15 results in Full-Duplex Communications and Networks
References
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp 312-330
-
- Chapter
- Export citation
2 - Signal Processing and Theoretical Limits
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp 50-89
-
- Chapter
- Export citation
-
Summary
Mitigation of self-interference is the prime challenge in making full-duplex technology feasible in wireless communications. However, self-interference poses a difficult problem in general and is not limited to full-duplex communications. A wireless transceiver may suffer from self-interference even when transmit and receive frequencies are not the same, because signal from the transmitted chains leaks into the receiver chain inside the device. For this reason, 2G GSM handsets do not transmit and receive during the same time slot, although uplink and downlink frequencies are different, as in Frequency Division Duplex (FDD) systems. This is possible because one frame in GSM is divided into eight time slots, each slot is used by a different user, and uplink and downlink transmissions by one user take place in different slots.
Limiting the number of available slots for communication naturally limits the data rates of a user, and in the next generation this limitation was relaxed. In 3G WCDMA, handsets are able to transmit and receive simultaneously in FDD mode. This causes self-interference inside a device when the transmitted signal leaks in the circuit board to the receiver chain. As a matter of fact, 3G handsets require a duplex filter, making the implementation more complex and expensive than that of GSM. Thus, mitigation of self-interference is required also when using different frequencies for transmission and reception in handsets.
On the base station side, co-siting of WCDMA-TDD and WCDMA-FDD base stations operating on adjacent frequencies is not considered commercially viable [110] due to the mutual interference. This has been further observed in [111], concluding that 4G LTE-TDD adjacent band interference is harmful to uplink LTE-FDD received signals when the LTE-TDD network operates in the 1900–1920 MHz band and LTE-FDD above the 1920 MHz band. Thus, self-interference is also a problem in base stations when transmit and receive bands are different but close to each other. Therefore, selfinterference cancellation techniques developed for in-band full-duplex communication can be useful for wireless transceivers in general.
In this chapter, we first present a system model of a wireless communication link employing a full-duplex transceiver. Then we present signal processing techniques in digital baseband that aim to mitigate self-interference. In addition to digital cancellation, other necessary techniques to mitigate self-interference are antenna isolation between transmitter and receiver antennas, and different analog cancellation techniques.
1 - Basics of Communication Systems
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp 1-49
-
- Chapter
- Export citation
-
Summary
A wireless network refers to a telecommunications network in which interconnections between nodes are implemented without the use of wires. Wireless networks have experienced unprecedented growth over the past few decades, and they are expected to continue to evolve in the future. Seamless mobility and coverage ensure that various types of wireless connections can be made anytime, anywhere. In this chapter, we first introduce some advanced types of wireless communication techniques. Next, we study the heterogeneous wireless networks, and finally we provide the reader with some necessary background on the state of the art in full-duplex development.
Advanced Wireless Communication Technology
In this section, we discuss some of the current state of the art in wireless technologies related to D2D communications.
OFDM/OFDMA Technology
Orthogonal Frequency-Division Multiplexing (OFDM) is a technique for transmitting multiple digital signals simultaneously over a large number of orthogonal subcarriers. Based on the fast Fourier transform algorithm to generate and detect the signal, data transmission can be performed over a large number of carriers that are spaced apart at precise frequencies. The frequencies (or tones) are orthogonal to each other. Therefore, the spacing between the subcarriers can be reduced and hence high spectral efficiency can be achieved. OFDM transmission is also resilient to interference and multipath distortion which causes Inter-Symbol Interference (ISI).
OFDM transmitter and receiver block diagrams are shown in Fig. 1.1 and Fig. 1.2, respectively. s[n] is a serial stream of binary digits to transmit. After serial-to-parallel conversion, the data is split into N streams. Each stream is then coded to X0, … XN−1 with possibly different modulation methods (such as PSK and QAM) depending on the subchannel condition. An inverse FFT is computed on each set of symbols, giving a set of complex time-domain samples. These samples are then quadrature-mixed to passband in the standard way: the real and imaginary components are first converted to the analog domain using digital-to-analog Converters (DACs); the analog signals are then used to modulate cosine and sine waves, respectively, at the carrier frequency, fc. These signals are then summed to yield the transmission signal, s(t). The receiver picks up the signal r(t), which is s(t) transmitted through radio channels and contaminated by noise. Then r(t) is quadrature-mixed down to baseband using cosine and sine waves at the carrier frequency.
