Preface

5G wireless technology is developing at an explosive rate and is one of the biggest areas of research within academia and industry. In this rapid development, signal processing techniques are playing the most important role. In 2G, 3G and 4G, the peak service rate was the dominant metric for performance. Each of these previous generations was defined by a standout signal processing technology that represented the most important advance made. In 2G, this technology was time-division multiple access (TDMA); in 3G, it was code-division multiple access (CDMA); and in 4G, it was orthogonal frequency-division multiple access (OFDMA). However, this will not be the case for 5G systems - there will be no dominant performance metric that defines requirements for 5G technologies. Instead, a number of new signal processing techniques will be used to continuously increase peak service rates, and there will be a new emphasis on greatly increasing capacity, coverage, efficiency (power, spectrum, and other resources), flexibility, compatibility, reliability and convergence. In this way, 5G systems will be able to handle the explosion in demand arising from emerging applications such as big data, cloud services, and machine-to-machine communication.

A number of new signal processing techniques have been proposed for 5G systems and are being considered for international standards development and deployment. These new signal processing techniques for 5G can be categorized into four groups:

1. new modulation and coding schemes
2. new spatial processing techniques
3. new spectrum opportunities
4. new system-level enabling techniques.

The successful development and implementation of these technologies for 5G will be challenging and will require huge effort from industry, academia, standardization organizations and regulatory authorities.

From an algorithm and implementation perspective, this book aims to be the first single volume to provide a comprehensive and highly coherent treatment of all the signal processing techniques that enable 5G, covering system architecture, physical (PHY)-layer (down link and up link), protocols, air interface, cell acquisition, scheduling and rate adaption, access procedures, relaying and spectrum allocation. This book is organized into twenty-three chapters in five parts.

Part 1: Modulation, Coding and Waveform for 5G

The first part, consisting of eight chapters, will present and compare the detailed algorithms and implementations of all major candidate modulation and coding schemes for 5G, including generalized frequency division multiplexing (GFDM), filter-bank multi-carrier (FBMC) transmission, universal filtered multi-carrier (UFMC) transmission, bi-orthogonal frequency division multiplexing (BFDM), spectrally efficient frequency division multiplexing (SEFDM), the faster-than-Nyquist signaling (FTN) based time-frequency packing (TFP), sparse code multiple access (SCMA), multi-user shared access (MUSA) and non-orthogonal multiple access (NOMA).

With a focus on FBMC, GFDM, UFMC, BFDM and TFP, Chapter 1 presents a comprehensive introduction to these waveform generation and modulation schemes by covering the basic principles, mathematical models, step-by-step algorithms, implementation complexities, schematic processing flows and the corresponding application scenarios involved.

Chapter 2 is devoted to the FTN data transmission method, with the emphasis on applications that are important for future 5G systems. What is explored in this chapter mainly includes time-FTN methods with non-binary modulation and multi-subcarrier methods that are similar in structure to OFDM. In either, there is an acceleration processing in time or compacting in frequency that makes signal streams no longer orthogonal. FTN can be combined with error-correcting coding structures to form true waveform coding schemes that work at high-bit rates per Hertz and second. As a matter of fact, FTN based systems can potentially double data transmission rates.

The technical evolution from OFDM to FBMC is addressed in Chapter 3, covering the principles, algorithms, designs and implementations of these two schemes. This chapter first presents the details of OFDM-based schemes and the major shortcomings that prevent them from being employed in 5G. Through introduction of synthesis and analysis filter banks, prototype filter design and the corresponding polyphase implementation, Chapter 3 then extensively deals with the working principles of FBMC and compares it with OFDM in terms of performance - power spectral density and out of band power radiation - and complexity - number of fast Fourier transforms and filter banks. One can also see from this chapter that OFDM is a special case of FBMC.

