5G and Beyond Wireless Communication Networks
Description
A comprehensive and up-to-date survey of 5G technologies and applications
In 5G and Beyond Wireless Communication Networks, a team of distinguished researchers deliver an expert treatment of the technical details of modern 5G wireless networks and the performance gains they make possible. The book examines the recent progress in research and development in the area, covering related topics on fundamental 5G requirements and its enabling technologies.
The authors survey 5G service architecture and summarize enabling technologies, including highly dense small cell and heterogeneous networks, device-to-device communications underlaying cellular networks, fundamentals of non-orthogonal multiple access in 5G new radio and its applications. Readers will also find:
* A thorough introduction to 5G wireless networks, including discussions of anticipated growth in mobile data traffic
* Comprehensive explorations of dense small cell and heterogeneous networks
* Practical discussions of the most recent developments in 5G research and enabling technologies
* Recent advancement of non-orthogonal multiple access and its role in current and future wireless systems
Perfect for graduate students, professors, industry professionals, and engineers with an interest in wireless communication, 5G and Beyond Wireless Communication Networks will also benefit undergraduate and graduate students and researchers seeking an up-to-date and accessible new resource about 5G networks.
More details
Other editions
Additional editions
Persons
Haijian Sun, PhD, is an Assistant Professor in the School of Electrical and Computer Engineering at the University of Georgia in Athens, USA. His research interests include wireless communications for 5G and beyond, efficient edge computing, wireless security, and wireless for distributed learning.
Rose Qingyang Hu, PhD, is a Professor in the Department of Electrical and Computer Engineering at Utah State University in Logan, USA. Her research interests include next-generation wireless communications, wireless network design and optimization, and more.
Yi Qian, PhD, is a Professor in the Department of Electrical and Computer Engineering at the University of Nebraska-Lincoln in Omaha, USA. His research interests include cyber security and communication network security, computer networks, and wireless networks.
Content
About the Authors xi
Preface xiii
Acknowledgments xv
1 Introduction to 5G and Beyond Network 1
1.1 5G and Beyond System Requirements 1
1.1.1 Technical Challenges 2
1.2 Enabling Technologies 3
1.2.1 5G New Radio 3
1.2.1.1 Non-orthogonal Multiple Access (NOMA) 3
1.2.1.2 Channel Codes 5
1.2.1.3 Massive MIMO 5
1.2.1.4 Other 5G NR Techniques 6
1.2.2 Mobile Edge Computing (MEC) 6
1.2.3 Hybrid and Heterogeneous Communication Architecture for Pervasive IoTs 7
1.3 Book Outline 8
2 5G Wireless Networks with Underlaid D2D Communications 11
2.1 Background 11
2.1.1 MU-MIMO 11
2.1.2 D2D Communication 11
2.1.3 MU-MIMO and D2D in 5G 12
2.2 NOMA-Aided Network with Underlaid D2D 12
2.3 NOMA with SIC and Problem Formation 14
2.3.1 NOMA with SIC 14
2.3.2 Problem Formation 15
2.4 Precoding and User Grouping Algorithm 15
2.4.1 Zero-Forcing Beamforming 16
2.4.1.1 First ZF Precoding 16
2.4.1.2 Second ZF Precoding 16
2.4.2 User Grouping and Optimal Power Allocation 16
2.4.2.1 First ZF Precoding 17
2.4.2.2 Second ZF Precoding 18
2.5 Numerical Results 18
2.6 Summary 19
3 5G NOMA-Enabled Wireless Networks 21
3.1 Background 21
3.2 Error Propagation in NOMA 22
3.3 SIC and Problem Formulation 22
3.3.1 SIC with Error Propagation 23
3.3.2 Problem Formation 24
3.4 Precoding and Power Allocation 25
3.4.1 Precoding Design 25
3.4.2 Case Studies for Power Allocation 26
3.4.2.1 Case I 26
3.4.2.2 Case II 27
3.5 Numerical Results 27
3.6 Summary 30
4 NOMA in Relay and IoT for 5G Wireless Networks 31
4.1 Outage Probability Study in a NOMA Relay System 31
4.1.1 Background 31
4.1.2 System Model 32
4.1.2.1 NOMA Cooperative Scheme 32
4.1.2.2 NOMA TDMA Scheme 34
4.1.3 Outage Probability Analysis 35
4.1.3.1 Outage Probability in NOMA Cooperative Scheme 35
4.1.4 Outage Probability in NOMA TDMA Scheme 36
4.1.5 Outage Probability with Error Propagation in SIC 37
4.1.5.1 Outage Probability in NOMA Cooperative Scheme with EP 38
4.1.5.2 Outage Probability in NOMA TDMA Scheme with EP 38
4.1.6 Numerical Results 39
4.2 NOMA in a mmWave-Based IoT Wireless System with SWIPT 41
4.2.1 Introduction 41
