一篇博士论文Turbo-Equalization for QAM Constellation

scsunjia · 发表于 2008-11-1 09:47:30

马上注册，结交更多好友，享用更多功能，让你轻松玩转社区。

您需要登录才可以下载或查看，没有账号？注册

x

[size=120%]Turbo-Equalization for QAM Constellation
Author:  PAUL F.PETIT
Publisher: The University of South Australia
Number Of Pages: 229
Publication Date: 2002

Content:
1  Introduction
  1.1  Introduction ............................1
  1.2  Models and Assumptions .....................2
  1.3  Bandwidth efficiency .......................3
  1.4  Coding and Coded Modulation.................3
  1.5  Multipath fading............................4
  1.6  Monte-Carlo Simulation......................5
  1.7  Organisation of this thesis.................5
  1.8  Summary of Contributions ....................5

2 Equalization
  2.1  Introduction ................................7
  2.2  Probability of error for QAM, without ISI......8
  2.3  Channels and Channel models ..................10
  2.4  Linear Equalizers .........................10
      2.4.1 Symbol Spaced Linear Equalizer .............11
      2.4.2 Fractionally Spaced Linear Equalizer.......13
  2.5 Decision Feedback Equalizers....................14
      2.5.1 The Ilnbia,sed MMSE-DFE ................16
      2.5.2 Adaptive Behaviour of DFE ...............16
      2.5.3 Finite Length Filters ...................17
  2.6  Equalizer Simulation Results ...................17
  2.7  Error Propagation in DFE ....................20
      2.7.1 Analysis of Error Propagation for Large M......22
      2.7.2 Simulations Showing DFE Error Propagation......23
  2.8  Proposals to Reduce Error Propagation .............25
      2.8.1 Parallel DFE...............................27
      2.8.2 Precoding...................................27
  2.9  Summary..............................................28

3  Capacity and Losses of the ISI channel
  3.1  Introduction ............................29
  3.2  Channel Frequency Responses ..................30
  3.3  Capacity calculation .........................31
      3.3.1    The UMMSE-DFE and Capacity.............32
      3.3.2    Constrained Capacity and Shaping Gain .......34
      3.3.3    Coding lain ...........................35
  3.4  Zeros of ISI Channels .......................35
  3.5  Other measures of the ISI Channel..............35
      3.5.1 Intersymbol Interference Power...........35
      3.5.2 Channel Dispersion ....................36
  3.6  Minimum distance and bounds for MLSE ............36
  3.7  Summary ........................................39

4  Iterative Decoding and Turbo-Equalization
  4.1  Introduction ............................41
  4.2  Iterative Decoding.......................42
      4.2.1 Interleavers......................43
      4.2.2 Block Size .........................44
      4.2.3 Recursive Systematic Convolutional Codes .......45
      4.2.4 Other Aspects of Turbo Codes ..............45
      4.2.5 Spectrally Efficient Turbo-Coding Schemes .......47
      4.2.6 Analogue Turbo Decoding....................47
  4.3  MAP Decoding.............................48
      4.3.1 Description of the MAP Algorithm..............49
      4.3.2 Comparison with Sub-optimal Decoding Algorithms ...52
  4.4  Analysis of Convergence of Iterative Decoding.........53
  4.5  Using Lob-Likelihoods to Estimate the BER ..........54
  4.6  Review of literature on Turbo-Equalization ...........55
      4.6.1 Turbo-Equalization with Trellis-Based Equalizers........55
      4.6.2 Turbo-Equalization with Reduced-Complexity Equalizers ..57
      4.6.3 Turbo-Equalization with Iterative Decoding..............58
      4.6.4 Turbo-Equalization and Spectrally Efficient Modulation...59
  4.7  Summary .................................59

5  Reduced Complexity Turbo-Equalization
  5.1  Introduction ............................61
  5.2  Introduction to the GLL Turbo-Equalizer ............62
      5.2.1 Receiver Structure.........................63
      5.2.2 Interference Canceller ...................65
      5.2.3 Demapping.................................66
      5.2.4 Re-mapping.................................68
  5.3  Results and Observations on the GLL Scheme..........69
      5.3.1 Observations regarding remapping .............71
      5.3.2 BER before Decoding.........................72
      5.3.3 Some Difficult Channels ....................74
      5.3.4 SNR Threshold for Convergence ..............78
      5.3.5 Noise estimation after the Initial Equalizer .......80
  5.4  The performance of the Interference Canceller............81
      5.4.1 IC Noise Estimation ......................82
      5.4.2 Using the Max-Lob-MAP decoder .............85
      5.4.3 Analysis and Proposals to Improve the IC.......86
      5.4.4 Proposals for a Modified IC for TEQ..............87
      5.4.5 Other observations regarding IC modification ......88
      5.4.6 Using the DFE as the IC .................88
  5.5  Further Analysis of the GLL-TEQ ................89
      5.5.1 Exact versus approximate demapping..........89
      5.5.2 The DFE as Initial Equalizer.................94
      5.5.3 Predicting the SNR required for convergence ......94
  5.6  Turbo-equalization with concatenated codes................94
      5.6.1 Combination with serial concatenated codes..........94
      5.6.2 Combination with Differential Inner Encoder ......95
  5.7  Bit interleaving versus symbol interleaving...........96
      5.7.1 Interleaving with other constellations and codes ....99
  5.8  Summary and Conclusions ....................100

