79,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Wissenschaft und neue Technologien
- Serie: Wiley - IEEE
- Sprache: Englisch
5G and Beyond Wireless Communication Networks
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.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 319
Veröffentlichungsjahr: 2023
Ähnliche
Table of Contents
Cover
Table of Contents
Title Page
Copyright
About the Authors
Preface
Acknowledgments
1 Introduction to 5G and Beyond Network
1.1 5G and Beyond System Requirements
1.2 Enabling Technologies
1.3 Book Outline
Note
2 5G Wireless Networks with Underlaid D2D Communications
2.1 Background
2.2 NOMA‐Aided Network with Underlaid D2D
2.3 NOMA with SIC and Problem Formation
2.4 Precoding and User Grouping Algorithm
2.5 Numerical Results
2.6 Summary
3 5G NOMA‐Enabled Wireless Networks
3.1 Background
3.2 Error Propagation in NOMA
3.3 SIC and Problem Formulation
3.4 Precoding and Power Allocation
3.5 Numerical Results
3.6 Summary
4 NOMA in Relay and IoT for 5G Wireless Networks
4.1 Outage Probability Study in a NOMA Relay System
4.2 NOMA in a mmWave‐Based IoT Wireless System with SWIPT
5 Robust Beamforming in NOMA Cognitive Radio Networks: Bounded CSI
5.1 Background
5.2 System and Energy Harvesting Models
5.3 Power Minimization‐Based Problem Formulation
5.4 Maximum Harvested Energy Problem Formulation
5.5 Numerical Results
5.6 Summary
6 Robust Beamforming in NOMA Cognitive Radio Networks: Gaussian CSI
6.1 Gaussian CSI Error Model
6.2 Power Minimization‐Based Problem Formulation
6.3 Maximum Harvested Energy Problem Formulation
6.4 Numerical Results
6.5 Summary
7 Mobile Edge Computing in 5G Wireless Networks
7.1 Background
7.2 System Model
7.3 Problem Formulation
7.4 Numerical Results
7.5 Summary
8 Toward Green MEC Offloading with Security Enhancement
8.1 Background
8.2 System Model
8.3 Computation Efficiency Maximization with Active Eavesdropper
8.4 Numerical Results
8.5 Summary
Note
9 Wireless Systems for Distributed Machine Learning
9.1 Background
9.2 System Model
9.3 FL Model Update with Adaptive NOMA Transmission
9.4 Scheduling and Power Optimization
9.5 Scheduling Algorithm and Power Allocation
9.6 Numerical Results
9.7 Summary
Note
10 Secure Spectrum Sharing with Machine Learning: An Overview
10.1 Background
10.2 ML‐Based Methodologies for SS
10.3 Summary
11 Secure Spectrum Sharing with Machine Learning: Methodologies
11.1 Security Concerns in SS
11.2 ML‐Assisted Secure SS
11.3 Summary
12 Open Issues and Future Directions for 5G and Beyond Wireless Networks
12.1 Joint Communication and Sensing
12.2 Space‐Air‐Ground Communication
12.3 Semantic Communication
12.4 Data‐Driven Communication System Design
Appendix A: Proof of Theorem 5.1
Bibliography
Index
End User License Agreement
List of Tables
Chapter 5
Table 5.1 Simulation parameters.
Chapter 6
Table 6.1 Simulation parameters
Chapter 9
Table 9.1 Summary of notations.
Table 9.2 Statistics of datasets.
List of Illustrations
Chapter 1
Figure 1.1 Four main goals for 5G.
Figure 1.2 NOMA principles: transmission and decoding stage.
Figure 1.3 Paradigm shift from cloud computing to mobile edge computing.
Figure 1.4 Wearable devices may have varying forms, from small medical senso...
Figure 1.5 A promising network architecture for pervasive IoT communication ...
Chapter 2
Figure 2.1 System model.
Figure 2.2 System capacity of two proposed ZF precoding methods vs. TDMA as ...
Figure 2.3 CUs capacity of two proposed ZF precoding methods vs. TDMA as the...
Chapter 3
Figure 3.1 UE rate with different precoding matrix as increases ().
Figure 3.2 Sum rate with different precoding matrix as increases ().
Figure 3.3 UE rate with different precoding matrix as increases ().
Figure 3.4 Sum rate with different precoding matrix as increases ().