4.2.2 System Model 41
4.2.2.1 Phase 1 Transmission 42
4.2.2.2 Phase 2 Transmission 44
4.2.3 Outage Analysis 45
4.2.3.1 UE 1 Outage Probability 45
4.2.3.2 UE 2 Outage Probability 45
4.2.3.3 Outage at High SNR 47
4.2.3.4 Diversity Analysis for UE 2 47
4.2.4 Numerical Results 47
4.2.5 Summary 48
5 Robust Beamforming in NOMA Cognitive Radio Networks: Bounded CSI 51
5.1 Background 51
5.1.1 RelatedWork and Motivation 52
5.1.1.1 Linear EH Model 52
5.1.1.2 Non-linear EH Model 53
5.1.2 Contributions 53
5.2 System and Energy Harvesting Models 54
5.2.1 System Model 54
5.2.2 Non-linear EH Model 55
5.2.3 Bounded CSI Error Model 55
5.2.3.1 NOMA Transmission 56
5.3 Power Minimization-Based Problem Formulation 56
5.3.1 Problem Formulation 57
5.3.2 Matrix Decomposition 59
5.4 Maximum Harvested Energy Problem Formulation 60
5.4.1 Complexity Analysis 61
5.5 Numerical Results 62
5.5.1 Power Minimization Problem 62
5.5.2 Energy Harvesting Maximization Problem 64
5.6 Summary 67
6 Robust Beamforming in NOMA Cognitive Radio Networks: Gaussian CSI 69
6.1 Gaussian CSI Error Model 69
6.2 Power Minimization-Based Problem Formulation 69
6.2.1 Bernstein-Type Inequality I 70
6.2.2 Bernstein-Type Inequality II 71
6.3 Maximum Harvested Energy Problem Formulation 72
6.3.1 Complexity Analysis 73
6.4 Numerical Results 73
6.4.1 Power Minimization Problem 74
6.4.2 Energy Harvesting Maximization Problem 76
6.5 Summary 79
7 Mobile Edge Computing in 5G Wireless Networks 81
7.1 Background 81
7.2 System Model 82
7.2.1 Data Offloading 83
7.2.2 Local Computing 83
7.3 Problem Formulation 83
7.3.1 Update pk, tk, and fk 85
7.3.2 Update Lagrange Multipliers 86
7.3.3 Update Auxiliary Variables 86
7.3.4 Complexity Analysis 87
7.4 Numerical Results 87
7.5 Summary 90
8 Toward Green MEC Offloading with Security Enhancement 91
8.1 Background 91
8.2 System Model 92
8.2.1 Secure Offloading 92
8.2.2 Local Computing 93
8.2.3 Receiving Computed Results 93
8.2.4 Computation Efficiency in MEC Systems 93
8.3 Computation Efficiency Maximization with Active Eavesdropper 94
8.3.1 SCA-Based Optimization Algorithm 94
8.3.2 Objective Function 95
8.3.3 Proposed Solution to P4 with given (¿¿, ß¿) 96
8.3.4 Update (¿¿, ß¿) 97
8.4 Numerical Results 97
8.5 Summary 100
9 Wireless Systems for Distributed Machine Learning 101
9.1 Background 101
9.2 System Model 102
9.2.1 FL Model Update 102
9.2.2 Gradient Quantization 104
9.2.3 Gradient Sparsification 104
9.3 FL Model Update with Adaptive NOMA Transmission 104
9.3.1 Uplink NOMA Transmission 104
9.3.2 NOMA Scheduling 105
9.3.3 Adaptive Transmission 106
9.4 Scheduling and Power Optimization 107
9.4.1 Problem Formulation 107
9.5 Scheduling Algorithm and Power Allocation 108
9.5.1 Scheduling Graph Construction 108
9.5.2 Optimal scheduling Pattern 109
9.5.3 Power Allocation 110
9.6 Numerical Results 111
9.7 Summary 114
10 Secure Spectrum Sharing with Machine Learning: An Overview 115
10.1 Background 115
10.1.1 SS: A Brief History 116
10.1.2 Security Issues in SS 118
10.2 ML-Based Methodologies for SS 119
10.2.1 ML-Based CRN 119
10.2.1.1 Spectrum Sensing 120
10.2.1.2 Spectrum Selection 122
10.2.1.3 Spectrum Access 123
10.2.1.4 Spectrum Handoff 125
10.2.2 Database-Assisted SS 125
10.2.2.1 ML-Based EZ Optimization 126
10.2.2.2 Incumbent Detection 126
10.2.2.3 Channel Selection and Transaction 127
10.2.3 ML-Based LTE-U/LTE-LAA 127
10.2.3.1 ML-Based LBT Methods 128
10.2.3.2 ML-Based Duty Cycle Methods 129
10.2.3.3 Game-Theory-Based Methods 129
10.2.3.4 Distributed-Algorithm-Based Methods 130
10.2.4 Ambient Backscatter Networks 131
10.2.4.1 Information Extraction 131
10.2.4.2 Operating Mode Selection and User Coordination 132
10.2.4.3 AmBC-CR Methods 133
10.3 Summary 134
11 Secure Spectrum Sharing with Machine Learning: Methodologies 135
11.1 Security Concerns in SS 135
11.1.1 Primary User Emulation Attack 135
11.1.2 Spectrum Sensing Data Falsification Attack 135
11.1.3 Jamming Attacks 136
11.1.4 Intercept/Eavesdrop 137
11.1.5 Privacy Issues in Database-Assisted SS Systems 137
11.2 ML-Assisted Secure SS 138
11.2.1 State-of-the-Art Methods of Defense Against PUE Attack 138
11.2.1.1 ML-Based Detection Methods 138
11.2.1.2 Robust Detection Methods 140
11.2.1.3 ML-Based Attack Methods 141
11.2.2 State-of-the-Art Methods of Defense Against SSDF Attack 142
11.2.2.1 Outlier Detection Methods 143
11.2.2.2 Reputation-Based Detection Methods 143
11.2.2.3 SSDF and PUE Combination Attacks 144
11.2.3 State-of-the-Art Methods of Defense Against Jamming Attacks 144
