of states M of the Markov chain is known. Also, for convenience, we assume that pm = 1/M , for m = 1, . . . , M . B.4 Maximum A Posteriori Channel Estimates Based on HMMs ln f Xh | y (0), . . . , y (N - 1)C = - å N - 1 m =0 ln f ( y (m )) - NP ln (2p ) - 1 2 ln X |Sxx | |Shh | C (B.3) - å N - 1 m =0 1 2 H [ y (m ) - h - m x ]T S - 1xx [ y (m ) - h - m x ] + (h - m h ) T S - 1hh (h - m h )J The maximum a posteriori (MAP) channel estimate, obtained by setting the derivative of the log posterior function ln f H | y Xh | y C to zero, is hˆ MAP = (Sxx + Shh ) - 1Shh ( y - m x ) + (Sxx + Shh ) - 1Sxx m h (B.4) where 383Appendix B y = 1 N å N - 1 m =0 y (m ) (B.5) is the time-averaged estimate of the mean of observation vector. Note that for a Gaussian process, the MAP and conditional mean estimate are identical. The conditional PDF of a channel h averaged over all HMMs can be expressed as f H |Y Xh |Y C = å V i =1 å s f H |Y,S,M Xh |Y, s, Mi CPS |M Xs |Mi CPM (Mi ) (B.6) where PM (Mi ) is the prior pmf of the input words. Given a sequence of N P -dimensional observation vectors Y = [ y (0), . . . , y (N - 1)], the posterior pdf of the channel h along a state sequence s of an HMM Mi is defined as [2]. It can be shown that [1] the MAP estimate along state s, on the left- hand side of (B.3), can be obtained as hˆ MAP (Y, s, Mi ) = å N - 1 m =0 3 å N - 1 k =0 XS- 1xx , s (k ) + S- 1hh C4 - 1 S - 1xx , s (m ) [ y (m ) - m x , s (m ) ] + 3 å N - 1 k =0 XS - 1xx , s (k ) + S- 1hh C - 14 S- 1hh m h (B.7) The MAP estimate of the channel over all state sequences of all HMMs can be obtained as hˆ (Y ) = å V i =1 å S hˆ MAP (Y, s, Mi )PS |M Xs |Mi CPM (Mi ) (B.8) A MAP differs from maximum likelihood in that MAP includes the prior PDF of a channel. This pdf can be used to confine the channel estimate within a desired subspace of the parameter space. Assuming that the channel input vectors are statistically independent, the posterior PDF of the channel given the observation sequence Y = y (0) . . . y (N - 1) is 384 Interference Analysis and Reduction for Wireless Systems fH | Y (h / y (0) . . . y (N - 1) = PN - 1 m =0 1 f y ( y (m )) f x ( y (m ) - h ) fH (h ) Assuming also that the channel input x (m ) is Gaussian, f x (x (m )) = N (x , m x , Sxx ), with mean vector m x and covariance matrix Sxx , and that the channel h is also Gaussian, fN (h ) = N (h, m h , Shh ), with mean vector m h and covariance matrix Shh , the logarithm of the posterior PDF is given by (B.3). B.4.1 Notation Let D = (d1 , . . . , dp )¢ . Denote the sequence of observations ( y1 , . . . , yT ) as YT . Let Y k 2 k 1 = ( y k 1 , . . . , y k 2 )¢ . Let XT = (i1 , . . . , iT )¢ and ST = (s1 , . . . , sT )¢ . Let S k 2 k 1 = (sk 1 , . . . , sk 2 )¢ and X k 2 k 1 = (xk 1 , . . . , xk 2 )¢ . B.4.2 Estimation Objectives Let f0 = XA , q , D, s 2e , s 2n C denote the true parameter vector that charac- terizes the narrowband interference—auto regressive (AR) signal—and the spread-spectrum signal (Markov chain). Given the observations Yk = ( y1 , . . . , y k ), our aim is twofold. 1. State estimation. Compute estimates of the narrowband interference ik and the spread-spectrum signal sk . 2. Parameter estimation. Derive a recursive estimator f (k ) for f0 , where f (k ) = XA (k ), q (k ), D (k ), s(k )e , s(k )n C, for k > 1, given the observations Yk . For maximum generality, the HMM-KF algorithm we present allows for estimation of some or all of the parameters of f (k ), depending on which parameters are known a priori. In the CDMA signal models q , se , and sn are assumed known. For such models, the HMM-KF algorithm provides state estimates and parameter estimates of D . The HMM-KF algorithm cross couples two recursive EM algorithms, one algorithm for an HMM and the other for a noisy AR model [5, 6]. 1. At time k , the KF and recursive estimate maximize (EM) parameter estimator for the narrowband interference yield estimates of the 385Appendix B state of ik , process noise variance s 2 e , observation noise variance s 2n , and the AR coefficients d1 , . . . , dp , with p the order of the autoregressive process (narrowband interference). 2. The HMM filter and recursive EM parameter estimator for the spread-spectrum signal gives online estimates of the state of sk , transition probability matrix A and Markov chain level q . B.4.3 Spread-Spectrum Signal Estimator Using Recursive HMMs At time k , the predicted narrowband interference ik |k - 1 and variance pi k |k - 1 of the predicted error wk ” ik - ik |k - 1 obtained from the KF is available. Therefore, the HMM to be estimated is (HMM signal model) y k - ik |k - 1 = sk + wk + nk (B.9) It is assumed that the Kalman predicted error wk is modeled as a zero- mean white Gaussian process with variance pi k |k - 1 and is independent of the observation noise nk . The recursive HMM estimator recursively updates the state and parame- ter estimates of the HMM. The recursive HMM parameter vector estimate at k is denoted as f (k ) HMM ” XA (k ), q (k ), s2(k )n C (B.10) Given the signal model (B.3), the state and adaptive parameter estima- tion procedure for the spread-spectrum signal sk , is presented next. B.4.3.1 State Estimation Define the symbol PDF bm ( y k ) ” f X y k | Yk - 1 , Ik |k - 1 , sk = q (k - 1)m , Sk - 1 , f (k - 1)HMM C, (B.11) m ˛ {1, . . . , M } = 1 Ö 2p X pi k |k - 1 + s2 (k - 1) n C · expS - X y k - ik |k - 1 - q (k - 1)m C2 2 X pi k |k - 1 + s2 (k - 1) n C D which is obtained directly from (B.2), and the assumptions on the noises wk , nk ? q (k - 1) m , and s 2(k - 1) n are the estimates at time k - 1 of the m th Markov chain level and the observation noise variance, respectively. 