Preface ...................................................... xiii
Acknowledgments ................................................ xv
Abbreviations ................................................ xvii
Nota Mon .................................................... xxiii
1 Introduction ................................................. 1
1.1 Multiantenna wireless channels .......................... 2
1.2 MIMO system model ....................................... 4
1.3 MIMO communication with CSIR-only ....................... 5
1.3.1 Slow fading channels ............................. 5
1.3.2 Fast fading channels ............................. 6
1.4 MIMO communication with CSIT and CSIR ................... 7
1.5 Increasing spectral efficiency: quadrature amplitude
modulation (QAM) vs MIMO ................................ 9
1.6 Multiuser MIMO communication ........................... 11
1.7 Organization of the book ............................... 12
References .................................................. 14
2 Large MIMO systems .......................................... 16
2.1 Opportunities in large MIMO systems .................... 16
2.2 Channel hardening in large dimensions .................. 17
2.3 Technological challenges and solution approaches ....... 19
2.3.1 Availability of independent spatial dimensions .. 20
2.3.2 Placement of a large number of antennas and RF
chains .......................................... 20
2.3.3 Low complexity large MIMO signal processing ..... 21
2.3.4 Multicell operation ............................. 23
References .................................................. 24
3 MIMO encoding ............................................... 25
3.1 Spatial multiplexing ................................... 25
3.2 Space-time coding ...................................... 27
3.2.1 Space-time block codes .......................... 28
3.2.2 High-rate NO-STBCs .............................. 29
3.2.3 NO-STBCs from CDAs .............................. 30
3.3 Spatial modulation (SM) ................................ 31
3.3.1 SM .............................................. 31
3.3.2 SSK ............................................. 32
3.3.3 GSM ............................................. 33
References .................................................. 38
4 MIMO detection .............................................. 40
4.1 System model ........................................... 43
4.2 Optimum detection ...................................... 44
4.3 Linear detection ....................................... 45
4.4 Interference cancelation ............................... 47
4.5 LR-aided linear detection .............................. 48
4.5.1 LR-aided detection .............................. 49
4.5.2 SA .............................................. 51
4.6 Sphere decoding ........................................ 54
References .................................................. 59
5 Detection based on local search ............................. 62
5.1 LAS .................................................... 65
5.1.1 System model .................................... 65
5.1.2 Multistage LAS algorithm ........................ 66
5.1.3 Complexity ...................................... 71
5.1.4 Generation of soft outputs ...................... 71
5.1.5 Near-optimal performance in large dimensions .... 73
5.1.6 Decoding of large NO-STBCs using LAS ............ 76
5.2 Randomized search (RS) ................................. 81
5.2.1 RS algorithm .................................... 81
5.2.2 Performance and complexity ...................... 83
5.3 Reactive tabu search (RTS) ............................. 85
5.3.1 RTS algorithm ................................... 87
5.3.2 RTS algorithm versus LAS algorithm .............. 91
5.3.3 Performance and complexity of RTS ............... 92
5.3.4 LTS ............................................. 96
5.3.5 R3TS ........................................... 100
5.3.6 Lower bounds on ML performance using RTS ....... 103
References ................................................. 107
6 Detection based on probabilistic data association (PDA) .... 110
6.1 PDA in communication problems ......................... 111
6.2 PDA based MIMO detection .............................. 112
6.2.1 Real-valued bit-wise system model .............. 112
6.2.2 Iterative procedure ............................ 113
6.2.3 Complexity reduction ........................... 115
6.3 Performance results ................................... 116
6.3.1 Performance in large V-BLAST MIMO .............. 117
6.3.2 PDA versus LAS performance in NO-STBC MIMO ..... 118
References ................................................. 120
7 Detection/decoding based on message passing on graphical
models ..................................................... 123
7.1 Graphical models ...................................... 123
7.1.1 Bayesian behef networks ........................ 123
7.1.2 Markov random fields ........................... 124
7.1.3 Factor graphs .................................. 125
7.2 BP .................................................... 127
7.2.1 BP in communication problems ................... 128
7.2.2 BP algorithm on factor graphs .................. 129
7.2.3 BP algorithm on pair-wise MRFs ................. 129
7.2.4 Loopy BP ....................................... 130
7.2.5 Damped BP ...................................... 130
7.3 Apphcation of BP in MIMO - an example ................. 131
7.3.1 MIMO-ISI system model .......................... 131
7.3.2 Detection using BP ............................. 131
7.3.3 Performance and complexity ..................... 135
7.4 Large MIMO detection using MRF ........................ 138
7.4.1 MRF BP based detection algorithm ............... 138
7.4.2 MRF potentials ................................. 139
7.4.3 Message passing ................................ 140
7.4.4 Performance .................................... 141
7.4.5 Complexity ..................................... 143