11.2.3.1 ML-Based Anti-Jamming Methods 145
11.2.3.2 Attacker Enhanced Anti-Jamming Methods 146
11.2.3.3 AmBC Empowered Anti-Jamming Methods 148
11.2.4 State-of-the-Art Methods of Defense Against Intercept/Eavesdrop 149
11.2.4.1 RL-Based Anti-Eavesdropping Methods 149
11.2.5 State-of-the-Art ML-Based Privacy Protection Methods 150
11.2.5.1 Privacy Protection for PUs in SS Networks 150
11.2.5.2 Privacy Protection for SUs in SS Networks 151
11.2.5.3 Privacy Protection for ML Algorithms 151
11.3 Summary 153
12 Open Issues and Future Directions for 5G and Beyond Wireless Networks 155
12.1 Joint Communication and Sensing 155
12.2 Space-Air-Ground Communication 155
12.3 Semantic Communication 156
12.4 Data-Driven Communication System Design 156
Appendix A Proof of Theorem 5.1 157
Bibliography 161
Index 181
1
Introduction to 5G and Beyond Network
We have witnessed an unprecedented development of wireless technology for the past few decades. Starting from 1980s, when the first mobile phone was released, major wireless technology advanced almost every decade. From first generation (1G) to 4G. The invention of smart devices, such as phones, tablets, and home appliances, is the main driving force for the ever-increasing mobile traffic today. It is not surprising that mobile traffic increased 10-fold between 2014 and 2019 globally. The mobile data traffic is expected to grow much faster than fixed IP traffic in the upcoming years [34]. Wireless technologies dramatically changed the way people interact, communicate, and collaborate, especially at post-Covid era. The need for faster, more efficient and secure, and intelligent communication technique remains strong. While the current wireless communication systems such as 4G long term evolution (LTE) have been pushed to their theoretic capacity limit, different air interface and radio access technologies including heterogeneous network (HetNet) [76, 77], multiuser multi-input multi-output (MU-MIMO) [105], and device-to-device (D2D) communication [51] have become potential paradigms to fulfill the gap between demands from end users and the capacity that current air interface can provide.
1.1 5G and Beyond System Requirements
In their pioneering work [10], Andrews et al. evaluated the requirements for 5G. In short, 5G wireless communication system should provide 1,000 times aggregate data improvement over 4G, support for as low as 1 ms round-trip latencies, 10 times longer battery life for low-power devices, and also support 10,000 times or more low-rate devices in a single macro cell, see Figure 1.1 for a brief illustration. Due to those high requirements, the transformation from 4G to 5G cannot be simply fulfilled by extensions of current technologies. In general, 5G and beyond system should support or deliver the following aspects. Notably, (i) more bandwidth. Currently commercial cellular systems use frequencies below 6 GHz (sub-6 GHz); in fact, there is abundant bandwidth in the millimeter-wave (mmWave) band, for example in 28 GHz and above, which can provide more bandwidth that previously have not been applied in cellular networks. (ii) More antennas. Higher frequency also brings smaller form factor of large antenna arrays. Additionally, the signal processing techniques in terms of massive MIMO and transceiver design also improved significantly. (iii) New radios (NR). The physical layer in 5G will change dramatically, specifically the 5G NR, which includes the new multiple access technology, the new air interface, and a combination of several existing techniques. (iv) New schemes. It is expected that ultra dense networks (UDN) will be heavily deployed. The density of small base station (BS), such as micro BS, femto cell, and pico cells, will be much higher than that in 4G. But they share the similarity in terms of deploying BSs with different powers to provide seamless coverage, as well as performance improvements from short-range communications. (v) High intelligence. It is expected that beyond 5G systems should support higher level of intelligence. Emerging applications such as Artificial intelligence (AI), semantic communication, and robots will surely benefit from AI-friendly wireless technology. (vi) Pervasive wireless. It is anticipated that each person will carry more personal devices for enhanced life style and health monitoring. To support ubiquitous wireless connectivity, those devices need be connected. Current network architecture can hardly support such high number of devices simultaneously.