5 - Full-Duplex OFDMA Communications
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp 138-178
-
- Chapter
- Export citation
-
Summary
In this chapter, we study the resource allocation and scheduling problem for a Full- Duplex (FD) Orthogonal Frequency-Division Multiple-Access (OFDMA) network where an FD Base Station (BS) simultaneously communicates with multiple pairs of UpLink (UL) and DownLink (DL) Half-Duplex (HD) users bidirectionally. We aim to find the optimal pairing of UL and DL users for each FD link and the allocation of the OFDM subcarriers and power to different pairs such that the sum rate of the network is maximized. This is a traditional combinatorial problem and the optimal approach requires an exhaustive search, which becomes prohibitively complicated as the number of users and subcarriers increase. In this chapter, we introduce two alternative ways of solving such a complex problem. In the first approach we formulate the problem as a mixed-integer nonlinear programming problem and solve it by using the dual method and Sequential Parametric Convex Approximation (SPCA). In the second approach, we introduce a low-complexity distributed approach based on matching theories and present an efficient near-optimal matching algorithm for resource allocation in FD-OFDMA systems.
FD-OFDMA Model
System Model
We consider a multi-user FD-OFDMA system, as shown in Fig. 1, consisting of one FD BS and multiple uplink (UL) and downlink (DL) half-duplex (HD) users, each with a single antenna. The UL and DL users are paired to form an FD transceiver unit in which the UL user acts as a Tx (transmit) user and the DL one acts as an Rx (receive) user, which communicate with the BS simultaneously. Note that each subcarrier is assigned to at most one transceiver unit only, but each transceiver unit can utilize more than one subcarrier. Let KSC = ﹛1, 2, · · · K﹜ denote the set of subcarriers,MT = ﹛U1,U2, · · ·, UM﹜ the set of UL Tx users, NR = ﹛D1,D2, · · ·, DN﹜ the set of DL Rx users, and (Um,Dn) a transceiver unit consisting of Um and Dn. We assume that the numbers of Tx users, Rx users, and the subcarriers are not necessarily the same.
We consider a block fading channel, for which the channel remains constant within a time slot, but varies independently from one to another. The channel coefficient from the BS to user Dn, and that from user Um to the BS, on subcarrier k are denoted by hB,n,k and hm,B,k, respectively.
3 - Full-Duplex System Hardware Implementation
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp 90-108
-
- Chapter
- Export citation
-
Summary
In this chapter we outline recent hardware development of full-duplex transceivers. The focus is on the development of compact full-duplex transceivers suitable for wireless communication systems. Isolation between the transmitter and receiver chains is limited when using the same antenna for transmitting and receiving. This kind of antenna sharing requires excellent performance of analog and digital cancellation techniques to mitigate self-interference. Therefore, full-duplex technology is more likely to be implemented first in access points and relays that may accommodate separate transmitter and receiver antenna arrays to improve passive isolation. This makes the task of analog and digital cancellation easier, although they are still needed because passive isolation is generally not enough in the case of compact devices. Wireless relays may be preferred, because they have inherent traffic symmetry irrespective of uplink and downlink traffic ratio, and the gains from full-duplex operation are largest when the traffic is symmetric.
Full-Duplex Technology Used in Wireless Systems
In-band full-duplex technology has been utilized in wireless communications systems for a long time in continuous-wave radars and in-channel repeaters in broadcast and cellular systems. Not surprisingly, self-interference has been one of the key challenges in these systems.
Continuous Wave (CW) radar systems use either two separate antennas or one shared antenna to transmit and receive simultaneously [160], whereas pulsed radar systems switch off the transmitter while radar returns are collected. The conventional CW radars of the 1940s and 1950s achieved isolation between the transmitter and receiver through antenna separation-based path-loss in separate-antenna systems, or through the use of circulators in shared-antenna systems. Because only mild levels of isolation could be achieved using these techniques, keeping self-interference to a manageable level required the transmission power to be greatly limited, which then limited the detectable range of targets. This restriction of CWradar to nearby targets (i.e., short ranges) turned out to be useful, because detecting nearby targets with a pulsed radar system would require on/off switching times that are impractically small. Thus, the operational range of CW radars matched their needs.
In the 1960s, an analog circuit-based form of self-interference cancellation was proposed to increase the dynamic range of CW radars [161].