386 Interference Analysis and Reduction for Wireless Systems Define the nonnormalized filtered Markov state density a k (m ), the filtered conditional mean (CM) state estimate sCMk |k , and the filtered MAP state estimate sMAPk |k , respectively, as a k (m ) ” f Xsk = q (k - 1)m , Yk | Ik |k - 1 , f (k - 1)HMM C (B.12) sCMk |k ” E Hsk | Yk , Ik |k - 1 , f (k - 1)HMM J (B.13) s MAPk |k = q (k - 1) j (B.14) Remark: sMAPk |k is discrete valued, s CM k |k is continuous. The nonnormalized filtered Markov state density a k (m ) is recursively computed as follows: a k (n ) = bn ( y k ) å M m =1 a (k - 1)mn a k - 1 (m ) (B.15) a1 (n ) = p 0 n bn ( y1 ) (B.16) The normalized filtered Markov state density g k (m ) is computed from a k (m ) and given by g k (m ) ” f Xsk = q (k - 1)m | Yk , Ik |k - 1 , f (k - 1)HMM C = am (m ) å N n =1 a k (n ) (B.17) The filtered CM state estimate sCMk |k and the associated conditional variance (CV )ps k |k ” E HXsk - sCMk |k C2 | Yk , Ik |k - 1 , f (k - 1)HMM J, which is the expected error in the estimate of sCMk |k , are given by sCMk |k = å M m =1 q (k - 1)m g k (m ) (B.18) ps k |k = å M m =1 Xq (k - 1)m C2g k (m ) - XsCMk |k C2 (B.19) 387Appendix B B.4.3.2 Parameter Estimation The received power levels are time varying, and if asynchronous transmission is used it may be necessary to estimate A and q . To estimate these parameters, the recursive EM algorithm is used. The recursive EM algorithm will be summarized next. At time k the parameter vector estimate is updated as f (k ) = f (k - 1) + (Icom (f (k - 1) ))- 1S (f (k - 1) ) (B.20) where Icom (f (k - 1) ) and S (f (k - 1) ) are the Fisher information matrix (FIM) of the complete data and the incremental score vector at time k , respectively, given by Icom (f (k - 1) ) = Icom (f (k - 2) ) + V (f (k - 1) ) (B.21) S (f (k - 1) ) ” ¶ Lk (f ) ¶ f |f = f (k - 1) (B.22) where V (f (k - 1) ) ” ¶ 2Lk (f ) ¶ f2 |f = f (k - 1) (B.23) Lk (f ) ” E Hln f XZk | Zk - 1 , f C | Zk ,obs , f (k - 1)J (B.24) where Zk ” XZk ,obs , Zk ,mis C denotes the complete data and Zk ,obs and Zk ,mis are the observed and missing data, respectively. In this case, Zk ,obs = Yk and Zk ,mis = Sk . Thus, given Ik |k - 1 , which are obtained from the recursive KF, we can determine LHMMk (f ) from (B.17) LHMMk (f ) = E Hln f X y k , sk | Yk - 1 , Sk - 1 , f C | Yk , f (k - 1)HMM J = - 1 2 ln X2p X pi k |k - 1 + s2n CC - å M m =1 g k (m ) · X y k - ik |k - 1 - qm C2 2 X pi k |k - 1 + s2n C · å M m =1 z k (m , n ) å M n =1 ln amn (B.25) 388 Interference Analysis and Reduction for Wireless Systems where z k (m , n ) ” f Xsk - 1 = q (k - 1)m , sk = q (k - 1)n | Yk , Ik |k - 1 , f (k - 1)HMM C denotes the normalized filtered joint probability that the Markov chain is in state qm at k - 1 time and in state qn at time k . It is shown in [7] that z k (m , n ) = bn ( y k )a (k - 1) mn a k - 1 (m ) å M m =1 å M n =1 bn ( y k )a (k - 1) mn a k - 1 (m ) , m , n ˛ {1, . . . , M } (B.26) Ignoring the terms ¶ 2LHMMk (f ) / ¶ s 2 n ¶ qm for all m = 1, . . . , M , the reestimation equations for f (k ) HMM are decoupled, then the evaluation of Icom Xf (k - 1)HMM C and S Xf (k - 1)HMM C in (B.13) yields Icom Xf (k - 1)HMM C = blockdiag X IA (k - 1) , Iq (k - 1) , I s 2 (k - 1) n C (B.27) S Xf (k - 1)HMM C = XS ¢A (k - 1) , S ¢q (k - 1)e , S ¢s 2(k - 1)n C (B.28) Thus, f (k ) HMM is updated as follows. B.4.4 Transition Probabilities The update equation for a (k )mn is somewhat complicated by the two constraints a (k )mn ‡ 0 and å M n =1 a (k )mn = 1. An elegant way of ensuring both constraints are met is to use the following differential geometric approach. Let a (k )mn = (g (k ) mn ) 2 Then g (k )mn has merely the equality constraint that å N m =1 (g (k )mn ) 2 = 1. Then computing IA and SA by projecting the derivatives to the tangent space yields g (k )mn = g (k - 1) mn + I - 1 g (k - 1)mn Sg (k - 1)mn (B.29) a (k )mn = Xg (k )mn C2 å M m =1 å M n =1 X g (k )mn C2 (B.30) 389Appendix B where Sg (k - 1)mn = 2Sz k (m , n ) g (k - 1)mn - g k - 1 (m )g (k - 1) mn D (B.31) Ig (k - 1)mn = rIg (k - 2)mn + 2Sz k (m , n )X g (k - 1)mn C2 + g k - 1 (m )D (B.32) B.4.5 Levels of the Markov Chain The update equation for q (k )m for m ˛ {1, . . . , M } is given by q (k )m = q (k - 1) m + I - 1 q (k - 1)m Sq (k - 1)m (B.33) where Sq (k - 1)m = X y k - ik |k - 1 - q (k - 1)m Cg k (m ) s2 (k - 1) n + pi k |k - 1 (B.34) Iq (k - 1)m = rIq (k - 2)m + g k (m ) s2 (k - 1) n + pi k |k - 1 (B.35) B.4.6 Observation Noise The update equation for s2 (k ) n is given by s2 (k ) n = s 2(k - 1) n + I - 1 s 2(k - 1)n Ss 2(k - 1)n (B.36) where Ss 2(k - 1)n = å M m =1 X y k - ik |k - 1 - q (k - 1)m C2g k (m ) 2Xs2(k - 1)n + pi k |k - 1 C2 - 1 2Xs2(k - 1)n + pi k |k - 1 C (B.37) 390 Interference Analysis and Reduction for Wireless Systems Is 2(k - 1)n = rIs 2(k - 2)n + å M m =1 X y k - ik |k - 1 - q (k - 1)m C2g k (m ) Xs2(k - 1)n + pi k |k - 1 C3 (B.38) - 1 2Xs2(k - 1)n + pi k |k - 1 C2 With no forgetting factor ( r = 1), and if we ignore the error in ik |k - 1 (i.e., pi k |k - 1 = 0 for all k ), then update equation for the observation noise is given by s2 (k ) n = s 2(k - 1) n + 1 k 1 å M m =1 X y k - ik |k - 1 - q (k - 1)m C2g k (m ) - s2(k - 1)n 2 (B.39) Conditional mean estimates of xk are given by KF [4]: x˙ k |k - 1 = F (k - 1)xk - 1 |k - 1 (B.40) Pk |k - 1 = F (k - 1)Pk - 1 |k - 1F (k - 1)¢ + Gs2 (k - 1) e G ¢ (B.41) uk |k - 1 = Hxk |k - 1 (B.42) hk = HPk |k - 1H ¢ + ps k |k + s 2(k - 1) n (B.43) xk |k = xk |k - 1 + Pk |k - 1H ¢ (hk ) - 1 (uk - uk |k - 1 ) (B.44) xk |k = xk |k - 1 + Pk |k - 1H ¢ (hk ) - 1 (uk - uk |k - 1 ) (B.45) Pk |k = Pk |k - 1 - Pk |k - 1H ¢ (hk ) - 1HPk |k - 1 (B.46) where F (k ) is the estimate of F in G = (1 01 · p )¢ , H = (1 01 · p ) at the k th time instant and xk |k - 1 = E Hxk | Yk - 1 , Sk - 1 | k - 1 , f (k - 1)KF J (B.47) TE AM FL Y Team-Fly® 391Appendix B xk - 1 |k - 1 = E Hxk - 1 | Yk - 1 , Sk - 1 | k - 1 , f (k - 2)KF J (B.48) Pk |k - 1 = E HXxk - xk |k - 1 C Xxk - xk |k - 1 CT | Yk - 1 , Sk - 1 | k - 1 , f (k - 1)KF J (B.49) xk |k = E Hxk | Y1 , Sk | k , f (k - 1)KF J (B.50) Pk |k = E HXxk - xk |k C Xxk - xk |k CT | Yk , Sk | k , f (k - 1)KF J (B.51) The estimate of the narrowband interference ik |k is given by the first element of the vector xk |k , while the error covariance pi k |k is given by the element (1,1) of Pk |k . The parameter estimation procedure given the follow- ing subsection requires the evaluation of quantities such as ik - m ik - n (k - 1) ” E Hik - m ik - n | Yk , Sk | k , f (k - 1)KF J (B.52) References [1] Vaseghi, Saeed V., Advanced Digital Signal Processing and Noise Reduction, Second Edition, New York: John Wiley, 2000. [2] Rabiner, L. R., and B. H. Juang, ‘‘An Introduction of Hidden Markov Models,’’ IEEE ASSP Magazine, 1986. [3] Young, S. J., ‘‘HTK: Hidden Markov Model Tool Kit,’’ Cambridge University Engi- neering Department, 1999. [4] Einstein, A., ‘‘Investigation on the Theory of the Brownian Motion,’’ NY: Dover, 1956. [5] Chui, C. K., and G. Chen, Kalman Filtering, Third Edition, Berlin: Springer, 1999. [6] Krishnamurthy, V., and A. Logothetis, ‘‘Adaptive Nonlinear Filters for Narrowband Interference Suppression in Spread Spectrum CDMA Systems,’’ IEEE Transactions on Comm., Vol. 47, 1999. [7] Krishnamurthy, V., and J. B. Moore, ‘‘Online Estimation of Hidden Markov Parameters Based on the Kullback-Leibler Information Measure,’’ IEEE Trans. Signal Processing, Vol. 41, Aug. 1993, pp. 2557–2573. About the Author Peter Stavroulakis received his B.S. and Ph.D. from New York University in 1969 and 1973, respectively, and his M.S. from the California Institute of Technology in 1970. He joined Bell Laboratories in 1973 and remained there until 1979, when he joined Oakland University in Rochester, Michigan, as an associate professor of engineering. He worked at Oakland University until 1981, when he joined AT&T International and, subsequently, NYNEX International. In 1990, he joined the Technical University of Crete (TUC), Greece, as a full professor of electrical engineering. His work at Bell Labs and Oakland University resulted in the publication of an IEEE (reprinted) book, Interference Analysis of Communication Systems, and the publication of a number of papers in the general area of telecom systems. His book on interference analysis is still referenced in textbooks and relevant international technical journals. He is also the author of four other books—two in distrib- uted parameter systems theory, published by Hutchinson and Ross; one in wireless local loops, published by John Wiley in 2001; and one in third generation mobile telecommunications systems, published by Springer in 2001. He has also served as a guest editor for three special journal issues— one for the Journal of Franklin Institute on Sensitivity Analysis and the other two for the International Journal of Communication Systems on Wireless Local Loops and the International Journal of Satellite Systems on Interference Suppres- sion Techniques. While at AT&T and NYNEX, Professor Stavroulakis worked as a technical director with the responsibility of leading a team that dealt with technoeconomic studies on various large national and international telephone 393 394 Interference Analysis and Reduction for Wireless Systems systems and data networks. When he joined TUC, he led the team for the development of the Technology Park of Chania, Crete, and has had various administrative duties besides his teaching and research responsibilities. Profes- sor Stavroulakis is the founder of the Telecommunication Systems Institute of Crete, a research center for the training of Ph.D. students in telecommuni- cations, associated with and in close collaboration with various research centers and universities in Europe and the United States. He now has a very large research team, the work of which is funded by various public and private sources, including the European Union. He is a member of the editorial board of the International Journal of Communication Systems and has been a reviewer for many technical international journals. He has orga- nized more than eight international conferences in the field of communication systems. His current research interests are focused on the application of various heuristic methods on telecommunications, including neural networks, fuzzy systems, and genetic algorithms and also in the development of new modulation techniques applicable to mobile and wireless systems. Professor Stavroulakis is a member of many technical societies and presently is a senior member of IEEE. Index Absolute signal phase, 58 Advanced mobile phone service, 4, 41 Advanced radio data information service,Absorption, 49 Access point, 35, 38 7, 10 ALOHA protocol, 143Access techniques, 132–33 Acquisition search rate, 136 Amplitude fading, 215 Amplitude modulation, 89–90, 95, 104,ACTS program, 41–42 Adaptive algorithm, 197, 201–02 107, 110, 223 interference, 95–96Adaptive array antenna, 61, 202, 204, 344–47 noise, 97–99 Amplitude-shift keying, 109Adaptive carrier tracking, 167–68 Adaptive equalization, 293, 297–99 Analog modulation, 88–92 Analog signal, 243–49Adaptive filter, 72, 296 Adaptive interference canceler, 294, Analog-to-digital conversion, 102, 152 Analog transmission, 86–88342–44 Adaptive interference canceling equalizer, interference, 93–97 noise, 92–93, 97–101330–31, 334 Adaptive multistage PIC, 354–58 Angle diversity, 182, 183 Angle modulation, 90–92Additive noise, 49, 156, 214 Additive white Gaussian noise, 73–74, Antenna direction, 49 Antenna diversity, 135, 143108, 125–26, 128, 130, 156, 160, 186, 187, 228, 230, 271, Antenna height reduction, 282 Antifrequency-selective fading, 142295, 307, 319, 320, 322, 330, 340, 357 Antipodal signaling, 114, 116–17 A priori estimation error, 330, 333, 335Ad hoc network, 31 Adjacent channel