Massive Connectivity
Description
Learn to support more devices and sensors in Internet of Things applications through NOMA and machine-type communication
Non-orthogonal multiple access (NOMA) has held much interest due to its ability to provide a higher spectral efficiency--such as more bits per unit bandwidth in Hertz--than other, orthogonal multiple access schemes. The majority of this research focuses on the application of NOMA to downlink channels (from base station to users) in cellular systems as its use for uplink (users to base station) is somewhat circumscribed. However, NOMA has recently been employed in contention-based uplink access, which has shown an improvement in performance that allows an increase in the number of users that can be supported. As a result, NOMA is promising for machine-type communication (MTC) in 5G systems and beyond, making it a key enabler of the Internet of Things (IoT).
Massive Connectivity provides an in-depth, comprehensive view of the benefits and drawbacks of uplink NOMA random access (RA) systems. This text offers a basic introduction and description of uplink NOMA RA systems before considering the possibilities for evolution of the scheme as attempts are made to derive the most benefits and overcome any weaknesses. The book further presents key performance analysis while also highlighting game-theoretic views. In essence, by describing the essential properties of stable and high-throughput yielding RA systems, the book demonstrates that uplink NOMA can fulfill these required properties.
Massive Connectivity readers will also find:
* An extensive literature survey on RA systems and their applications since the 1970s
* Recent advances in random access for massive connectivity
* Retransmission control algorithms for NOMA random access
* Discussion of how uplink NOMA random access systems can be integrated into the existing long-term evolution (LTE) or upcoming 5G cellular networks
Massive Connectivity is a useful reference for field engineers and academics, as well as experts for random access systems for IoT applications.
More details
Other editions
Additional editions
Person
Jinho Choi, PhD, is a Professor with the School of Information Technology at Deakin University, Australia.
Content
Preface xiii
1 Introduction 1
1.1 Machine-Type Communication 1
1.2 Non-Orthogonal Multiple Access 3
1.3 NOMA for MTC 4
1.4 An Overview of Probability and Random Processes 6
1.4.1 Review of Probability 6
1.4.2 Random Variables 7
1.4.3 Random Processes 14
1.4.4 Markov Chains 15
2 Single-User and Multiuser Systems 19
2.1 A Single-User System 19
2.1.1 Signal Representation 20
2.1.2 Transmission of Signal Sequences 21
2.1.3 ML Decoding 23
2.1.4 ML Decoding over Fading Channels 26
2.1.5 Achievable Rate 28
2.2 Multiuser Systems 33
2.2.1 Broadcast Channels 34
2.2.2 Multiple Access Channels 37
2.3 Further Reading 41
3 OMA and NOMA 43
3.1 Orthogonal Multiple Access 43
3.1.1 Time Division Multiple Access 43
3.1.2 Frequency Division Multiple Access 46
3.1.3 Orthogonal Frequency Division Multiple Access 47
3.2 Non-Orthogonal Multiple Access 51
3.2.1 Downlink NOMA 52
3.2.2 Uplink NOMA 57
3.3 Power and Rate Allocation 60
3.3.1 System with Known Instantaneous CSI 60
3.3.2 System with Unknown Instantaneous CSI 67
3.4 Code Division Multiple Access 73
3.4.1 DS-CDMA 74
3.4.2 Multiuser Detection Approaches 78
3.5 Further Reading 84
4 Random Access Systems 87
4.1 ALOHA Systems 88
4.1.1 Single Channel Random Access 88
4.1.2 Multi-Channel S-ALOHA 90
4.2 Throughput Analysis 91
4.2.1 Pure ALOHA 91
4.2.2 Slotted ALOHA 92
4.2.3 Multichannel ALOHA 94
4.3 Analysis with a Finite Number of Users 98
4.3.1 A Markov Chain 98
4.3.2 Drift Analysis 100
4.4 Analysis with an In_nite Number of Users 102
4.4.1 Constant Re-transmission Probability 102
4.4.2 Adaptive Re-transmission Probability 104
4.5 Fast Retrial 107
4.6 Multiuser Detection 108
4.6.1 Compressive Random Access 108
4.6.2 Throughput Analysis 110
4.7 Further Reading 114
5 NOMA-based Random Access 117
5.1 NOMA to Random Access 117
5.1.1 S-ALOHA with NOMA 118
5.1.2 More Power Levels 122
5.2 Multichannel ALOHA with NOMA 127
5.2.1 Multichannel ALOHA with NOMA and Throughput Analysis 128
5.2.2 Channel-Dependent Selection 132
5.3 Opportunistic NOMA 137
5.3.1 System Model 137
5.3.2 Throughput Analysis 140
5.3.3 Opportunistic NOMA for Channel Selection 147
5.4 NOMA-based Random Access with Multiuser Detection 152