8 - Full-Duplex Cognitive Radio Networks
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp 247-290
-
- Chapter
- Export citation
-
Summary
With the proliferation of wireless services and the ever increasing data rate demands, spectrum resources have become more and more scarce. As a promising technique to increase the efficiency of spectrum utilization, Cognitive Radio (CR) has great potential to meet such a requirement by allowing unlicensed users (Secondary Users, SUs) to coexist in licensed bands allocated to licensed users (Primary Users, PUs). In conventional CR systems, the spectrum sensing is performed at the beginning of each time slot before the data transmission, which is also known as the “listen-before-talk” protocol. While this protocol has worked well in CR networks, two inherent problems still exist: 1) transmission time reduction due to sensing, and 2) sensing accuracy impairment due to data transmission.
While a great many works have discussed the design of sensing interval and duration to minimize the impact of the above problems, we, on the other hand, manage to apply FD technology to CR networks to bypass the problems. Specifically, FD technology enables simultaneous sensing and transmission for SUs. In other words, neither sensing nor data transmission needs to be interrupted by the other. In this way, SUs can react promptly to the PUs’ access and departure, and possibly fully utilize spectrum opportunities for data transmission.
In this chapter, we first provide some preliminaries of cognitive radio and some existing works. Then, we elaborate the proposed FD CR protocol, named the “Listen- And-Talk” (LAT) protocol, and provide detailed analysis about the parameter design and a unique trade-off between secondary transmit power and secondary throughput. Based on the basic LAT protocol that evolves only one pair of SUs, we extend the scenario to cooperative spectrum sensing and dynamic spectrum access, respectively, in the next two sections. Finally, we list some key challenges in the design and implementation of FD CRNs.
Cognitive Radio Basics
The existing and new wireless technologies, such as smart phones, wireless computers, and WiFi home and business networks, are rapidly consuming the radio spectrum. Unlike the wired Internet, the wireless world has a limited number of links to distribute. The usage of radio spectrum resources and the regulation of radio emissions are coordinated by national regulatory bodies like the Federal Communications Commission (FCC). These bodies assign spectrum to licensed holders, also known as Primary Users (PUs), on a long-term basis for large geographical regions.
Index
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp 331-334
-
- Chapter
- Export citation
9 - Full-Duplex Random Access Networks
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp 291-311
-
- Chapter
- Export citation
-
Summary
Traditionally, the spectrum allocation policy grants fixed spectrum bands to licensed users for exclusive access, which worked well over the past decades. However, with the proliferation of wireless services and data volume in recent years, spectrum scarcity, as a major drawback of this static spectrum allocation policy, has been unveiled. Meanwhile, it is observed that a significant amount of the licensed spectrum is rather underutilized.
As an opposite of the conventional static spectrum management policy, the concept of dynamic spectrum access has been proposed to increase the flexibility in spectrum usage, so as to alleviate the spectrum scarcity problem and improve spectrum utilization. In the existing literature, Dynamic Spectrum Access (DSA) models can be categorized as follows: exclusive-use, shared-use, and commons models. In the exclusive-use model, a licensed user can grant an unlicensed user the spectrum access rights to have exclusive access to the spectrum. In a shared-use model, an unlicensed user accesses the spectrum opportunistically without interrupting a licensed user. In a commons model, an unlicensed user can access the spectrum freely. DSA can be implemented in a centralized or a distributed network architecture. DSA can be optimized globally if a central controller is available in the network. On the other hand, when a central controller is not available, distributed algorithms would be required for dynamic spectrum access. Issues related to spectrum trading, such as pricing, will also need to be considered for dynamic spectrum access, especially with the exclusive-use model. For DSA-based radio networks, MAC protocols designed for traditional wireless networks have to be modified to include spectrum sensing, spectrum access, as well as spectrum trading between a licensed user and an unlicensed user.
Most current dynamic spectrum access paradigms are designed for HD devices with the basic assumption that data transmission and reception of any devices must be separated in the time or frequency domain. Recently, with the rapid development of self-interference suppression techniques, FD communication rapidly extends its application to different scenarios. In FD communications where co-channel simultaneous data transmission and reception becomes possible, many more possible communication modes among communication devices arise. For example, two FD devices can perform bidirectional transmission to each other; one FD devices can receive data from a source, and transmits data to another destination concurrently on the same channel.