interference, 96, 118, Ardis, 39 Asynchronous transfer mode, 10, 35221–23, 240, 284 Adjacent channel interference cancellation, Asynchronous transfer mode wireless access communication, 41351–54 Adjacent channel protection, 351 Atmospheric effects, 49 395 396 Interference Analysis and Reduction for Wireless Systems Autocorrelation function, 244–45, 255, 271, 296, 333, 340, 344, 347, 354270–71, 338, 363 Automatic frequency controller, 167–68 Bit error rate average, 160–65 Bit rates, wireless, 26Automatic gain controller, 167–68 Automatic repeat request, 75 Bit-timing interval, 342 Blind cancellation algorithm, 335–40Autoregressive coefficient, 311–12 Autoregressive signal, 384 Blind equalization, 198, 299–301, 323 Block codes, 75–77Average signal power, 97 Averaging window, 18, 61 Blocking matrix, 343–44 Block interleaver, 72–73 Bandlimited signal, 170, 228, 231, 374 Bluetooth, 33–35 Bandpass filter, 93, 107, 118, 122, 134 BPF filter, 167, 168 Bandpass noise, 97, 99, 108 Branch, 177 Bandpass signal, 127, 156 Branch metric, 199 Bandwidth efficiency, 106–7, 132 Broadband adaptive homing ATM Bandwidth expansion, 294 architecture, 41 Base-Chaudhuri-Hocquenghem codes, 77 Broadband integrated services digital Base station, 18, 26, 51, 54, 55, 60 network, 41, 43 Base station antenna, 282 Broadband radio access network, 37–39 Base station power control, 286 Business premises network, 39 Base station spreading code, 136 Base transceiver station, 256 C-450 system, 6 Call admission control, 284Bayesian finite state process, 379 Bayes’ theorem, 306 Call blocking probability, 18 Call dropping probability, 18Beam pattern, 22–23 Beam-to-beam interference, 265 Call setup, 143 Capacity, 320–21Beamwidth, 281 Bello functions, 70–72 Carrier regeneration, 167–68 Carrier-to-cochannel interference, 259–61Bent pipe, 27, 29 Bessel-Fourier expansion, 370–71 Carrier-to-interference ratio, 14, 142, 214, 239, 242, 246–49, 285, 344,Bessel function, 66, 159, 252, 367, 370–71 345 analog signal, 246–49Binary amplitude modulation, 161 Binary frequency shift keying, 164–65, 187 digital signal, 250–51, 253–56, 258–59, 265–66Binary Hamming code, 75–76 Binary modulation, 102, 109–10 Carrier-to-noise ratio, 23, 25, 253 Cartesian form, 373Binary phase amplitude modulation, 233 Binary phase-shift keying, 113–15, 131, Cavity coupling, 22 Cell, 13163–64, 187, 190, 191, 193, 206, 295 Cell-loading factor, 277 Cell splitting, 18, 19Bit-energy-to-noise-power spectral density, 104 Cellular concept, 13–14 Cellular digital packet data, 7, 10, 41, 77Bit-energy-to-noise ratio, 128 Bit error probability, 160–66, 206 Cellular network types, 19–20 Cellular radio spread spectrumBit error rate, 47, 65, 71, 75, 87, 106, 121, 130–32, 151, 174, performance, 129 Center of gravity, 69184–85, 186, 187, 188, 190, 397Index Central limit theorem, 308 Conditional cochannel interference probability, 214–15CEPT, 30, 32 Conditional mean state estimate, 386, 390Channel access control sublayer, 31, 32 Conditional probability density function,Channel assignment, 143–44 156–60Channel coding, 48, 72, 72–82, 102, 105 Constant envelope, 118, 119, 342types, 74–82 Constant sampling rate, 201Channel equalization, 299–303 Constrained minimum mean square, 335,Channel estimator, 349–50 337, 338–40Channel state information, 184 Constraint coefficient condition, 342Chatter, 61 Constraint length, 78Chip duration, 201 Constructive addition, 55–56Clipped-soft-decision mapping, 349 Continuous-phase frequency shift keying,Cluster size, 15–17, 277, 283 121Cochannel cell, 15 Continuous-phase signal, 121Cochannel interference, 14, 20, 96, 143, Continuous-time digital communications,214–21, 259–61, 277, 278–79, 316–17281, 282–84, 292, 330, 331, Continuous-time message, 102, 134333, 334, 335, 340–42 Continuous-wave modulation, 88–89, 168Cochannel interference cancellation, Convergence, 296344–37 Convolutional code, 75, 77–82, 235Code division multiple access, 7, 60, 72, Convolutional interleaver, 72, 73133, 135–46, 152–53, 207, Convolutional noise, 303–4255–58, 317 Cooperation in the Field of Scientific andcellular system, 256–58 Technical Research group, 43Code division multiplexing pilot signal, Copolarization, 181169, 170–75 Correlated shadowing, 59–62 Code domain, 133 Correlation statistics distribution Code time division multiple access, 207–9 convolution, 201 Coding. See Channel coding Cosmic radiation, 49 Coding gain, 77, 82 Costas loop, 167 Coherence bandwidth, 62 Cross-correlation, 135, 177, 201, 275 Coherence time, 62 Cross-polarization, 181 Coherent detection, 108–9, 111–12, 113, CSMA/CA subframe, 36 118, 120, 151, 152, 160, 163, CT2/CT2+ systems, 9 165–67, 176, 188, 249–50, 352 Cumulative distribution function, 58, Combiner/combining, 177, 329–30. See 184, 188–89, 218 also Maximum ratio combining Customer premises network, 41 Comfort noise, 288 Cut-off rate in bits/user, 208 Communication channel number, 87 Cyclic codes, 76–77 Communication Research Laboratory, 41 Cyclic frequency hopping, 292 Complementary channel, 291 Complex envelope, 124–25, 127–28 Data compression, 102 Complex Fourier series, 368–69 Data-rate reduction, 72 Composite gamma/log-normal shadowing, DC block, 93, 96 67 Decision-directed equalization, 303–4 Decision feedback, 293Computer and communication research, 41 398 Interference Analysis and Reduction for Wireless Systems Decision feedback equalizer, 196–98, 200, Diversity, 71, 129, 135, 143, 175–92, 286–87, 294, 330207–9, 304, 330 Decision logic block, 115–16 Diversity combining, 182–92 DML receiver, 271–72Decision threshold, 130–31 Decision variable, 128 Domestic premises network, 38–39 Doppler shift/spread, 56–57, 70, 182Decorrelation filter, 181 Demodulation, 92–100, 108, 109–10, Dot product, 139–40 Double sideband, 98–99, 117115–16, 120, 131, 243–44, 268–69 Downlink channel, 134–35, 257–59, 285 Downlink