6  Reduced Complexity Trellis Equalizers
  6.1  Introduction ..........................103
  6.2  MLSE and MAP Equalizers ...............104
  6.3  SISO-MAP Algorithm .......................105
      6.3.1 SISO-MAP decoder ..................105
      6.3.2 MAP Equalizer ....................109
  6.4  Sub-optimal Trellis-Based Equalizers........110
  6.5  Reduced State and Reduced Search Detection....110
  6.6  The M-Algorithm................................112
      6.6.1 The M-Algorithm Versus RSSE ...........113
      6.6.2 The M-Algorithm Versus the T-Algorithm .....114
      6.6.3 The M-Algorithm Applied to ISI Channels .....114
  6.7  Soft Output M-Algorithm Equalizer .................115
      6.7.1  The SISO-MAP Version of M-Algorithm ......117
  6.8 Forward Filters for Reduced-Search Equalizers .......119
  6.9 Time-Reversal of Data..................................121
  6.10 Complexity, Speed and Delay............................121
  6.11 Estimating the Performance of Reduced-State Algorithms....122
  6.12 Summary and Conclusions .................................123

7  Turbo-equalization Using the M-Algorithm
  7.1  Introduction ............................125
  7.2  Structure of the MMAP Turbo-Equalizer ............126
      7.2.1 Prefilter for the MMAP Equalizer ............127
      7.2.2 Scaling Down the a priori Information .........127
  7.3  Full-State MAP-TEQ Performance ...............128
      7.3.1 Observations .......................132
      7.3.2 Performance Close to Capacity on Some Channels ...132
  7.4  Performance of MMAP Turbo-Equalizer ............132
      7.4.1 Equalizer and Decoder performance, without iteration...133
      7.4.2 Poor Quality Soft Outputs................133
      7.4.3 Using the Max-Lob-MAP Decoder ............136
      7.4.4 The Proportion of State Probability Rejected .. ...136
  7.5  Leakage of a priori Information .................137
      7.5.1 Indications of Leakage of API ..............139
      7.5.2 Modified MMAP Equalizer to Reduce Leakage .....139
      7.5.3 Leakage Prevention using the Lee Algorithm ......144
  7.6  Simulation Results with MMAP Equalizer ...........144
      7.6.1 Observations and Conclusions ..............152
  7.7  Some Strategies to Reduce Complexity .............153
      7.7.1 Further State Reduction for Later Iterations ......153
      7.7.2 Combination with an Interference Canceller .........154
      7.7.3 Separate I and Q Equalizers ...............154
  7.8  Strategies to Improve Performance ..................154
  7.9  Summary and Conclusions ....................155

8  Conclusions
  8.1  Summary of Results.......................157
  8.2  Areas for Further Research ....................159
  8.3  Concluding Remarks .......................159

A  Channel Data and Characteristics
  A.l  Introduction ............................161
  A.2  Time Domain Channel Impulse Responses ...........161
  A.3  Frequency Domain Channel Characteristics ...........162
  A.4  The SNR of the Unbiased MMSE-DFE .............162
  A.5  Zeros of Channels .........................168
  A.6  Partial Enemy of Channels ....................169

B  Bounds on DFE error propagation for QAM
  B.l  Introduction ..........................173
  B.2  Applying the bound to PAM .................174
  B.3  Bit Error rate for QAM ................... 176

C Additional Simulation results

D  Simulation details
  D.l  Introduction ............................181
  D.2  Simulation Software and Hardware ...............181
  D.3  LE and DFE Simulations .....................182
  D.4  Simulations for the GLL Turbo-Equalizer . ...........182
  D.5  Simulations for MMAP Turbo-Equalizers ............183

Bibliography

Turbo-Equalization for QAM Constellation.pdf (1.87 MB, 下载次数: 606 )