Preface
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp xi-xvi
-
- Chapter
- Export citation
-
Summary
Overview
With more and more new multimedia-rich services being introduced and offered to a rapidly growing population of global subscribers, there is an ever-increasing demand for higher data rate wireless access, making more efficient use of this precious resource a crucial need. As a consequence, new wireless technologies such as Long Term Evolution (LTE) and LTE-Advanced have been introduced. These technologies are capable of providing high-speed, large-capacity, and guaranteed quality-of-service (QoS) mobile services.With the technological evolution of cellular networks, new techniques, such as small cells, have also been developed to further improve the network capacity by effectively reusing the limited radio spectrum. However, all existing wireless communication systems deploy half-duplex (HD) radios which transmit and receive the signals in two separate/orthogonal channels. They dissipate the precious resources by employing either time-division or frequency-division duplexing.
Full-duplex (FD) systems, where a node can send and receive signals at the same time and frequency resources, can offer the potential to double spectral efficiency; however, for many years it has been considered impractical. This is because the signal leakage from the local output to input, referred to as self-interference, may overwhelm the receiver, thus making it impossible to extract the desired signal. How to effectively eliminate self-interference has remained a long-standing challenge. Recently, there has been significant progress in self-interference cancellation in FD systems, which presents great potential for realizing FD communications for the next generation of cellular networks.
This book provides state-of-the-art research on FD communications and cellular networks covering the physical, MAC, network, and application layer perspectives. The book also includes fundamental theories based on which FD communications will be built. In addition to the self-interference cancellation signal processing algorithms, the book discusses physical layer algorithms, radio resource allocation and network protocols in the practical design and implementation of centralized and distributed FD wireless networks. Main applications such as FD cognitive radio networks, FD cooperative networks, and FD heterogeneous networks are explored.
The key features of this book are as follows:
• A unified view of FD communications and networking;
• A comprehensive review of the state-of-the-art research and key technologies of FD communications networks;
• Coverage of a wide range of techniques for design, analysis, optimization, and application of FD communications networks;
• An outline of the key research issues related to FD communications and networking.
Dedication
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp v-vi
-
- Chapter
- Export citation
6 - Full-Duplex Heterogeneous Networks
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp 179-204
-
- Chapter
- Export citation
-
Summary
The continuous growth in the numbers of mobile users and the increase in data communication make deploying new techniques with increased data rates a crucial need. In other words, modern network topologies are required to give the required performance boost. Since decreasing the number of users in each cell, as well as decreasing the distance between the users and the base station, increases the total achieved capacity, therefore, small cells are required. However, it is well known that large cells are the cost-effective way to serve fast-moving users. Therefore, one solution to match these demands is to employ Heterogeneous Networks (HetNets). The main idea of HetNets is to densify the existing macro-cell by adding a mix of pico, femto and relay base stations to which users will be offloaded from the macro base station. This will help in increasing the network capacity in hot spots and in giving better coverage for both outdoor and indoor areas not covered by the macro network.
Furthermore, both the efficiency of utilizing the available communication resources and the capacity achieved by HetNets can be further improved by deploying Full Duplex (FD). Theoretically speaking, allowing the network nodes to simultaneously transmit and receive data at the same channel and in the same time slot achieves double the capacity achieved by Half Duplex (HD) communication when using the same resources. However, it must be mentioned that FD was considered unfeasible for a long time as the increase in networks’ capacity offered by deploying FD is restricted by the ability to suppress the self-interference which is caused by the node's transmission on the node's reception and results in a great degradation in the received Signal-to-Interference Noise Ratio (SINR). Recent evolution in self-interference cancellation techniques revived the attention to FD. As mentioned, adding FD communication to HetNets will improve the network efficiency and capacity. Accordingly, efficient resource allocation techniques are needed to fully benefit from the available resources and to overcome the challenges that arise from deploying FD to HetNets.
