Title Page
Abstract
Contents
Chapter 1. Introduction 18
1.1. Background 18
1.2. Motivations 19
1.2.1. Novel Vector Quantization Algorithm 21
1.2.2. Novel Codebook Generation Criteria 22
1.2.3. Practical Implementation of Novel Codebooks for Interference Reduction in Het-Nets 23
1.3. Dissertation Organization 25
Chapter 2. State of the Art of Codebook Design Technologies 27
2.1. Problem Statement 28
2.2. Classifications of Codebooks 29
2.2.1. Online Codebook vs. Offline Codebook 29
2.2.2. Scalar Quantization vs. Vector Quantization 31
2.2.3. Direct Quantization vs. Indirect Quantization 33
2.3. Open Problems of Existing Codebook Design Strategies 34
Chapter 3. Vector Quantization for Limited Feedback Generation 36
3.1. Limited Feedback in Wireless Communication Systems 37
3.1.1. Why Limited Feedback? 37
3.1.2. Usage of Limited Feedback 37
3.1.3. Generation of Limited Feedback 39
3.2. Basic Concept of VQ Algorithms 40
3.3. Conventional VQ Algorithms 43
3.3.1. The LBG Algorithm 43
3.3.2. The Enhanced LBG Algorithm 45
3.4. The Proposed VQ Algorithms 48
3.4.1. Initial Codebook Generation Algorithm using Linear Interpolation 48
3.4.2. The Proposed ELBG-M VQ Algorithm 51
3.4.3. Considerations and Discussions 56
3.5. Simulation Results and Comparisons 64
Chapter 4. Codebook Design for Precoding in MIMO Interference Channels 67
4.1. Codebook Design Strategies using Scalar Quantization 68
4.1.1. Amplitude Quantization of Channel Elements 69
4.1.2. Phase Quantization of Channel Elements 73
4.1.3. Achievable Rate of MISO Beamforming with the Proposed SQ Codebooks 73
4.1.4. Numerical Results of MISO Beamforming with the Proposed SQ Codebooks and Discussions 78
4.1.5. Simulation Results of MIMO Precoding Employing the Proposed SQ Codebooks 80
4.2. Codebook Design Strategies using Vector Quantization 84
4.2.1. Codebook Design Criterion for Singular Values of MIMO Channel Matrix 86
4.2.2. Codebook Design Criterion for Singular Matrices of MIMO Channel Matrix 91
4.2.3. Achievable Rate of MIMO Precoding Systems with the Proposed VQ Codebook 94
4.2.4. Numerical Results of MIMO Precoding Systems with the Proposed VQ Codebook and Discussions 105
4.3. Minimization of the Number of Feedback Bits 107
Chapter 5. Interference Reduction with Limited Feedback in LTE-A Het-Net 116
5.1. Network Scenarios and Dominant Interference 116
5.1.1. Inter-cell Interference in Cellular Networks 118
5.1.2. Intra-cell Interference in Het-Nets 118
5.1.3. Interference with CCC and Stack RUs 120
5.2. Codebook Design Strategies for Different Networks 121
5.3. Interference Reduction with the Proposed Codebooks 122
5.3.1. Simulation Results in Cellular Networks 122
5.3.2. Simulation Results in Het-Nets 130
5.3.3. Simulation Results with CCC and Stack RUs 138
5.4. Feasibility of the Proposed Codebooks in Practical Systems 141
5.4.1. Performance Enhancement by Employing the Proposed Codebooks 141
5.4.2. Comparison of Computational Complexity 145
Chapter 6. Conclusions and Future Works 148
6.1. Conclusions 148
6.2. Future Works 150
참고문헌 153
Table 3.1. Comparison of execution time 61
Table 3.2. Comparison of iteration numbers 61
Table 5.1. General simulation parameters 123
Table 5.2. SQ performance comparison 129
Table 5.3. VQ performance comparison 131
Table 5.4. General simulation parameters 132
Table 5.5. Comparison of SQ and VQ performance 140
Table 5.6. Comparison between LTE codebook and the proposed codebooks 145
Figure 2.1. An illustration of channel quantization 28
Figure 2.2. General classification of codebooks 30
Figure 2.3. General codebook generation procedure 34
Figure 3.1. Block diagram of SU-MIMO system with limited feedback 38
Figure 3.2. High-level flow chart of the LBG algorithm 45