satellite, 23, 25Destructive addition, 55–56 Differential coding, 105 Dual mode carrier recovery, 168 Dual path-loss law, 20Differential detection, 151 Differentially coherent detection, 160, 188 Duplexer, 132, 145, 146, 148 Differentially noncoherent detection, 188 Effective isotropic radiated power, 22–25 Differential modulation, 105 Eigenvalue/eigenvector, 318, 321, 323 Differential phase shift keying, 106, Electrical appliance interference, 49 115–16, 131, 132, 174 Embedded training sequence, 143 Differential pulse-code modulation, 106 Energy efficiency, 106 Digital advanced mobile phone service, Energy signal/waveform, 363, 365–66, 142–43 376 Digital cordless system 1800, 7, 8, 11 Energy spectral density function, 376 Digital distortion techniques, 152 Enhanced total communication system, 5 Digital European cordless, 7, 9, 134, 145, Ensemble-averaged inverse-matrix least 255 squares, 334 Digital mobile radio, 344–47 Envelope-and-phase equation, 93, 100 Digital modulation, 104–29 Envelope detection, 93, 100 Digital signal, 249–72 Equal gain combining, 186–88, 192 Digital-to-analog conversion, 152 Equalization, 72, 135, 195–206, 296–304, Digital transmission, 102–4 330–35, 351 Direct coding, 105 Ergodic hidden Markov model, 380 Directional antenna, 71, 182, 262–64 Ergodic signal, 364 Directional diversity, 182, 183 Erlang B formula, 283 Direct mode, managed, and unmanaged, 40 Error correction, 48, 74, 75 Direct reduction, 288–92 Error detection, 48, 74 Direct sequence code division multiple Error propagation, 196 access, 152, 170, 181, 268, 295, Estimate-maximize algorithm, 308, 310, 335 384–85, 387–88 Direct sequence frequency hopping, 293 Europe, 4, 7, 39, 39–42 Direct sequence spread spectrum, 124, European Telecommunication Standards 126–29, 133, 135 Organization, 30, 37, 38, 39 Dirichlet conditions, 374 Excess delay, 69 Discontinuous transmission, 148, 287–88 Exponential modulation interference, Discrete modulation, 105–6 96–97, 99–100 Discrete-time Markov process, 379 Extra-large zone indoor system, 53–54 Discrete-time message, 102 Distortion combat, 293–94 Fading, 18, 19, 50, 142, 151, 292. See also Fast fading; Frequency-Distortion mitigation, 292–304 399Index selective fading; Shadowing; Frequency division multiplexing, 87, 169 pilot signal, 168–69, 170, 171Time-selective fading Frequency division multiplexing/frequencyFano sequential decoding algorithm, 82 modulation, 268–71Fast fading, 18, 19, 50, 62–63, 66, Frequency domain, 13368–70, 128–29, 174, 294, 330, Frequency-domain description, 374, 375,354, 357 376Fast frequency hop, 125, 129, 290 Frequency domain model, 70–72Feedback decoding, 82 Frequency hopping, 36, 181, 289–92, 317Feedback filter, 196 Frequency hopping spread spectrum,Feedforward filter, 196, 330 124–25, 129, 133, 290, 293Finite-response filter, 289 Frequency modulation, 4, 58–59, 88,Finite-state Bayesian model, 299 90–92, 96, 99, 104, 107Finite-state shift register, 78 Frequency reuse, 13–17, 214, 330First generation system, 4, 5–6 Frequency-selective fading, 62, 129, 142,Fisher information matrix, 387 145, 149, 152, 181, 192–95,Fixed-chip duration, 201 200–1, 293, 331Fixed-network access point, 40 compensation algorithms, 207–9Fixed-service ML system, 268–71 Frequency shift keying, 4, 31, 106,Fixed-service DML system, 271–72 119–21, 131, 132, 164–65, 290Fixed-service frequency division multiplex/ Frequency-time orthogonalization, 317frequency modulation, 266, 268 Functional cell-loading factor, 277Fixed-service microwave link, 268 Future public land mobileFixed telephone network, 1 telecommunications system, 9,Flat fading compensation, 165–67 12–13Flat Rayleigh fading, 195 Forward code division multiple access Gauss-Hermite formula, 219–21 channel, 138–41 Gaussian frequency shift keying, 34 Forward error correction, 75–77, 139, 144 Gaussian interpolation technique, 170 Forward link, 136 Gaussian minimum shift keying, 31, 351 Forward-link interference, 277–80 Gaussian noise, 108, 225, 228, 234, 247, Fourier series, 367, 368–70 306, 308, 316–17, 319, 320 Fourier transform, 70, 71, 110, 150–51, Gaussian observation likelihood, 299, 301 180, 237, 271, 299, 372–74, Gaussian random process, 249 376 Generalized likelihood ratio, 181 Fourier transform pair, 374 Generalized packet radio service, 41 Fourth-generation system, 13 Generalized selection combining, 192 Fractional cell-loading factor, 277, 282–84 Generalized switched diversity combining, Frame synchronization, 142 192 Free distance, 82 General packet radio service, 7 Frequency detector, 93, 96 Generator matrix, 78 Frequency deviation, 90–91, 96, 121 Generator polynomial, 76, 77, 78–79 Frequency diversity, 153, 181, 183 Geosynchronous orbit, 27, 28 Frequency division duplex, 12, 134, 145, Global positioning system, 26 146–48 Global system for mobile Frequency division multiple access, 4, 133, communications, 4, 7, 8, 11, 142, 200134, 141–46, 351, 354 400 Interference Analysis and Reduction for Wireless Systems Golay codes, 76 Infrared data association, 35 Inner receiver, 208Group-interference cancellation unit, 347–48 In-phase channel, 115–16, 117, 118, 122, 169, 172, 341, 349Guard time, 134, 151, 152 Input delay spread function, 68 Hadamard codes, 76, 307 Instantaneous frequency/phase, 90, 184, Hadamard matrix, 76 290 Hadamard-Walsh sequences, 139–40 Instantaneous phase error, 167 Half-power beamwidth, 22–24 Integral equation, 317 Hamming codes, 75–76 Intercarrier interference, 152 Hamming distance, 76 Interference avoidance, 316–24, 340 Handover blocking probability, 18 Interference-canceling equalizer, 331–35 Handover/handoff, 14, 17–18, 27, 61, 136 Interference cancellation, 276 Handover probability, 18 Interference estimation/elimination, 305–8 Handover rate, 18 Interference projection, 255 Hard decision coding, 81–82 Interference suppression, 276, 288–89 