5.4.1 Compressive Random Access 152
5.4.2 Layered CRA 154
5.4.3 Performance under Realistic Conditions 159
5.5 Further Reading 161
6 Application of NOMA to MTC in 5G 163
6.1 Machine-Type Communication 163
6.1.1 IoT Connectivity 163
6.1.2 Random Access Schemes for MTC 164
6.2 A Model with Massive MIMO 168
6.2.1 Massive MIMO 168
6.2.2 Two-step Random Access with Massive MIMO 173
6.2.3 Throughput Analysis 174
6.3 NOMA for High-Throughput MTC 177
6.3.1 Co-existing Preambles and Data Packets 178
6.3.2 Maximum Throughput Comparison 180
6.3.3 Limitations 184
6.4 Layered Preambles for Heterogeneous Devices 185
6.4.1 Heterogeneous Devices in MTC 185
6.4.2 Design of Layered Preambles 187
6.4.3 Performance Analysis 189
6.5 Further Reading 195
7 Game-Theoretic Perspective of NOMA-based Random Access 197
7.1 Background of Game Theory 197
7.1.1 Normal-Form Games 198
7.1.2 Nash Equilibrium 200
7.1.3 Mixed Strategies 200
7.2 Random Access Game 202
7.2.1 Normal-Form and NE 203
7.2.2 Mixed Strategies 204
7.3 NOMA-ALOHA Game 204
7.3.1 Single-Channel NOMA-ALOHA Game 205
7.3.2 Multichannel NOMA-ALOHA Game 216
7.4 Fictitious Play 221
7.4.1 A Model for Fictitious Play 221
7.4.2 Convergence 223
7.5 Evolutionary Game Theory and Its Application 227
7.5.1 Population Games 227
7.5.2 Replicator Dynamics and Evolutionary Stable State 228
7.5.3 Stability of the Replicator Dynamics 231
7.5.4 Application to NOMA 232
7.6 Further Reading 234
Index 247
1
Introduction
Two main topics are covered in this book. One is machine-type communication (MTC) and the other is non-orthogonal multiple access (NOMA). Each topic has its own foundations and applications. In this chapter, we briefly explain each of them, and then explain why both topics should be covered in this book.
1.1 Machine-Type Communication
It may not be easy to imagine our daily life and business without the Internet although it began to appear as a backbone network in the 1970s to interconnect a small number of academic and military networks. The Internet is a network of networks and allows to exchange information between servers, computers, mobile phones, and so on restlessly. The Internet-of-Things (IoTs) is a natural extension of the Internet as machines, devices, and sensors are connected to the Internet to exchange information without human intervention in a number of applications such as smart factory.
As the number of devices connected to the Internet grows, their connectivity becomes important. Private and public networks can be used for their connectivity. For example, for smart home applications, a private network can be used at home to allow a small number of devices to be connected. For smart city applications, a large-scale public network would be preferable. Thus, the deployment of IoT networks depend on applications.
MTC is to support communications between machines or devices without human intervention. Unlike human-type communication (HTC), MTC mainly focuses on uplink transmission rather than downlink transmission (this is one of the main differences between MTC and HTC, where it can be seen that MTC's design principles must be different from those of HTC) and will support sporadic traffic in the form of short data packets. As a result, in order to keep signaling overhead low, the random access channel (RACH) procedure is used for MTC in Long-Term Evolution (LTE) systems. In the fifth generation (5G) system, a new random access scheme consisting of two steps, which is more efficient than the RACH procedure in LTE consisting of four steps, has been standardized.
MTC can provide connectivity for a large number of devices in a cell, paving the way for IoT applications to interact with devices deployed over a large area via cellular systems. This means that MTC becomes essential in various IoT applications such as smart cities and intelligent transport systems.
Furthermore, the global number of connected devices is expected to exceed 500 billion by 2030, while the human population is predicted to be 8.5 billion by the United Nations (UN). This means that IoT devices will outnumber human population by approximately 60-folds in 2030, and these devices will be used in a variety of IoT applications requiring heterogeneous connectivity demand. As a result, MTC will play a more prominent role in 5G and beyond (i.e. the sixth generation (6G)) and thus new MTC schemes need to be developed to meet the diverse requirements for future IoT applications.
