본문 바로가기 주메뉴 바로가기
국회도서관 홈으로 정보검색 소장정보 검색

결과 내 검색

동의어 포함

목차보기

Title Page

Contents

Abstract 16

Chapter 1. Introduction 20

1.1. Background 20

1.2. Outline and Contributions 23

1.3. Common Notations 26

Chapter 2. A Vector Perturbation with User Selection for Multiuser MIMO Downlink 27

2.1. Introduction 27

2.2. System Model and Previous Works 28

2.3. Vector Perturbation with Virtual Users 31

2.3.1. Motivation 31

2.3.2. Virtual User Selection 32

2.3.3. Modified Sphere Encoding for Searching Optimal γ 35

2.3.4. Lower Bound of the Sum Rate Loss 40

2.3.5. Comments on Complexity 41

2.3.6. Scheduling for Fairness 43

2.4. Extension to Multi-user MIMO Downlink Scenario 44

2.5. Simulation Results and Discussions 49

2.6. Chapter Summary and Conclusions 55

Chapter 3. A MMSE Vector Precoding with Block Diagonalization for Multiuser MIMO Downlink 56

3.1. Introduction 56

3.2. Multi-user MIMO Downlink 57

3.2.1. System Model 57

3.2.2. BD Algorithm 58

3.2.3. BD-VP Algorithm 60

3.3. BD-MVP 61

3.3.1. The BD-MVP 61

3.3.2. BD-MVP with Geometric Mean Decomposition 64

3.3.3. The Sum Rate of the BD-MVP 65

3.3.4. MSE Performance between the BD-VP and BD-MVP 67

3.3.5. Comments on Complexity 68

3.3.6. Proof of Joint Optimization Problem 69

3.3.7. Proof of (3.16) 71

3.4. Simulation Results and Discussions 72

3.5. Chapter Summary and Conclusions 76

Chapter 4. Antenna Grouping based Feedback Reduction Technique 77

4.1. Introduction 77

4.2. MIMO Beamforming 79

4.2.1. System Model and Conventional Beamforming 79

4.2.2. Limited Feedback 82

4.3. Antenna Grouping based Feedback Reduction 83

4.3.1. AGB Algorithm 83

4.3.2. Antenna Group Pattern Generation 87

4.3.3. Quantization Distortion Analysis 89

4.3.4. Further Application of AGB Algorithm 97

4.4. Simulations 99

4.4.1. Simulation Setup 99

4.4.2. Simulation Results 100

4.5. Chapter Summary and Conclusions 104

Chapter 5. Antenna Group Selection based User Scheduling for Massive MIMO Systems 112

5.1. Introduction 112

5.2. System Model 113

5.3. Antenna Group Selection based User Scheduling 115

5.3.1. AGS Algorithm 116

5.3.2. Codebook Generation at User Terminal 117

5.3.3. Basestation Precoding 118

5.4. Sum Rate Analysis of Antenna Group Selection 118

5.4.1. Ergodic Sum Rate Analysis under Limited Feedback Systems 121

5.4.2. Comments on Scheduling for Fairness 123

5.5. Simulations 125

5.6. Chapter Summary and Conclusions 129

Chapter 6. Conclusions 130

Acronyms 132

Bibliography 134

List of Tables

Table 2.1: Operation of the proposed method. 39

Table 2.2: Computational complexity (flops) of the proposed method and standard vector perturbation (average over 10?channel realizations). 42

Table 2.3: Modified PFS algorithm. 45

Table 3.1: Time complexity for {4, 4} x 8 system (running time for 10⁴ symbols). 75

Table 4.1: Summary of the antenna group pattern generation. 90

Table 4.2: Singular value ratio vs. Nt and p.(이미지참조) 95

Table 4.3: Pattern distribution per fading block on temporally and spatially correlated channels (Nt = 8,Ng = 4, Bp=5).(이미지참조) 104

List of Figures

Figure 2.1: Illustration of γ reduction using the vector perturbation. Lattice points with gray dots represent s + τℓ' in (a) and P(s + τℓ') in (b). We can observe from (b) that ℓ = -1 provides smaller γ than ℓ = 0. 30

Figure 2.2: a) E[ץ ]of the proposed method using (2.10) and (2.11) and Na,min = 3 and Nc = 10) as well as standard vector perturbation for 4 x 4 multi-user MISO downlink system. (b) E[γ] of the proposed method as a function of the number of virtual candidates Nc (SNR=...(이미지참조) 34

Figure 2.3: Performance of 4 x 4 multi-user downlink system (є = αRrg)(이미지참조) 50

Figure 2.4: Performance of 4 x 4 multi-user downlink system with Nv variation (є = 0.15Rorg)(이미지참조) 51

Figure 2.5: Performance of 4 x 4 multi-user downlink system with various σe2,h.(이미지참조) 52

Figure 2.6: Performance of {4, 4} x 8 multi-user downlink system (є₁ = є₂ = 0.15R1,org)(이미지참조) 54

Figure 3.1: The transceiver structure of the proposed BD-MVP technique. 62

Figure 3.2: Achievable sum rates of DPC, conventional BD algorithms (BD, BD-WF and BD-VP), THP based schemes (BD-GMD and BD-UCD), and the proposed BD-MVP. 73

Figure 3.3: BER performance for {4, 4} x 8 multi-user MIMO systems. 74

Figure 4.1: CSI feedback in the multi-user downlink system. 80

Figure 4.2: Illustration of the AGB algorithm for Nt = 8, Ng=4. The reduced dimension channel vector hr is obtained by expressing subgroup elements of h as a representative value. Note that ^hr is the quantized version of ~hr(i) and hr is expanded from ^hr in order to measure(이미지참조) 84

Figure 4.3: Feedback packet structure 85

Figure 4.4: Example of antenna group patterns (Nt = 16, Ng = 8, Np = 3). Antenna elements belonging to the same pattern are mapped to one representative value.(이미지참조) 86

Figure 4.5: A block diagram of the AGB algorithm in the multi-user downlink system. 87

Figure 4.7: Sum rate as a function of the correlation coefficient α. 106

Figure 4.8: Sum rate as a function of the number of feedback bits B (Nt = 32, Ng=24).(이미지참조) 107

Figure 4.9: Sum rate as a function of the number of transmit antennas Nt (B = Nt, Ng = Nt/2).(이미지참조) 108

Figure 4.10: Sum rate comparison between the NTCQ and analytical approximation of RVQ. 109

Figure 4.11: Average beamforming gain as a function of number of feedback bits (a) ULA channel (b) UPA channel. 110

Figure 4.12: Sum rate as a function of the number of pattern bits (Nt = 16, Ng=8).(이미지참조) 111

Figure 5.1: Example of antenna groups (Nt = 12, Ng = 3, J = 4).(이미지참조) 115

Figure 5.2: CSI feedback in the antenna group selection based user scheduling in the multi-user downlink systems. 116

Figure 5.3: Sum rate as a function of the number of feedback bits B when Nt = 32, Ng = 8, p = 10 dB.(이미지참조) 124

Figure 5.4: Sum rate as a function of the number of antenna elements per group Ng (Nt = 32, K = 4).(이미지참조) 127

Figure 5.5: Sum rate as a function of the number of feedback bits B (Nt = 64, Ng = 16, K=4).(이미지참조) 128