Hard limiter, 167 Interim standard 54/136, 7, 8 Hata’s equation, 51–52 Interim standard 95, 7, 8 Hermitian operation, 296 Interleaving, 72–73 Hidden Markov model, 297, 299–303, Intermediate frequency, 101 379–91 Intermediate frequency filtering, 351 Hidden Markov model Kalman filter, Intermodulation interference, 223–28 308–16, 381–82, 384–85 Intermodulation product, 223 Hidden-terminal problem, 30 International mobile telecommunications High bit rate, 31, 32, 33, 35 2000, 9 High pass filter, 342, 355 International Standards Organization, 31 HIPERACCESS, 38–39 International Telecommunications Union, 9 HIPERLAN, 30–39, 37 Internet protocol, 28 type 1, 30–33 Internet service provider, 36 type 2, 35, 37, 39 Intersatellite link, 28–29 type 3, 38 Intersymbol interference, 69, 142–43, type 4, 38 195, 196, 228–39, 296–97, HIPERLINK, 38–39 330, 333, 351 Home radio frequency, 35–37 Intersymbol interference cancellation, Hopping. See Frequency hopping 344–47 Hybrid diversity, 184, 191–92 Inverse discrete Fourier transform, 150–51 Hybrid interference cancellation, 347–50 Inverse fast Fourier transform, 151 Inverse filter, 299, 304IEEE 802.11 standard, 30, 34, 35, 36, 37 IEEE 802.15 standard, 35 Inverse Fourier transform, 150–51, 373 Iridium system, 27, 28Implicit diversity, 182 Inband interference, 221 ISM 2.4 band, 34 Incremental metric, 199 Iterative reduction, 322–324 Indirect cochannel interference Japan, 4, 7, 41, 43cancellation, 340–42 Japan total access communications system, 6Indirect reduction, 277–88 Kalman filter, 308–16, 331, 381–82,Indoor communication system, 43, 52–55 Infrared, 43 384–85, 390 TE AM FL Y Team-Fly® 401Index Kalman gain, 197, 334 Matched filtering, 108, 111, 155, 164, 174, 230, 294, 296, 319, 321,Kronecker delta function, 366 335–40, 342–43, 352 Lagrange multiplier, 238, 321, 338 Maximum a posteriori channel estimate, 302, 382–91Laguerre functions, 371–72 Maximum likelihood decision rule, 155,Laguerre series expansion, 371–72 181Land-mobile radio, 39, 181 Maximum likelihood estimation, 198,Large-zone indoor system, 54 301, 302, 331, 354LBR data application, 66 Maximum likelihood sequence estimation,Least mean square-based carrier 143, 195, 198–200, 207–9,regeneration, 168 330, 331–33, 344–46, 347, 352Least mean squares algorithm, 296, 304, Maximum ratio combining, 158, 185–86,344, 354, 355–58 192, 200, 205, 330Least mean squares blind equalization, 198 pilot-aided, 186Legendre polynomial, 367 Mean channel power, 177–78Limiter-discriminator detection, 119 Mean delay, 69Linear equalization, 143, 303 Mean square error, 233, 235Linear feedback shift register, 77, 138–39 Medium access control, 31–22Linear filter, 319 Medium Earth orbit, 27–29Linear interference cancellation, 335–40 Message bandwidth, 86Linear modulation, 96, 97–99, 107–19 Message modulation, 100Linear receiver filter, 232–33 Message polynomial, 76–77Linear reduction, 294–96 Metric combining, 330 Line of sight, 47, 53, 65–66, 215 Metricom system, 39 Line spectrum, 370–72 Microcellular radio network, 19–20, 21, Loading factor, 282–84 215, 258–59 Local area network, 1, 37, 41 Microscopic diversity, 183 Local area network access point, 35 Microstrip antenna, 181 Local loop, 7 Microzone indoor system, 55 Local mean power, 215, 216, 219 Middle-zone indoor system, 54–55 Logic table, 79 Millimeter wave, 43–44 Lognormal shadowing, 19, 56, 67 Minimum mean square error, 294–96, 306 Low bit rate, 31, 32 Minimum mean square estimation, Low Earth orbit, 27–29 197–98, 319, 321, 322–24, Lower sideband, 98 330, 346 Low-noise receiver, 241 Minimum shift keying, 121–23 Lowpass filter, 86, 93, 174 Mobile broadband system, 13, 43–44 Mobile communications system, 214, 239 Macrocell environment diversity, 178–80 Mobile network access point, 40 Macrocellular radio network, 19, 21, Mobile satellite system, 26–29, 44, 258–59 265–66 Magnitude-phase form, 373 Mobile station, 18, 51, 256 MAP state estimate, 386 Mobile station power control, 286 Markovian state prior, 299 Mobile switching center, 14, 284–86 M-ary frequency shift keying, 124–25, 164 Mobile-terminating request, 143 Mobile terminating unit, 40M-ary phase shift keying, 106, 166, 167 402 Interference Analysis and Reduction for Wireless Systems Mobile-unique code, 141 Nonadaptive interference reduction, 294 Noncoherent detection, 109–10, 115–16,Mobitex, 7, 10, 39 Modulation index, 89, 92, 121–22 119–21, 125, 132, 159, 164, 188Multicarrier code division multiple access, 153 Nonfrequency-selective fading, 330 Nonlinear decision feedback, 143Multicarrier direct sequence code division multiple access, 153 Nonlinear equalizer, 195 Nonlinear estimator, 303Multicarrier system, 148–49, 153, 181 Multicell environment, 294–95 Nonlinear modulation, 119–23, 223 Nonlinear reduction, 304–16Multihop call, 28 Multimedia application, 41, 66 Non-line of sight, 63, 65, 215 Nonpilot signal-aided techniques, 167–68Multimedia mobile access point, 35 Multinomial theorem, 270 Nonreturn-to-zero, 117 Nonselective frequency fading, 62–65Multipath diversity, 129 Multipath fading. See Fast fading Nonzero frequency shift, 65 Nordic mobile telephone, 4, 5Multipath propagation, 55–73 Multiple access, 132–33 Nordic mobile telephone 450, 4, 5 Nordic mobile telephone 900, 4, 5Multiple accessing scheme, 207–9 Multiple access interference, 294–95, 304, Normalized reuse distance, 15–16 Normal probability distribution, 130289, 349, 354–55 Multiple amplitude modulation, 160–61 NTACS, 6 Nyquist interpolation technique, 170Multiple amplitude shift keying, 110–12 Multiple symbol differentially coherent Nyquist rate, 102, 375 detection, 160 Object protocol, 35Multiple user interference avoidance, 320 Observation noise, 313, 389–91Multiplicative noise, 49–50, 214 Offset quadrature phase shift keying,Multipoint communication network, 198 118–19Multistage detection, 350 Okumura