Chapter 4
Figure 4.1 NOMA cooperative scheme.
Figure 4.2 NOMA TDMA scheme.
Figure 4.3 Theorem 4.1 and 4.2. bps/Hz.
Figure 4.4 Theorem 4.3. bps/Hz. and .
Figure 4.5 Theorem 4.4. bps/Hz. and .
Figure 4.6 System model.
Figure 4.7 Power‐in‐power‐out response in the non‐linear energy harvest mode...
Figure 4.8 Outage performance for both UEs with comparison to analytical res...
Figure 4.9 Outage performance for UE 2 as the function of .
Chapter 5
Figure 5.1 (a) An illustration of the system model. (b) The power splitting ...
Figure 5.2 The empirical CDF of the minimum transmit power of the CBS under ...
Figure 5.3 The minimum transmit power of the CBS vs. the required SNR of SUs...
Figure 5.4 Impact of the number of CBS antennas on the minimum transmitted p...
Figure 5.5 Impact of channel uncertainties and on the overall minimum tr...
Figure 5.6 Average maximum EH power under different interferences tolerated ...
Figure 5.7 Average maximum EH power vs. the minimum SNR required by the SUs,...
Figure 5.8 Average total EH power vs. the number of SUs for dBm, bit/s/H...
Chapter 6
Figure 6.1 The empirical CDF of the minimum transmit power of the CBS under ...
Figure 6.2 The minimum transmit power of the CBS vs. the required SNR of SUs...
Figure 6.3 Impact of the number of CBS antennas on the minimum transmitted p...
Figure 6.4 Impact of channel uncertainties and on the overall minimum tr...
Figure 6.5 Average maximum EH power under different interferences tolerated ...
Figure 6.6 Average maximum EH power vs. the minimum SNR required by the SUs,...
Figure 6.7 Average total EH power vs. the number of SUs for dBm, bit/s/H...
Chapter 7
Figure 7.1 Performance comparison of different schemes.
Figure 7.2 Performance comparison of our proposed scheme and the binary offl...
Figure 7.3 Trade‐off between offloading and local computing.
Chapter 8
Figure 8.1 Secure MEC partial offloading model.
Figure 8.2 Time sharing offloading scheduling.
Figure 8.3 Iterative algorithm convergence analysis.
Figure 8.4 Computation efficiency vs. required computation bits under differ...
Figure 8.5 Computation efficiency vs. required computation bits under differ...
Chapter 9
Figure 9.1 An illustration of the proposed scheme. (a) A general FL model up...
Figure 9.2 Scheduling diagram.
Figure 9.3 A scheduling graph example.
Figure 9.4 Test accuracy comparison under different scenarios, when , , an...
Figure 9.5 Test accuracy comparison between original TDMA‐based FedAvg and N...
Figure 9.6 Test accuracy on FEMNIST datasets: Test accuracy comparison vs co...
Figure 9.7 Test accuracy on
Sent140
datasets: Test accuracy comparison vs co...
Chapter 10
Figure 10.1 Spectrum sharing paradigm.
Figure 10.2 ML‐based methodologies for SS.
Figure 10.3 Key steps in CRN.
Figure 10.4 The reinforcement learning cycle.
Figure 10.5 LBT‐based method.
Figure 10.6 Duty cycle‐based method.
Figure 10.7 The coexistence of LTE‐U and WiFi in an unlicensed spectrum.
Figure 10.8 AmBC network.
Chapter 11
Figure 11.1 Secure issues in SS network.
Figure 11.2 ML‐assisted secure SS.
Figure 11.3 Illustration of PUE attacks.
Figure 11.4 Illustration of SSDF attack.
Figure 11.5 System model for attacker's learning.
Figure 11.6 Anti‐jamming attack in AmBC‐CRN.
Guide
Cover
Table of Contents
Title Page
Copyright
About the Authors
Preface
Acknowledgments
Begin Reading
Bibliography
Index
Wiley End User License Agreement
Pages
iii
iv
xi
xiii
xiv
xv
1
2
3
4
5
6
7
8
9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
69
70
71
72
73
74
75
76
77
78
79
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
155
156
157
158
159
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
181
182
183
184
5G and Beyond Wireless Communication Networks
Haijian Sun
University of GeorgiaAthens, GA, USA
Rose Qingyang Hu
Utah State UniversityLogan, UT, USA
Yi Qian
University of Nebraska‐LincolnOmaha, NE, USA
This edition first published 2024
© 2024 by John Wiley & Sons Ltd
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions.
