Iterative Detection Book

	Preface	xi
	Introduction	xvii
1.	Overview of Non-Iterative Detection	1
1.1	Decision Theory Framework	1
1.1.1	The Bayes Decision Rule	3
1.1.2	Composite Hypothesis Testing	4
1.1.3	Statistical Sufficiency	6
1.2	MAP Symbol and Sequence Detection	7
1.2.1	The General Combining and Marginalization Problem and Semi-Ring Algorithms	15
1.2.2	Detection with Imperfect CSI	23
1.3	Data Detection for an FSM in Noise	28
1.3.1	Generic FSM Model	28
1.3.2	Perfect Channel State Information	31
1.3.3	Detection with Imperfect CSI	45
1.4	Performance Bounds Based on Pairwise Error Probability	49
1.4.1	An Upper Bound using Sufficient Neighborhood Sets	51
1.4.2	Lower Bounds Based on Uniform Side Information	54
1.4.3	An Upper Bound for MAP-SqD	59
1.4.4	A Lower Bound for MAP-SyD	64
1.5	Chapter Summary	67
1.6	Problems	68
2.	Principles of Iterative Detection	77
2.1	Optimal Detection in Concatenated Systems	77
2.2	The Marginal Soft-Inverse of a System	85
2.2.1	Some Common Subsystems	88
2.3	Iterative Detection Conventions	95
2.3.1	Summary of a General Iterative Detector	98
2.3.2	Explicit Index Block Diagrams	100
2.4	Iterative Detection Examples	101
2.4.1	Normalization Methods and Knowledge of the AWGN Noise Variance	101
2.4.2	Joint "Equalization" and Decoding	105
2.4.3	Turbo Codes	111
2.4.4	Multiuser Detection	120
2.5	Finite State Machines SISOs	128
2.5.1	The Forward-Backward Fixed-Interval SISO	130
2.5.2	Fixed-Lag SISOs	131
2.5.3	Forward-Only (L[superscript 2]VS) FL-SISO	133
2.5.4	Sliding Window SISOs	136
2.5.5	A Tree-Structured SISO	138
2.5.6	Variations on Completion and Combining Windows	142
2.5.7	Soft-Output Viterbi Algorithms	143
2.6	Message Passing on Graphical Models	144
2.6.1	Optimality Conditions for Message Passing	146
2.6.2	Revisiting the Iterative Detection Conventions	156
2.6.3	Valid Configuration Checks	161
2.6.4	Other Graphical Models	169
2.7	On the Non-uniqueness of an Iterative Detector	175
2.7.1	Additional Design Guidelines	181
2.8	Summary and Open Problems	182
2.9	Problems	184
3.	Iterative Detection for Complexity Reduction	193
3.1	Complexity Reduction Tools	193
3.1.1	Operation Simplification	194
3.1.2	Decision Feedback Techniques	194
3.2	Modified Iterative Detection Rules	199
3.2.1	Altering the Convergence Rate	199
3.2.2	Modified Initialization Schemes	202
3.3	A Reduced-State SISO with Self-Iteration	204
3.3.1	Reduced-State SISO Algorithm	205
3.3.2	Example Applications of the RS-SISO	209
3.4	A SISO Algorithm for Sparse ISI Channels	213
3.4.1	Sparse ISI Channel	213
3.4.2	Existing Algorithms for S-ISI Channels	217
3.4.3	The Sparse SISO Algorithms for S-ISI Channels	218
3.4.4	Features of the S-SISOs	223
3.4.5	Design Rules for the S-SISO Algorithms	223
3.4.6	Using the Sparse SISO Algorithms	229
3.4.7	On Performance Bounds for S-ISI Channels	231
3.5	Summary and Open Problems	234
3.6	Problems	235
4.	Adaptive Iterative Detection	239
4.1	Exact Soft Inverses--Optimal Algorithms	242
4.1.1	Separate Sequence and Parameter Marginalization	243
4.1.2	Joint Sequence and Parameter Marginalization	244
4.2	Approximate Soft Inverses--Adaptive SISO Algorithms	246
4.2.1	Separate Sequence and Parameter Marginalization	246
4.2.2	Joint Sequence and Parameter Marginalization	248
4.2.3	Fixed-Lag Algorithms	250
4.2.4	Forward Adaptive and Forward-Backward Adaptive Algorithms	252
4.3	TCM in Interleaved Frequency-Selective Fading Channels	253
4.4	Concatenated Convolutional Codes with Carrier Phase Tracking	259
4.4.1	SCCC with Carrier Phase Tracking	259
4.4.2	PCCC with Carrier Phase Tracking	262
4.5	Summary and Open Problems	268
4.6	Problems	269
5.	Applications in Two Dimensional Systems	273
5.1	Two Dimensional Detection Problem	273
5.1.1	System Model	273
5.1.2	Optimal 2D Data Detection	274
5.2	Performance Bounds for Optimal 2D Detection	276
5.2.1	Finding Small Distances	281
5.3	Iterative 2D Data Detection Algorithms	283
5.3.1	Iterative Concatenated Detectors	283
5.3.2	Distributed 2D SISO Algorithms	290
5.4	Data Detection in POM Systems	294
5.4.1	POM System Model	294
5.4.2	Existing Detection Algorithms	296
5.4.3	The Performance of Iterative Detection Algorithms	297
5.5	Digital Image Halftoning	300
5.5.1	Baseline Results	301
5.5.2	Random Biasing	302
5.5.3	Larger Filter Support Regions	306
5.5.4	High Quality and Low Complexity using 2D-GM2	307
5.6	Summary and Open Problems	308
5.7	Problems	310
6.	Implementation Issues: a Turbo Decoder Design Case Study	315
6.1	Quantization Effects and Bitwidth Analysis	316
6.1.1	Quantization of Channel Metrics	316
6.1.2	Bitwidth Analysis of the Forward/Backward State Metrics	320
6.1.3	Soft-Out Metric Bitwidths	323
6.2	Initialization of State Metrics	326
6.3	Interleaver Design and State Metric Memory	328
6.4	Determination of Clock Cycle Time and Throughput	330
6.5	Advanced Design Methods	333
6.5.1	Block-level Parallelism	333
6.5.2	Radix-4 SISO Architectures	334
6.5.3	Fixed and Minimum Lag SISOs	335
6.5.4	Minimum Half Window (Tiled) SISOs	335
6.5.5	Sliding Window SISOs	336
6.5.6	Tree SISOs	336
6.5.7	Low-Power Turbo Decoding	337
6.6	Problems	338
	References	341
	Index	357