curve, 51Multistage PIC, 354–58 Omnidirectional antenna, 261–62, 278Multitone approximation, 65 One-dimensional microcell, 20Multitone code division multiple access, One-path model, 331153 One-step interference cancellation, 307Multiuser detection, 294, 319, 329–30, 349 On-off keying, 109–10, 130–31Multiuser interference, 181, 275–76 Operation and management, 17 Operation and management handover, 17Nakagami fading, 66, 67, 162, 189, 190, 191, 192, 215, 217 Optimum combining, 201–6 Orthogonal coding, 139–40, 170, 200–1Narrowband channel simulations, 64–65 Narrowband fast fading, 62–65 Orthogonal cover code, 139–40 Orthogonal decomposition, 296Narrowband filtering, 351 Narrow-beam adaptive antenna, 277 Orthogonal frequency division multiplexing, 72, 148–53,Narrow-beam antenna, 277–281 Near-far interference, 144, 239–41, 284 255–56, 293 Orthogonal function, 366Nippon Electric Company, 41 Nippon Telephone and Telegraph, 4, 6 Orthogonalizing matched filter, 335–40, 342–43Noise power ratio, 227 Noise types, wireless communication, 49 Orthogonal series representation, 366–72 403Index Orthogonal signaling, 74, 131–32, 133, Pilot symbol–aided techniques, 169–70, 172, 295144, 164, 284–85 Orthogonal spreading codes, 200–1 Pilot tone–aided techniques, 168–69, 170, 171Outdoor large-zone system, 51–52 Out-of-band interference, 117–18, 119, Ping-pong effect, 347 Plain old telephone service, 7134, 221 Output correlation component, 70 Point-to-point connection, 35 Polar coordinate system, 373 Packet data network, 40–41 Polarization diversity, 180–81, 183 Packet-switched network, 1 Power control, 60, 144, 277, 284–86 Parallel detection, 276, 350 Power delay profile, 68–70 Parallel interference cancellation, 294, Power efficiency, 106 295, 347, 354–58 Power signal/waveform, 363, 365–66, Parallel-to-serial conversion, 102 376–77 Parameter estimation, 310–11, 384, Power spectral density, 104, 110, 114, 387–88, 391 118, 121, 122–23, 128, 187, Parity bit, 75, 77 243–45, 270–271, 376–77 Parseval’s theorem, 375–76 Power waveform Path diversity, 182, 183 Predetection filter, 97 Path loss, 47, 50, 51–55 Predetection noise spectrum, 97–98 Peak power, 144 Private branch exchange, 54 Peak-to-average power, 152 Private mobile radio, 39 Peak-to-mean power ratio, 153 Probability density function, 57–58, 67, Personal access communication service, 7, 9 180, 184, 189, 192, 215–18, Personal communication system, 1, 132 222, 226, 251–52, 299–301, Personal digital cellular, 7, 8, 145 306, 383–84, 385 Personal handy phone system, 7, 9, 145 conditional, 156–60 Phase amplitude modulation, 233 Processing gain, 123, 127, 129, 137–38, Phase detector, 93, 96, 243, 247 290, 295 Phase deviation, 90, 91, 96 Process noise, 312–13 Phase-encoding scheme, 168 Pseudonoise sequence, 124–26, 135, 137, Phase lock loop, 167–68 172, 174, 265, 268, 354, 355, Phase modulation, 90–91, 95, 96, 99, 356, 381 104, 107, 223 Pseudorandom hopping, 292, 293 Phase shift keying, 106, 112–19, 149, Public access mobile radio, 39 249, 251–53 Public Safety Radio Communication Phase-sweeping method, 288 Project, 40 Phasor construction, 94–95, 96, 100 Public switched telephone network, 14 Physically realizable waveform, 361–66 Pulse code modulation, 88, 89, 102, 105 Physical sublayer, 31–32, 34 Pure-combining diversity, 183–92 Picocellular radio network, 20, 21, 215 Piconet, 34 Quadrature amplitude modulation, 106, 111–12, 149, 150, 151, 166,Pilot-aided maximum-ratio combining, 186 169–70 Quadrature-carrier equation, 93Pilot code–aided techniques, 169, 170–75 Pilot signal–aided techniques, 168–75 Quadrature (Cartesian) form, 373 404 Interference Analysis and Reduction for Wireless Systems Quadrature channel, 116, 117, 118, 122, Reverse code division multiple access, 141 Reverse-link interference, 280–81169, 172, 341, 349 Quadrature Fourier series, 369–70 Rice distribution, 20, 64, 65–66, 69, 178, 215Quadrature modulation, 127 Quadrature phase shift keying, 116–19, Root mean square delay, 19, 69–70 Rural path-loss model, 52167, 169, 349 Quasi-synchronous operation, 61 Sampling, 102, 109–10Quenching, 109–10 Sampling theorem, 374–75 Satellite personal communication system,Radiocomm-2000, 6 Radio frequency, 101, 106 27, 29 Satellite system, 20, 22–29, 214, 249,RAKE receiver, 126, 129, 158, 182, 201, 205, 336, 349, 355, 358 265–66 Satellite television industry, 26RAM mobile data, 7, 10, 39 Random-access channel, 143 Scanning receiver, 286 Scattering, 66, 70Random data modulation, 121 Random signal, 58, 363–65 Schur concave/Schur convex, 321 Seamless wireless network, 41Rayleigh density function, 58 Rayleigh fading, 19, 63, 65, 66, 67, 69, Second generation system, 4, 7–9, 72, 142 Selection method, 286160, 162, 166, 177, 178, 181, 182, 184–85, 187, 189, 190, Selective combining, 184–85, 191, 200 Self-recovering equalization, 198192, 195, 215, 222, 295, 331, 340–41, 350 Serial detection, 350 Serial processing, 276Rayleigh’s energy theorem, 375 Received average signal power, 104 Serial-receiver correlation, 59–60, 61 Serial-to-parallel converter, 117Received bit energy, 104 Receive filter coefficient, 236 Seven-cell cluster, 261–64 Shadowing (slow fading), 50, 55–62, 67,Receiver complexity versus performance, 208–9 183, 294 correlated, 59–62Receiver filter, 134 Recursive algorithm, 196–97, 304 Shannon’s theory, 48 Shared wireless access protocol, 36–37Recursive estimate-maximize algorithm, 384–85, 387–88 Shift register, 77, 78, 138 Shot noise, 49Recursive hidden Markov model, 385–88 Recursive least squares, 197 Sidelobe level, 281 Sidelobe regeneration, 118–19Recursive least-squares maximum likelihood sequence estimation, Signal, 361, 3563 Signal envelope, 57330–34 Recursive narrowband interference Signal phase, 57–58 Signal processing, analog, 87estimation, 308–16 Redundancy coding, 291 Signal projection, 255 Signal-to-interference