Figure 1.1 Four main goals for 5G.
1.1.1 Technical Challenges
The above promising technologies are able to deliver ambitious goals of 5G, but they ultimately encounter some challenges. First of all, even though high-frequency bands have major vacancy, mmWave signals are notorious for weak penetration and vulnerable blockage; hence, the transmission characteristics are big concerns. Moreover, studies also have shown mmWave signals have high attenuation due to atmospheric gaseous, rain, concrete structure, glasses, even foliage. The real-world deployment of such mmWave systems needs to be carefully studied and planned. Secondly, from the transceiver design perspective, higher-frequency signals impose challenges in circuit design, materials, and heating issues. Nyquist theorem sets the lower boundary for sampling rate in communication systems. With wide bandwidth in mmWave spectrum, sampling rate can reach up to 10 Gbit/s level, and high-speed circuit design becomes very difficult. It is also reported that the energy efficiency for components (power amplifier, analog-to-digital converter, digital-to-analog converter) in high frequency is low, only around 10%. One of the major concerns from network operators is that power consumption will hike due to 5G. Furthermore, the low efficiency in these components also brings thermal issues in hand-held devices, degrading user experiences. Thirdly, with mmWave band, performance gain largely comes from large-scale antenna array, current design can integrate hundreds of antenna elements in a small area (due to small wavelength of mmWave signals). Even though this can facilitate the beamforming, which generates narrow but stronger signals toward desired direction, the overhead for channel estimation, precoding, and beam tracking is too large. Fourthly, in UDN networks, since the transmitter density is high, signals can cause higher interferences with each other. The problem will be more severe with high-density users in the same area. Challenges in mobility management, interference management, and heterogeneity nature of devices are severe. Lastly, it is expected to support intelligent applications in beyond 5G systems. For example, conventional communication systems are transparent of message (i.e. they are only responsible for transmitting bits but do not know any further info). Semantic communication, on the other hand, has knowledge of the underlying message, and the communication scheme can be dynamically changed to fit different needs of the message. Besides, ubiquitous wireless signals open door for sensing applications, such as localization, monitoring, and healthcare. In recent years, intelligent communication system has been proposed to accommodate these needs. A notable example is wireless federated learning system to cater the distributed machine learning. However, a deep integration from wireless design perspective is strongly desired.
Recently, there are several emerging technologies which aim to deliver the goal of 5G and beyond, and address the challenges above. Specifically, in this book, our focus is on the physical layer techniques, such as 5G NR non-orthogonal multiple access (NOMA) and physical layer (PHY) mobile edge computing (MEC), high-level communication architecture for pervasive Internet of Things (IoT) devices, as well as wireless federated learning system design. We have conducted preliminary researches to address the challenges mentioned above. Specifically, we discuss how to utilize NOMA on improving aggregated data rate and supporting more devices simultaneously, propose schemes for wearable IoT communications, discuss the usage of MEC on helping with power consumption and latency, and analyze how wireless design can facilitate distributed machine learning. Below we briefly introduce each enabling technique.
1.2 Enabling Technologies
1.2.1 5G New Radio
1.2.1.1 Non-orthogonal Multiple Access (NOMA)
Initially proposed by NTT DOCOMO as an enhancement for LTE-advanced (LTE-A) in 2013, NOMA has been recognized as one of the most promising techniques for 5G due to its capability of supporting a higher spectral efficiency (SE) and native integration of massive connectivity. The basic principle of NOMA is that at the transmitter side, multiple signals are added up with different powers, forming a superimposed signal (SS). To ensure weak user's quality of service (QoS), at the receiver side, successive interference cancellation (SIC) is used to retrieve each user's signal sequentially from the SS. Specifically, a user can decode the strongest signal by treating other signals as interference. If the decoded signal is its own data, SIC stops. Otherwise, the receiver subtracts the decoded signal from the SS and continues to decode the next strongest signal. Notice that SS with SIC is not new; in information theory, this duo is a capacity-achieving technique in the uplink communication. However, the difference is in NOMA, the weak user has a stronger power, which is not information-theoretic optimal. Since its design philosophy may be combined with diverse transceivers, it has drawn tremendous attention in multiple-antenna systems and in downlink and uplink multi-cell networks. In contrast to classic orthogonal multiple access...