Easy and effective integration with massive multiple-input and multiple-output (MIMO) technology is a key requirement for a modulation and waveform generation scheme in 5G. Chapter 4 demonstrates that FBMC can serve as a viable candidate waveform in the application of massive MIMO. The chapter outlines the system model, algorithm formulation, self-equalization property and pilot contamination of FBMC for massive MIMO channels, and also shows that while FBMC offers the same processing gain as OFDM, it offers the advantages of: more flexible carrier aggregation (CA), higher bandwidth efficiency - because of the absence of cyclic prefix (CP) - blind channel equalization and larger subcarrier spacing, and hence less sensitivity to carrier frequency offset and lower peak-to-average power ratio (PAPR).

Chapter 5 presents a non-orthogonal multicarrier system, namely, spectrally efficient frequency division multiplexing (SEFDM), which packs subcarriers at a frequency separation less than the symbol rate while maintaining the same transmission rate per individual subcarrier. Thus spectral efficiency is improved in comparison with the OFDM system. By transmitting the same amount of data, the SEFDM system can conceptually save up to 45% bandwidth. This chapter also describes a practical experiment in which the SEFDM concept is evaluated in a CA scenario considering a realistic fading channel. On the other hand, SEFDM involves higher computation complexity and longer processing delays, mainly due to the requirement for complex signal detection. This suggests that advanced hardware implementation is still highly desirable, so as to make SEFDM a better fit to 5G.

As pointed out in Chapter 6, non-orthogonal multi-user superposition and shared access is a promising technology that can increase the system throughput and simultaneously serve massive connections. Non-orthogonal access allows multiple users to share time and frequency resources in the same spatial layer via simple linear superposition or code-domain multiplexing. This chapter overviews all major non-orthogonal access schemes, categorizing them into two groups:

• the non-spreading methods, where modulation symbols are one-to-one mapped to the time/frequency resource elements
• the spreading methods, where symbols are first spread and then mapped to time/frequency resources.

Their design principles, key features, advantages and disadvantages are extensively discussed in this chapter.

Chapter 7 is devoted to a new multiple access scheme, termed NOMA, which introduces power-domain user multiplexing and exploits channel differences among users to improve spectrum efficiency. This chapter also explains the interface design aspects of NOMA, for example multi-user scheduling and multi-user power control, and its combination with MIMO. The performance evaluation and ongoing experimental trials of downlink and uplink NOMA are reported. The simulation results and the measurements obtained from the testbed show that under multiple configurations the cell throughput achieved by NOMA is 30% higher than that of OFDMA.

With a tutorial style, Chapter 8 presents an overview of all the major multicarrier modulation (MCM) candidates for 5G, categorizing them into three groups:

• subcarrier filtered MCM using linear convolution
• subcarrier filtered MCM using circular convolution
• subband windowed MCM.

General comparisons of these candidate algorithms are made in this chapter, covering PAPR, OOB emission, processing and implementation complexity, spectrum efficiency, the requirement of CP, intercarrier interference, intersymbol interference, multipath distortion, orthogonality and the related effects of frequency offset and phase noise, synchronization requirements in both the time domain and the frequency domain, latency, mobility, compatibility and integration with other processing such as massive MIMO.

Part 2: New Spatial Signal Processing for 5G

The five chapters in Part 2 focus on new spatial signal processing technologies for 5G, mainly addressing massive MIMO, full-dimensional MIMO (FD-MIMO), three-dimensional MIMO (3D-MIMO), adaptive 3D beamforming and diversity, continuous aperture phased MIMO (CAP-MIMO) and orbital angular momentum (OAM) based multiplexing. Chapter 9 mainly deals with the principle, theory, algorithm, design, testing, implementation and prototyping on advanced computing and processing platforms for the massive MIMO technique, which will certainly be employed in 5G standards. Core processing blocks, such as downlink precoding, uplink detection and channel estimation, are reviewed first, after which the emphasis is put on the various hardware implementation issues of massive MIMO, covering radio frequency (RF) front-end calibration, baseband processing,...