optimization,Redundant bit, 48, 74, 75 Reed-Solomon codes, 77 316–20 Signal-to-interference plus noise ratio,Reflection, 49 Relative signal phase, 58 205, 255, 296, 339 Signal-to-interference ratio, 60, 214,Repeater satellite, 29 Research and development, 10 319–20 405Index Signal-to-noise ratio, 65, 67, 71, 74, Spread spectrum system, 123–29, 135–41, 169, 290, 329–3086–87, 88, 97–99, 100, 101, Spurious signal, 223104, 130, 131, 137–38, Square-law detector, 125160–65, 174, 184, 185–86, Stack sequential decoding algorithm, 82187, 188, 192, 193, 205, 222, Standardization, 12–13242, 247, 248, 280, 289, 306 State changes equation, 105digital signal, 250–51, 256–57 State diagram, 79, 80Signal-to-noise ratio combat loss, 294 State estimation, 310, 384, 385–86Signal-to-variation power, 255 State space model, 309–11Simulcast operation, 61 State transition, 79–81, 379–80, 381Single-channel per carrier, 87 Station-to-station link, 88Single-receiver correlation, 59–60 Stochastic-gradient blind equalization, 198Single sideband, 99, 117 Stochastic signal, 49, 57, 160, 316Site diversity, 61 Subband diversity, 200–1Site-to-site correlation, 60 Subspace-based estimation, 255Six-sector model, 263–64 Subtractive demodulation, 351–54Slot synchronization, 142 Subtractive interference cancellation, 295Slow fading. See Shadowing Suburban path-loss model, 52Slow frequency hop, 125, 143, 290 Successive interference cancellation, 294,Small-angle approximation, 100 347, 350, 351–53Small cell, 71 Sum capacity, 320–21Small-zone indoor system, 55 Superframe synchronization, 142Smart antenna, 277 Super-high-frequency band, 41, 43Smoothing filter, 350 Switch and stay combining, 188–91Soft decision coding, 81–82 Switch and stay diversity, 188–91, 192Softer handover, 144 Switching (scanning) receiver, 286 Soft handover, 61, 136, 144 Symbol error probability, 160–65 Source coding, 102 Symbol generator, 105 Space diversity, 178–80, 183, 286 Synchronous connection-oriented link, 35 Space division multiple access, 202–4 Synchronous detection, 93, 95–96 Space-time orthogonalization, 317 System for advanced mobile broadband Spatial domain, 133 applications, 41 Spatial filtering of interference reduction, 202 Tap coefficient, 343–44, 358 Specialized mobile radio, 39–40 Tap gain process, 66, 68–70 Spectral density equation, 104 Tap transversal filter, 333 Spectral efficiency, 208, 209 Tap weight, 139, 296 Spectral expansion, 363–65 Terrestrial mobile cellular Spectrum analyzer, 373 communications, 253–54 Spreading chips, 139 Thermal noise, 49, 249, 266, 308 Spreading codes, 135–36, 200–1, 295, Third generation system, 9–13, 330 307 Three-sector model, 262–63 Walsh, 170–75 Threshold detector, 113, 122 Spreading gain, 201 Time delay spread, 70 Spread spectrum diversity, 129 Time-discrete process, 197 Time diversity, 182, 183Spread spectrum signal estimator, 385–88 406 Interference Analysis and Reduction for Wireless Systems Time division code division multiple Unipolar-to-bipolar converter, 117 Universal mobile telecommunicationsaccess, 12 Time division duplex, 12, 35, 36, 134, system, 9–13, 37 Universal pilot code, 140145–48 Time division multiple access, 7, 34–35, Universal wireless personal communications, 936, 133, 134–35, 141–45, 146, 207, 253–55, 351 Unnecessary handover probability, 18 Uplink antenna pattern, 23Time division multiple access/frequency division multiple access, 143–44 Uplink channel, 134–35, 143 Uplink satellite power budget, 23–25Time division multiplexing, 169 Time division multiplexing pilot signal, Upper sideband, 98 Urban path-loss model, 51–52169–70, 172 Time domain, 133 User capacity, 321 User separation algorithm, 208Time-domain description, 374, 375 Time-domain orthogonality, 201 Variable transmission rate control, 142 Time sampling, 102 Viterbi algorithm, 81–82, 198–200, Time-selective fading, 152 207–9, 330, 333, 334 Time-variant impulse response, 68 Viterbi equalization, 293, 345–47 Time-variant transfer function, 70 Vocoder, 139 Timing synchronization, 148 Voice application, 66 Total access communication system, 4, 5 Voltage control oscillator, 167 Total excess delay, 69 Total square correlation, 321, 322–24 Walsh codes, 136, 139–40, 170–75, 307 Walsh spreading codes, 170–75Tracking mode, 333, 334 Traffic channel, 143 Wavelet-packet orthogonal code, 201 Welch bound equality, 323Training mode, 333, 334, 346 Training sequence, 198 White Gaussian noise generator, 64–65 Whitening filter, 319Trans-European trunked radio, 39–40 Transmission channel, 290 White noise, 353 Wide area wireless packet data system, 7, 10Transmitter interference, 49 Transparent tone in band, 169 Wideband code division multiple access, 12 Wideband fast fading, 66, 68–70Transversal combining, 330, 331 Transversal filter, 195, 330, 331, 333 Wideband system fading, 62 Wide-sense stationary scattering, 70Transversal filter equalizer, 298–99 Traveling wave tube, 20, 22 Wide-sense stationary signal, 363 Wiener filter, 304Traveling wave tube amplifier, 22 Tree diagram, 79 Wiener-Khintechine theorem, 270–71 Wiener solution, 203Trellis coded modulation, 82 Trellis diagram, 79, 81 Wireless access communications system, 7–9Two-dimensional microcell, 20 Wireless asynchronous transfer mode, 35,Two-path model, 331, 336–37 41Two-ray Rayleigh model, 195 Wireless broadband mobileTwo-sided spectrum, 370, 373 communication system, 41–43 Ultrahigh frequency, 43 Wireless broadband multimedia communication system, 13Unbalanced branches, 191 407Index Wireless communication channel, 48–50, Wireless local loop, 7, 29–30, 41, 44, 214, 266–72132 Wireless customer premises network, 41 X.25 protocol, 40 Wireless data network, 39–41, 44 Wireless evolution, 2–4 Zero-delay channel estimation, 347–50 Wireless local area network, 1, 7, 26, 30, Zero mean Gaussian noise, 308, 309 Zero variance envelope, 34236, 44