7 - Full-Duplex Cooperative Networks
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp 205-246
-
- Chapter
- Export citation
-
Summary
In this chapter, we introduce Full Duplex (FD) cooperative networks, consisting of one source node, multiple relays, and one destination. We assume that the relay works in a full duplex mode. Due to the residual self-interference in FD relay nodes, the analysis of FD relay systems will become essentially different from conventional relay systems. We will analyze the performance of FD Amplify-and-Forward (FA) relay systems. The performance of FD AF relay systems is limited by the residual self-interference. To overcome such a limitation, we will introduce an effective X-duplex relaying protocol, which switches adaptively between FD and Half-Duplex (HD) spatial multiplexing modes based on the channel conditions, such that the limitations of FD and HD can be overcome through such an adaptive protocol. Finally, we consider a general setup of multi-relay FD systems and introduce a joint antenna and relay selection scheme for such a network.
Cooperative Communication Basics
Before introducing FD relay networks, in this section, we first briefly introduce some basics of cooperative communications and relaying protocols.
The concept of cooperative communications can actually be traced back to the early work of Cover and ElGamal on achievable capacity of a relay network in 1979; it was rediscovered recently in relation to great potential applications in cellular and wireless sensor networks. The distributed nature of wireless networks provides a unique opportunity for cooperation and distributed signal processing. The design of efficient cooperative protocols and distributed signal processing techniques has been an important issue in implementing cooperative communications in wireless networks. Therefore, recent research on cooperative wireless networks has focused on designing relaying protocols, signaling and distributed coding. In particular, the design of efficient relaying protocols and distributed coding schemes has attracted significant attention and a number of novel relaying protocols and distributed coding schemes have been developed in the past several years. Capacity-approaching performance has been achieved by some elegantly designed distributed coding schemes.
When we talk about “cooperative networks” and “relay networks,” they have the following distinctions. In cooperative networks, each node acts as both a source and a relay node. That is, each node not only transmits its own information but also helps other nodes to transmit signals. In relay networks, the relay nodes are explicitly built nodes only for the purpose of relaying and forwarding information.
Contents
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp vii-x
-
- Chapter
- Export citation
4 - Full-Duplex MIMO Communications
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp 109-137
-
- Chapter
- Export citation
-
Summary
Thanks to the various state-of-the-art approaches for self-IC schemes, SI is no longer a critical bottleneck when implementing a practical FD system. In this chapter, we introduce FD MIMO communications, including some key techniques and some performance analysis. The FD MIMO advantages can be summarized as: efficient and flexible utilization of wireless communication resources; increasing the capacity of the communication networks; and guaranteeing reliable communication. These have all become crucial claims for the next generation of cellular networks. Full Duplex (FD) is a very promising technique that allows for efficient use of wireless communication resources, given that the self-interference level can be suppressed to an acceptable level.
In the following, we describe a few signal processing techniques of the FD MIMO system where two nodes bidirectionally communicate with each other. Firstly, we describe the mode switching between FD and half-duplex spatial-multiplexing (HDSM). By configuring the antennas as either transmit or receive antennas, the MIMO system can be configured as either an FD system or an HD-SM system; these are considered two important techniques for improving the spectral efficiency of MIMO systems. FD transmission suffers from self interference, while the performance of an SM system is greatly affected by spatial correlation. Therefore, there is an optimal trade-off between FD and SM, depending on the system setting and channel conditions. Then, we introduce the antenna pairing strategy, where each node is equipped with two antennas, used for either transmission or reception. Specifically, one transmit and receive antenna combination is selected based on two system performance criteria: 1) maximum sum-rate (Max-SR), and 2) minimum symbol-error-rate (Min-SER). We further extend our strategy to the scenario where each node is equipped with an arbitrary number of antennas. We describe bidirectional link selection schemes by selecting a pair of transmit and receive antenna at both ends for communications in each direction, to maximize the weighted sum-rate or minimize the weighted sum SER. Then, we introduce an X-Duplex scheme, where the antenna is adaptively configured based on the channel conditions. The X-Duplex scheme aims to maximize the instantaneous sum-rate of the system. Finally, we conclude this chapter and discuss some of the key challenges in FD MIMO communications.
FD MIMO Signal Processing
Mode Switching between Full-Duplex and Half-Duplex
Typically, there are two kinds of baseline for FD and HD mode switching: a fixed number of antennas and a fixed number of RFs.
Frontmatter
- Lingyang Song, Peking University, Beijing, Risto Wichman, Aalto University, Finland, Yonghui Li, University of Sydney, Zhu Han, University of Houston
-
- Book:
- Full-Duplex Communications and Networks
- Published online:
- 16 March 2017
- Print publication:
- 02 March 2017, pp i-iv
-
- Chapter
- Export citation