The right of Haijian Sun, Rose Qingyang Hu, and Yi Qian to be identified as the authors of this work has been asserted in accordance with law.
Registered Offices
John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats.
Trademarks: Wiley and the Wiley logo are trademarks or registered trademarks of John Wiley & Sons, Inc. and/or its affiliates in the United States and other countries and may not be used without written permission. All other trademarks are the property of their respective owners. John Wiley & Sons, Inc. is not associated with any product or vendor mentioned in this book.
Limit of Liability/Disclaimer of Warranty
While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.
Library of Congress Cataloging‐in‐Publication Data applied for:
Hardback ISBN: 9781119089452
Cover Design: Wiley
Cover Image: © timandtim/Getty Images
About the Authors
Haijian Sun, PhD, is an Assistant Professor in the School of Electrical and Computer Engineering at the University of Georgia, Athens, GA, USA.
Rose Qingyang Hu, PhD, is an Associate Dean for Research in the College of Engineering and a Professor in the Department of Electrical and Computer Engineering at Utah State University, Logan, UT, USA. She is a Fellow of IEEE.
Yi Qian, PhD, is an IEEE Fellow and is a Professor in the Department of Electrical and Computer Engineering at the University of Nebraska‐Lincoln, Lincoln, NE, USA. He is a Fellow of IEEE.
Preface
Motivated by the increasing demands for connectivity, modern wireless technologies have experienced rapid developments. Efforts from academia, industry, and government have pushed wireless research at an unprecedented level. As a complex solution, wireless systems comprise many components, from physical layer, to network and upper application layer. One of the most exciting innovations in the past decade is 5G physical layer, which includes new radio (NR), new spectrum, coding, etc. Inspired by recent physical layer research advances in 5G and beyond wireless systems, this book intends to present the state‐of‐art challenges and solutions for physical layer techniques that are already applied, or will be utilized in wireless systems. This book covers a variety of topics, primarily on the intersection of 5G and beyond system with NR, mobile edge computing, and machine learning, and spectrum sharing. Ultimately, we expect to deliver a more energy‐, spectral‐, and computation‐efficient wireless technology.
There are twelve chapters in this book. They can be categorized into three main topics. Chapters 1–6 focus on 5G new radio research, especially recent advancements and systematic research on non‐orthogonal multiple access (NOMA). Chapters 7–9 discuss the interactions of mobile edge computing and wireless technology. Chapters 10 and 11 focus on secure spectrum sharing in 5G and beyond era. Chapter 12 concludes this book and further discusses some future research directions. Below, we briefly summarize each chapter.
Chapter 1 presents an overview of 5G and beyond wireless system. We start by introducing system requirements and their technical challenges. Then the enabling technologies from NR, mobile edge computing, and heterogeneous communication architecture are illustrated.
Chapter 2 discusses the integration of 5G networks with device‐to‐device (D2D) communication. Specifically, the 5G system with underlaid D2D is presented. We show that such system can increase spectral efficiency, providing that resource allocation is properly designed.
Chapter 3 deals with NOMA‐enabled practical wireless networks. The highlight is the integration of error propagation, a well‐known issue in NOMA. It shows that error propagation can degrade system performance, depending on the residual value.
Chapter 4 presents 5G relay and IoT networks with NOMA. In the first part, we derive the outage probability in the relay system and show the potential of such a configuration. Then, in the second part, the IoT network with power transfer capability is considered.
Chapter 5 discusses the robust beamforming problem in cognitive radio system; we specifically illustrate the beamforming design when bounded channel estimation error is present.
Chapter 6 is a continuation of Chapter 5. It considers a more realistic channel estimation model, in which channel estimation error is modeled as Gaussian variable. Correspondingly, beamforming design also changes.
Chapter 7 presents mobile edge computing in 5G wireless networks. The system aims at reducing computing latency and offload computation tasks to nearby edge servers. With the goal of maximize computation efficiency, resource allocation optimization is proposed and designed.
Chapter 8 further considers security enhancements in mobile edge computing. Our security design focuses on physical layer, i.e. from wire‐tap channel perspective.
