Full-Duplex Device-to-Device Communication for 5G Network

With the rapidly growing of the customers? data traffic demand, improving the system capacity and increasing the user throughput have become essential concerns for the future fifth-generation (5G) wireless communication network. In this context, device-to-device (D2D) communication and in-band full-duplex (FD) are proposed as potential solutions to increase the spatial spectrum utilization and the user rate in a cellular network. D2D allows two nearby devices to communicate without base station (BS) participation or with limited participation. On the other hand, FD communication enables simultaneous transmission and reception in the same frequency band. Due to the short distance property of D2D links, exploiting the FD technology in D2D communication is an excellent choice to further improve the cellular spectrum efficiency and the users? throughput. However, practical FD transceivers add new challenges for D2D communication. For instance, the existing FD devices cannot perfectly eliminate the self-interference (SI) imposed on the receiver by the node?s own transmitter. Thus, the residual self-interference (RSI) which is tightly related to the transmitter power value highly affects the performance of FD transmission. Moreover, the FD technique creates additional interference in the network which may degrade its performance when compared with the half-duplex transmission. Thus, proper radio resource management is needed to exploit the benefits of FD and guarantee the quality of service (QoS) of the users. The works in this dissertation focus on the power allocation (PA) and channel assignment (CA) of a full-duplex device-to-device (FD-D2D) network. In particular, this thesis first addresses the PA problem and proposes a simple yet efficient centralized optimal PA framework, and next, it derives the optimal joint PA and CA scheme for an FD-D2D network. A simple sub-optimal algorithm for resource allocation named CATPA, based on CA followed by PA, is also derived and proposed. This dissertation also develops, in the end, an efficient decentralized PA using game theory tools that will be an essential part of future works in the context of distributed radio resource management.