1.2 Non-orthogonal Multiple Access
Various multiple access schemes have been used to support multiple users in a cellular system. For example, time division multiple access (TDMA) is adopted in the global system for mobile communications (GSM), which is regarded as a second generation (2G) system. In the third generation (3G) to 5G systems, orthogonal frequency division multiple access (OFDMA) is used. In general, most multiple access schemes used in cellular systems are orthogonal multiple access (OMA) schemes that allocate orthogonal channel resources to different users. To increase the spectral efficiency, NOMA schemes have been considered, where multiple users share the same channel resource.
NOMA has been extensively studied for cellular systems since the 2010s. In particular, for downlink transmissions, various NOMA schemes are studied using the difference in propagation loss between users near the center of the cell (where a base station (BS) is located) and users far from the center. The resulting NOMA is often referred to as power-domain NOMA.
It is necessary to distinguish between power-domain NOMA and NOMA in a broad sense. For example, code-division multiple access (CDMA) and interleave-division multiple access (IDMA) can be seen as NOMA schemes, because multiple users' signals can co-exist in a shared radio resource block, where one user signal can see the other users' signals as interfering signals. In CDMA, each user's signal is spread by a dedicated sequence, which is called the spreading sequence. Due to spreading sequences, CDMA has a bandwidth expansion. In particular, the bandwidth of CDMA increases by a factor of the processing gain or the length of the spreading sequence, while IDMA is a generalization of CDMA with forward error correcting codes. On the other hand, power-domain NOMA does not use spreading sequences. As a result, there is no bandwidth expansion and a high spectral efficiency can be achieved.
In power-domain NOMA, however, the transmit power levels and transmission rates should be carefully decided so that successive interference cancellation (SIC) can be used to remove other users' signals once they are decoded.
1.3 NOMA for MTC
In general, power-domain NOMA for downlink requires coordinated transmissions by a BS in terms of transmit powers and rates. Thus, it seems difficult to use power-domain NOMA for uplink as coordinated transmission by distributed users is not easy to implement. In other words, the gain of NOMA can be offset by excessive signaling overhead to perform coordinated transmissions by distributed users. As a result, the use of NOMA for random access seems quite challenging. On the contrary, NOMA is well-suited to random access as we will illustrate with two users.
Suppose that two users want to access a given channel without coordination. If two users always transmit their signals, they experience collisions and no user succeeds to transmit. Thus, they need to transmit with a certain probability. To this end, each user is to decide the access probability, denoted by , , for user . The probability that at least one user succeeds to transmit a packet is given by
If , , which is maximized when and the maximum of is , which is also the maximum average number of successfully transmitted packets. To consider random access with NOMA, we can assume two different power levels, and with , and the receiver is able to decode both the signals if one user transmits a signal with transmit power and the other with . Let be the probability that a user chooses the high transmit power when transmitting (with probability ). Then, the average number of successfully transmitted packets is given as follows:
where the first term is the average number of successfully transmitted packets when only one user transmits and the second term is the average number of successfully transmitted packets when two users transmit simultaneously. It is easy to show that maximizes . Then, we have . Thus, maximizes , which is 1. In other words, the maximum average number of successfully transmitted packets can be doubled if NOMA is used for uncoordinated transmissions of two users in uplink transmissions.
As more power levels are considered, the average number of successfully transmitted packets can increase. However, this comes at the cost of high transmit power by devices.
In this book, we mainly focus on the principles of NOMA and the application of NOMA to MTC. In particular, we discuss how NOMA can help improve the performance of random access in MTC once we present key ideas of random access including its stability. Game theory will also be used to understand the nature of random access where users compete for common radio resources in MTC.
1.4 An Overview of Probability and Random Processes
Prior to the main parts of this book, we present an overview of probability and random processes in this section, which can be used to see the required background in terms of probability and random processes. The reader is referred to well-known textbooks such as Papoulis and Pillai (2002); Ross (1995); Mitzenmacher and Upfal (2005) if not well equipped with theory of probability and random processes.
1.4.1 Review of Probability
A sample space is the set of all possible outcomes (or events) of an experiment. Let be an event , which is a subset of . A probability measure is a mapping from to the real line with the following properties:
- ,
- For a countable set of events, , if , for , then
The joint probability of two events and is expressed as , where the conditional probability of given is expressed as
The two events and are independent if and only if
and this implies .
In addition, for any two events and , we have
where the equality holds if . Thus, for a set of events, it can be shown that
(1.1) ...