Chapter 9 deals with an innovative wireless system to facilitate distributed machine learning as opposed to machine learning for wireless communication. We show that efficient information exchange via wireless can accelerate large‐scale distributed machine learning. A direct application is wireless federated learning.
Chapter 10 provides an overview for secure spectrum sharing with machine learning. While secure spectrum sharing is not a new topic, we have witnessed advancements in this area, especially with machine learning techniques.
Chapter 11 presents detailed machine learning methodologies for secure machine learning. This chapter illustrates several dominant attacks and their respective mitigation approaches.
Chapter 12 concludes this book and gives some emerging research directions in 5G and beyond wireless networks.
We hope our readers will enjoy this book.
January 2023
Haijian Sun
, University of Georgia
Rose Qingyang Hu
, Utah State University
Yi Qian
, University of Nebraska‐Lincoln
Acknowledgments
We would like to thank our families for their continuous support and love.
We would like to express our sincere gratitude and appreciation to our colleagues, students, and staff at University of Georgia, Utah State University, and University of Nebraska‐Lincoln who have supported us throughout the journey of writing and publishing this book. Your encouragement, feedback, and advice have been invaluable in shaping the final product, and we are truly grateful for your contributions.
We also would like to extend our heartfelt gratitude to Juliet Booker, Sandra Grayson, and Becky Cowan at Wiley who have played an integral role in bringing our new book to fruition. Thank you for your hard work, support, and guidance throughout the publishing process.
This book project was partially supported by the U.S. National Science Foundation under grants CNS‐2236449, CNS‐2007995, CNS‐2008145, ECCS‐2139508, and ECCS‐2139520.
Haijian Sun, Rose Qingyang Hu, and Yi Qian
1Introduction 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 (OMA), NOMA provides simultaneous access to multiple users at the same time and on the same frequency band, for example by using power‐domain multiplexing. It has been shown that NOMA is capable of achieving a higher SE and energy efficiency (EE) than OMA, such as OFDMA, time division multiple access (TDMA), and frequency domain multiple access (FDMA). Figure 1.2 shows the basic principle of NOMA with data encoding and decoding. and are the symbols for users 1 and 2, respectively. We also assume user 1 has a better channel than user 2. At the transmitter side, binary phase shift keying (BPSK) and quadratic phase shift keying (QPSK) modulation are applied, respectively, for the two users. Clearly, the average symbol power of is larger to compensate for the unfavorable channel. Actual transmitted symbol is simply the addition of these two. At the receiver side, symbols with the highest power will be decoded first, in this example, . Besides, since the received symbol is on the right side of y‐axis, for BPSK, it will be decoded as , and then removed from the composite signal, which only has left. Notice that the symbols can use the same modulation scheme as long as they have different power. Most NOMA works, however, do not consider any specific modulation, rather they apply the Gaussian coding and perform analysis based on information‐theoretic perspective.
Figure 1.2 NOMA principles: transmission and decoding stage.
The disadvantage of NOMA, however, lies in the following aspects. Firstly, NOMA requires a more complicated receiver structure to perform SIC; hence, the cost will be higher and receiver architecture will also be changed accordingly. Secondly, during SIC procedure, one user will decode signal from others; this will cause security and privacy concerns. Lastly, depending on implementation, this successive decoding will impose certain delays for users.
Starting from 3rd Generation Partnership Project (3GPP) LTE Release‐13, NOMA, as one of the techniques in multi‐user superposed transmission (MUST), has become part of the standardization. In 2017, with LTE Release‐14, 15 MUST schemes have been proposed for the uplink NR. Additionally, NOMA has attracted extensive attention from industry. NTT DoCoMo and MediaTek collaborated to have a field test of NOMA in Nov. 2017. With a simple scenario of one base station and three users, they were able to achieve 2.3 time spectral efficiency compared with current technology.1
Nevertheless, we have applied NOMA in many schemes and systematically studied its performance, for example NOMA with D2D, with MIMO, relay networks, and cognitive radio. More importantly, we have reviewed the fundamental principle of NOMA and pointed out the error propagation phenomenon. Furthermore, we have also considered the channel imperfection and its impact to NOMA performance.