7.5 Large MIMO detection using a factor graph ............. 143
7.5.1 Computation complexity ......................... 146
7.5.2 Performance .................................... 146
7.5.3 Vector GA (VGA) in PDA versus SGA in FG BP ..... 146
7.6 BP with the Gaussian tree approximation (GTA) ......... 148
7.7 BP based joint detection and LDPC decoding ............ 151
7.7.1 System model ................................... 152
7.7.2 Individual detection and decoding .............. 152
7.7.3 Joint detection and decoding ................... 153
7.7.4 Performance and complexity ..................... 155
7.8 Irregular LDPC codes design for large MIMO ............ 156
7.8.1 EXIT chart analysis ............................ 157
7.8.2 LDPC code design ............................... 160
7.8.3 Coded BER performance .......................... 163
References ................................................. 165
8 Detection based on MCMC techniques ......................... 169
8.1 Monte Carlo integration ............................... 169
8.2 Markov chains ......................................... 171
8.3 MCMC techniques ....................................... 173
8.3.1 Metropolis-Hastings algorithm .................. 173
8.3.2 Sinnilated annealing ........................... 175
8.3.3 Gibbs sampling ................................. 176
8.4 MCMC based large MIMO detection ....................... 177
8.4.1 System model ................................... 178
8.4.2 Conventional Gibbs sampling for detection ...... 179
8.4.3 Motivation for mixed-Gibbs sampling (MGS) ...... 180
8.4.4 MGS ............................................ 182
8.4.5 Effect of mixing ratio q ....................... 183
8.4.6 Stopping criterion ............................. 184
8.4.7 Performance and complexity of the MGS
algorithm ...................................... 186
8.4.8 Multirestart MGS algorithm for higher-order
QAM ............................................ 188
8.4.9 Effect of multiple restarts .................... 188
8.4.10 MGS with multiple restarts ..................... 190
8.4.11 Restart criterion .............................. 191
8.4.12 Performance and complexity of the MGS-MR
algorithm ...................................... 191
8.4.13 Performance of the MGS-MR as a function of
loading factor ................................. 193
References ................................................. 195
9 Channel estimation In large MIMO systems ................... 197
9.1 MIMO capacity with imperfect CSI ...................... 197
9.2 How much training is required? ........................ 198
9.2.1 Point-to-point MIMO training ................... 199
9.2.2 Multiuser MIMO training ........................ 201
9.3 Large multiuser MIMO systems .......................... 202
9.3.1 System model ................................... 202
9.3.2 Iterative channel estimation/detection in
frequency-flat fading .......................... 202
9.3.3 Iterative channel estimation/equalization in
ISI channels ................................... 208
9.3.4 Equalization using initial channel estimates ... 213
9.3.5 Equalization using the MGS-MR algorithm ........ 214
References ................................................. 216
10 Precoding In large MIMO systems ............................ 219
10.1 Precoding in point-to-point MIMO ...................... 219
10.1.1 SVD precoding .................................. 220
10.1.2 Pairing of good and bad subchannels ............ 221
10.1.3 Performance of X-codes and Y-codes ............. 226
10.2 Precoding in a multiuser MIMO downlink ................ 227
10.2.1 Linear precoding ............................... 227
10.2.2 Non-linear precoding ........................... 229
10.2.3 Precoding in large multiuser MISO systems ...... 230
10.2.4 Precoder based on norm descent search (NDS) .... 233
10.2.5 Complexity and performance ..................... 236
10.2.6 Closeness to sum capacity ...................... 237
10.3 Multicell precoding ................................... 239
10.3.1 System model ................................... 241
10.3.2 Precoding without BS cooperation ............... 244
10.3.3 Precoding with BS cooperation .................. 245
10.3.4 Performance .................................... 246
References ................................................. 248
11 MIMO channel models ........................................ 251
11.1 Analytical channel models ............................. 252
11.1.1 Spatial correlation based models ............... 252
11.1.2 Propagation based models ....................... 256
11.2 Effect of spatial correlation on large MIMO
performance: an illustration .......................... 260
11.2.1 Pinhole effect ................................. 261
11.2.2 Effect of spatial correlation on LAS detector
performance .................................... 262
11.3 Standardized channel models ........................... 264
11.3.1 Models in IEEE 802.11 WiFi ..................... 265
11.3.2 Models in 3GPP/LTE ............................. 267
11.4 Large MIMO channel measurement campaigns .............. 268
11.5 Compact antenna arrays ................................ 275
11.5.1 PIFA ........................................... 276
11.5.2 PIFAs as elements in compact arrays ............ 277
11.5.3 MIMO cubes ..................................... 278
References ................................................. 279
12 Large MIMO testbeds ........................................ 285
12.1 12 × 12 point-to-point MIMO system .................... 286
12.2 8 × 16 point-to-point MIMO system at 10 Gbps rate ..... 287
12.3 16 × 16 multiuser MIMO system ......................... 287
12.4 64 × 15 muUiuser MIMO system (Argos) .................. 288
12.5 32 × 14 multiuser MIMO system (Ngara) ................. 290
12.6 Summary ............................................... 293
References ................................................. 293
Author index .................................................. 297
Subject index ................................................. 303
| 
