Title Page
Abstract
Contents
Chapter 1. Introduction 14
1.1. Background 14
1.1.1. Basics of Reconfigurable Intelligent Surfaces 15
1.1.2. Basics of Terahertz Communications 16
1.1.3. Basics of Compressed Sensing Technique 18
1.1.4. Basics of Riemannian Manifold Optimization Technique 20
1.2. Contribution and Organization 22
Chapter 2. Parametric Sparse Channel Estimation for RIS-Assisted Terahertz Systems 24
2.1. Introduction 25
2.2. RIS-Assisted Wideband THz System Model 29
2.2.1. RIS-Assisted Wideband THz System Model 29
2.2.2. Near-Field Effect of RIS-Assisted THz Systems 32
2.2.3. RIS-Assisted Wideband THz Channel Model 33
2.2.4. Frequency-Wideband Effect of RIS-Assisted THz Systems 35
2.3. Sparse Channel Estimation for Near-Field RIS-Assisted THz Systems 36
2.3.1. Polar-Domain Sparse Mapping 37
2.3.2. Block-Sparse Representation 39
2.3.3. Block-Sparse Recovery and Channel Reconstruction 41
2.3.4. Computational Complexity Analysis 42
2.4. Practical Issues in Near-Field RIS-Assisted THz Channel Estimation 44
2.4.1. Polar Bin Design for RIS-assisted THz Channel Estimation 44
2.4.2. Phase Shift Control for RIS-assisted THz Channel Estimation 48
2.5. Simulation Results 52
2.5.1. Simulation Setup 52
2.5.2. Simulation Results 53
2.6. Summary 58
2.7. Proofs 60
2.7.1. Proof of Lemma 2 60
2.7.2. Proof of Lemma 2 61
2.7.3. Proof of Proposition 2 63
2.7.4. Proof of Lemma 3 64
Chapter 3. Energy-Efficient Power Control and Beamforming for Reconfigurable Intelligent Surface-Aided Uplink IoT Networks 65
3.1. Introduction 66
3.2. RIS-Aided Uplink IoT System Model 69
3.2.1. RIS-Aided Uplink IoT Network 70
3.2.2. Uplink Channel Estimation 71
3.2.3. Uplink Data Transmission 74
3.2.4. Uplink Power Minimization Problem Formulation 75
3.3. Riemannian Conjugate Gradient-based joint optimization Algorithm 76
3.3.1. Notions on Riemannian Manifolds 77
3.3.2. Joint RIS Phase Shift and BS Beamforming Optimization on Product Manifold 79
3.3.3. Uplink Transmit Power Minimization 84
3.4. Convergence and Computational Complexity Analysis of RCG-JO 85
3.4.1. Convergence Analysis of RCG-JO Algorithm 86
3.4.2. Computational Complexity Analysis of RCG-JO Algorithm 88
3.5. Simulation Results 89
3.5.1. Simulation Setup 89
3.5.2. Simulation Results 92
3.6. Summary 98
3.7. Proofs 98
3.7.1. Proof of Proposition 1 98
3.7.2. Proof of Theorem 1 100
3.7.3. Proof of Proposition 2 101
3.7.4. Proof of Lemma 6 103
Chapter 4. Towards Reconfigurable Intelligent Surfaces-Assisted Wideband Beamforming for 6G 105
4.1. Introduction 105
4.2. RIS-Assisted Wideband THz Systems 107
4.2.1. RIS-Assisted THz System Model 107
4.2.2. Conventional True Time Delay-based Phase shifter 109
4.2.3. Data Rate Maximization Problem Formulation 110
4.3. RIS-Assisted Wideband Beamforming 111
4.3.1. Frequency-Dependent Beamforming Optimization 112
4.3.2. RIS Phase Shift Vector Optimization 115
4.4. Numerical Results 116
4.4.1. Simulation Setup 116
4.4.2. Simulation Results 117
4.5. Summary 119
4.6. Proofs 120
4.6.1. Proof of Lemma 1 120
Chapter 5. Conclusion 122
Bibliography 124
초록 133
Table 2.1. Comparison of computational complexity. 42
Table 3.1. System parameters 91
Figure 1.1. Employment of RIS in wireless communication networks. 15
Figure 1.2. Extreme requirements of 6G. 17
Figure 2.1. Near-field RIS-assisted THz system model. 29
Figure 2.2. RIS-assisted wideband THz channel estimation protocol. 30
Figure 2.3. Illustration of the near-field RIS-assisted THz channel. 32
Figure 2.4. Overall description of the proposed PF-RCE algorithm. 39
Figure 2.5. Correlation between the near-field array response vectors 46
Figure 2.6. Comparison of polar bins generated by (a) the proposed method (see (2.55) and (2.57)) and (b) the conventional method in [1] when Qθ=5, Qr=4, and rmin=4....[이미지참조] 48
Figure 2.7. NMSE vs. SNR. 52
Figure 2.8. NMSE vs. number of subframes. 53
Figure 2.9. NMSE vs. number of subcarriers. 54
Figure 2.10. NMSE vs. system bandwidth. 55
Figure 2.11. NMSE vs. UE-RIS distance. 56
Figure 2.12. NMSE vs. number of reflecting elements. 57
Figure 2.13. Block-mutual coherence vs. number of iterations. 58
Figure 2.14. NMSE vs. SNR. 59
Figure 2.15. NMSE vs. number of paths. 60
Figure 3.1. Illustration of the RIS-aided uplink IoT network. 69
Figure 3.2. Illustration of uplink RIS-aided channel estimation consisting of two major procedures: 1) estimation of the partial channel information observed at the active... 71
Figure 3.3. Illustration of (a) the tangent space TXY, (b) the Riemannian gradient gradYf(X), and (c) the retraction operator RX(V) at the point X in the Riemannian...[이미지참조] 78
Figure 3.4. Locations of BS, RIS, and IoT devices for K = 2. 89
Figure 3.5. Uplink transmit power vs. number of reflecting elements N. 92
Figure 3.6. Uplink transmit power vs. rate requirement of devices Rkmin.[이미지참조] 93
Figure 3.7. Uplink transmit power vs. BS-devices horizontal distance dh.[이미지참조] 94
Figure 3.8. Uplink transmit power vs. noise power σk2.[이미지참조] 95
Figure 3.9. Uplink transmit power vs. number of IoT devices K. 96
Figure 3.10. Uplink transmit power vs. percentage of active reflecting elements. 97
Figure 3.11. Cumulative distribution of the number of iterations required to converge. 97
Figure 4.1. Structure of a N-TTD array. 110
Figure 4.2. Overall structure of the RIS-assisted wideband beamforming (RWB) scheme. 111
Figure 4.3. Achievable data rate vs. noise power σn2.[이미지참조] 117
Figure 4.4. Achievable data rate vs. number of reflecting elements M. 118
Figure 4.5. Achievable data rate vs. number of subcarriers S. 119