1.2.1.2 Channel Codes
Channel coding is instrumental for achieving higher capacity and reliability. For example, low‐density parity‐check (LDPC) has been extensively used in 4G, replacing convolutional and turbo codes in previous generations. In 5G NR, polar codes are identified as another promising capacity‐achieving coding technique. Polar codes have been adopted in 5G standardization process. For example, 3GPP incorporates polar codes for both uplink and downlink control information communication service, such as enhanced mobile broadband (eMBB), massive machine type communications (mMTC), and ultra‐reliable and low latency communications (URLLC). Channel codes for 5G NR should be flexible to support the variable rate and length for both data and control packets. To address that, LDPC has developed several variations, such as quasi‐cyclic (QC) LDPC codes for better rate matching and interleaving, as well as parallelism for efficient encoding and decoding [59]; Multi‐edge (ME) LDPC mainly for throughput improvement and can scale well in larger block lengths. On the other hand, newly introduced polar code takes advantage of channel polarization, a natural behavior due to signal propagation. Correspondingly, encoding is recursively performed by the channel transformation matrix and creates channels that are either perfectly noiseless or completely noisy. A detailed tutorial of polar codes can be found in [16].
1.2.1.3 Massive MIMO
Massive MIMO refers to applying large‐scale antenna elements at transmitter and/or receiver side, usually the number of antenna is hundreds or more. MIMO can exploit spatial diversity or multiplexing, and improve system reliability (for example, lower bit error rate) and throughput, respectively. Compared with legacy MIMO system, massive MIMO brings significant improvements in diversity and multiplexing to fully exploit wireless channel characteristics. One prominent aspect is massive MIMO can generate very narrow beams toward the receiver side. Hence, it can not only increase reception power, but also benefit network capacity and coverage, and ultimately provide better user experience.
These benefits come at a price. Like MIMO, performance gain from massive MIMO largely comes from beamforming and advanced signal processing techniques, which require channel information. If both transmitter and receiver have massive MIMO antennas, their channel is a matrix with hundreds by hundreds of elements. Overhead for accurate channel estimation is prohibitively large. For example, orthogonal pilots are usually applied to obtain channel information; in the case of massive MIMO, maintaining pilot orthogonality is difficult, not to mention practical challenges such as pilot contamination and offset (time and frequency). To address these challenges, prior works have studied robust beamforming design, such that the requirement for accurate channel information can be relaxed. Furthermore, signal processing in massive MIMO is also sophisticated. Traditional optimization methods for throughput maximization or bit error rate (BER) minimization become problematic due to high computation complexity, which hinders the deployment in mobile devices.
It is worth to note that other approaches such as applying out‐of‐band information, including vision, location, and geometry data to assist beamforming are also studied. Out‐of‐band information provides complementary details for assisting beamforming steering. These emerging solutions are primarily motivated and enabled by machine learning.
1.2.1.4 Other 5G NR Techniques
5G NR also has other innovations. Recent 3GPP releases 15, 16, and 17 gradually bring more flexibility and enhancement on several aspects. For example, dynamic slot structure caters to different communication needs, for either low‐latency or high data‐rate application. This structure allows for customized slot design, for examples, adding a longer or shorter cyclic prefix, changing the data frame length, or providing extra guard space. Another innovation is spectrum sharing. In contrast to static database‐aided spectrum sharing, which detects secondary users' interference and only allows them to access bands in an opportunistic way, current spectrum sharing is more dynamic, enabled by advanced machine learning‐based approach, hence is more efficient and accurate.
1.2.2 Mobile Edge Computing (MEC)
Due to the size, battery, and cost limitations, mobile devices can experience performance bottleneck when computation‐intensive tasks are added. More than one decade ago, people solved this problem by introducing the concept of cloud computing. Mobile devices do not perform large‐scale computation locally; instead, they send these tasks to remote servers for faster and more secure processing, storage, and sharing. The centralized nature of cloud‐based computing can reduce the expenditure cost while providing easier deployment process. However, cloud servers may be located in remote areas, which causes inevitably longer end‐to‐end transmission and processing delay.
MEC is a new alternative paradigm for the upcoming 5G systems. Instead of transmitting data to the remote servers for processing, MEC provides certain computation capacities locally, for example within the base station in the wireless cellular networks. This paradigm shift can effectively reduce long backhaul latency and energy consumption, as well as support a more flexible infrastructure in a cost‐effective way. MEC has attracted extensive research interests recently, not only in the architectural level, but also in specific tasks such as cooperative computation offloading. Computation offloading, which leverages the powerful MEC servers in proximity and sends the computation‐intensive tasks for further processing, is a desirable scheme to overcome the physical limitations of user devices (Figure 1.3).