Figure 3.3. High-level flow chart of the optimization step of ELBG algorithm 46
Figure 3.4. High-level flow chart of the optimization step of ELBG algorithm 49
Figure 3.5. High-level flow chart of the proposed VQ algorithm 54
Figure 3.6. Detailed flow chart of "ELBG-M" block in codebook optimization step 55
Figure 3.7. MQE performance of the LBG algorithm adopting the proposed RCG with different ρ values 60
Figure 3.8. MQE performance of the ELBG-M algorithm adopting with different γ values 60
Figure 3.9. High-level flow chart of combined VQ scheme 62
Figure 3.10. MQE performance of various of algorithms with codebook size of 16 65
Figure 3.11. MQE performance of various of algorithms with codebook size of 64 65
Figure 3.12. MQE performance of various of algorithms with codebook size of 256 66
Figure 4.1. Block diagram of MU-MIMO system with limited feedback 69
Figure 4.2. Data rate of scenario 1 for MISO beamforming systems 79
Figure 4.3. Data rate of scenario 2 & 3 for MISO beamforming systems 79
Figure 4.4. Basic scenarios of BER comparison of ZF precoding with SQ codebooks 82
Figure 4.5. BER comparison of ZF precoding with SQ codebooks for amplitude and phase respectively 82
Figure 4.6. BER comparison of ZF precoding with SQ codebooks when codebook size is varying 83
Figure 4.7. BER comparison of ZF precoding with SQ codebooks for LASQ+UPQ and NLASQ+UPQ 83
Figure 4.8. Joint distribution of singular values in 4? MIMO system 87
Figure 4.9. Distribution of the smallest singular value in 4? MIMO system and its Gaussian approximation 88
Figure 4.10. Distribution of the largest singular value in 4? MIMO system and its Gaussian approximation 88
Figure 4.11. Block diagram of a MIMO system employing linear precoding matrix F with limited feedback 92
Figure 4.12. Codeword partitions of the proposed codebook in M dimensional space 98
Figure 4.13. Ergodic rate of the spatial multiplexing system employing the proposed VQ codebook 106
Figure 4.14. Comparison between SQ codebooks and VQ codebooks 107
Figure 4.15. An example of 3-bit index for representing 8 different states 108
Figure 4.16. An example of 2/3-bit index for representing 12 different states 108
Figure 5.1. Single-class network model with two neighbor cells 117
Figure 5.2. Single-class network model with three neighbor cells 117
Figure 5.3. Overview of Het-Net model 119
Figure 5.4. Co-existence of a microcell and a femtocell 119
Figure 5.5. BER performance of SQ system employing BD with ZF detection in two-cell networks 124
Figure 5.6. BER performance of SQ system employing THP in two-cell and three-cell networks 124
Figure 5.7. BER performance of SQ system employing BD with THP in two-cell networks 125
Figure 5.8. BER performance of SQ system employing BD with ZF detection in three-cell networks 127
Figure 5.9. BER performance of SQ system employing BD with THP in three-cell networks 127
Figure 5.10. BER performance of VQ codebook employing BD with THP in two-cell and three-cell networks 130
Figure 5.11. Simplified Het-Net model 133
Figure 5.12. BER performance of SQ codebook employing BD in Het-Net 136
Figure 5.13. BER performance of SQ codebook employing BD with THP in Het-Net 136
Figure 5.14. Performance comparison between of SQ and VQ codebooks employing BD with THP in Het-Net 137
Figure 5.15. BER performance of SQ codebook employing BD with THP 139
Figure 5.16. BER performance of VQ codebook employing BD with THP 139
Figure 5.17. Performance comparison between GSP codebook and the proposed EDVQUS codebook 142
Figure 5.18. Performance comparison of various of MIMO techniques with perfect CSIT 142
Figure 5.19. Performance comparison between SQ codebooks and LTE codebook 144
Figure 5.20. Performance comparison between VQ codebooks and LTE codebook 144