Figure 1.3 Paradigm shift from cloud computing to mobile edge computing.
We see this paradigm shift in a more fundamental way. In cloud computing era, even though the data transmission speed is not high, the bottleneck comes mainly from the computation capacity. With Moore's law still being effective, performance of integrated circuit chips grows exponentially. On the other hand, communication technology makes the speed increase almost linearly. Since the goal is to reduce processing speed, it is more beneficial to perform task execution both locally and remotely.
In order to reduce latency as well as to improve system efficiency, we propose a joint processing scheme in which the total task can be divided into two parts, one for local computing and the other for offloading. To cope with the ever‐increasing concerns on energy efficiency, we evaluate the system performance by a new metric, computational efficiency (CE). It is defined as the total number of bits computed with the total energy consumption. The objective is to maximize each user's CE with time constraints (users should finish their task before a required time), energy constraint (each user is powered by battery; hence, the total energy should be below a threshold), and task constraint (each user should finish a minimum number of data bits). Later we show CE is a more appropriate method in terms of finding the balance of more tasks and less energy.
1.2.3 Hybrid and Heterogeneous Communication Architecture for Pervasive IoTs
Recent years have witnessed the unprecedented growth of wearable devices owing to the swift advances in chip design, computing, sensing, and communications technologies. While wearable devices are not new, the past few years have seen a surge in their large‐scale use and popularity. A wearable device or simply a wearable refers to a device that can be worn on the body. This rapid rise in popularity was spurred, in part, by technological innovation. Emerging system on chip (SoC) and system in package (SiP) have scaled down the printed circuit board (PCB) size, decreased power consumption, and most importantly, have made it possible to design wearables in a variety of desired shapes (Figure 1.4). Wearable devices provide easier access to information and convenience for their users. They have varying form factors, from low‐end health and fitness trackers to high‐end virtual reality (VR) devices, augmented reality (AR) helmets, and smart watches. These devices can collect data on heart rates, steps, locations, surrounding buildings, sleeping cycles, and even brain waves. Yet computing limitations continue to hinder wearables' ability to process data locally. As a result, most devices choose to offload their collected data to other powerful devices or to the clouds. This necessary communication plays a key role in wearable devices. Different applications provided by different wearables may have varying communication requirements. For example, while medical sensors have stringent requirements on latency and reliability, they have a low data rate need. On the other hand, AR/VR devices need both high throughput and low latency for a better user experience.
Figure 1.4 Wearable devices may have varying forms, from small medical sensors to entertainment helmets.
Wearable devices may not be able to take full advantage of current communication architecture, due to their potential cost and hardware complexity. On the other hand, wearable devices have succeeded in becoming more and more involved in everyday activities requiring voice, image, and video inputs. Human beings are generally sensitive to an approximate 100 audible delay and can catch visual delays of less than 10 . Furthermore, cell phones and tablets now use primarily touch interaction, a “tactile interaction” that requires a more rigorous delay control, such as 1 . A promising heterogeneous and hybrid network architecture is shown in Figure 1.5. It contains small BS (SBS), marco BS (MBS), remote radio head (RRH), and network slice.
Figure 1.5 A promising network architecture for pervasive IoT communication needs.
1.3 Book Outline
In face of several challenges by 5G and beyond system, this book focuses on technologies that can improve spectral, energy, and computation efficiency. As mentioned above, we mainly study physical layer techniques. Specifically, our first focus (Chapters 1–6) is to provide reader with latest research efforts on 5G NOMA. We have studied NOMA in a systematic way, including applying NOMA to address spectral efficiency and number of connected devices challenges under various network schemes. Our next focus (Chapters 7 and 8) is MEC. MEC is used to reduce computation delay, and we primarily investigate its role for computation offloading. Chapter 9 discusses the emerging wireless paradigm to facilitate distributed machine learning. Chapters 10 and 11 review secure spectrum sharing with machine learning techniques. Lastly, Chapter 12 concludes this book and discusses current and further research directions on 5G and beyond wireless systems.
Note
1
MediaTek Newsroom, Nov. 2017.
