Abstract:The past few years have seen a major change in cellular networks, as explosive growth in data demands requires more and more network capacity and backhaul capability. New wireless technologies have been proposed to tackle these challenges. One of the emerging technologies is device-to-device (D2D) communications. It enables two cellular user equip- ment (UEs) in proximity to communicate with each other directly reusing cellular radio resources. In this case, D2D is able to of oad data traf c from central base stations (BSs) and signi cantly improve the spectrum ef ciency of a cellular network, and thus is one of the key technologies for the next generation cellular systems. Radio resource management (RRM) for D2D communications and how to effectively exploit the potential bene ts of D2D are two paramount challenges to D2D communications underlaying cellular networks. In this thesis, we focus on four problems related to these two challenges. In Chapter 2, we utilise the mixed integer non-linear programming (MINLP) to model and solve the RRM optimisation problems for D2D communications. Firstly we consider the RRM optimisation problem for D2D communications underlaying the single carrier frequency division multiple access (SC-FDMA) system and devise a heuristic sub- optimal solution to it. Then we propose an optimised RRM mechanism for multi-hop D2D communications with network coding (NC). NC has been proven as an ef cient technique to improve the throughput of ad-hoc networks and thus we apply it to multi-hop D2D communications. We devise an optimal solution to the RRM optimisation problem for multi-hop D2D communications with NC. In Chapter 3, we investigate how the location of the D2D transmitter in a cell may affect the RRM mechanism and the performance of D2D communications. We propose two optimised location-based RRM mechanisms for D2D, which maximise the throughput and the energy ef ciency of D2D, respectively. We show that, by considering the location information of the D2D transmitter, the MINLP problem of RRM for D2D communications can be transformed into a convex optimisation problem, which can be ef ciently solved by the method of Lagrangian multipliers. In Chapter 4, we propose a D2D-based P2P le sharing system, which is called Iunius. The Iunius system features: 1) a wireless P2P protocol based on Bittorrent protocol in the application layer; 2) a simple centralised routing mechanism for multi-hop D2D communications; 3) an interference cancellation technique for conventional cellular (CC) uplink communications; and 4) a radio resource management scheme to mitigate the interference between CC and D2D communications that share the cellular uplink radio resources while maximising the throughput of D2D communications. We show that with the properly designed application layer protocol and the optimised RRM for D2D communications, Iunius can signi cantly improve the quality of experience (QoE) of users and of oad local traf c from the base station. In Chapter 5, we combine LTE-unlicensed with D2D communications. We utilise LTE-unlicensed to enable the operation of D2D in unlicensed bands. We show that not only can this improve the throughput of D2D communications, but also allow D2D to work in the cell central area, which normally regarded as a “forbidden area” for D2D in existing works. We achieve these results mainly through numerical optimisation and simulations. We utilise a wide range of numerical optimisation theories in our works. Instead of utilising the general numerical optimisation algorithms to solve the optimisation problems, we modify them to be suitable for the speci c problems, thereby reducing the computational complexity. Finally, we evaluate our proposed algorithms and systems through sophisticated numer- ical simulations. We have developed a complete system-level simulation framework for D2D communications and we open-source it in Github: https://github.com/mathwuyue/py- wireless-sys-sim .

An Autonomous Transmission Scheme Using Dueling DQN for D2D Communication Networks

A Deep Learning-Based Transmission Scheme Using Reduced Feedback for D2D Networks

A deep reinforcement learning-based D2D spectrum allocation underlaying a cellular network

Deep Reinforcement Learning for Joint Channel Selection and Power Control in D2D Networks

Machine Learning-Based Resource Optimization for D2D Communication Underlaying Networks

Double Deep Q-Network Based Distributed Resource Matching Algorithm for D2D Communication

Deep Multi-Agent Reinforcement Learning for Resource Allocation in D2D Communication Underlaying Cellular Networks

Hybrid Centralized-Distributed Resource Allocation Based on Deep Reinforcement Learning for Cooperative D2D Communications

Uplink channel reusing selection optimization for Device-to-Device communication underlaying cellular networks

Deep Reinforcement Learning-Based Dynamic Spectrum Access for D2D Communication Underlay Cellular Networks

Deep Reinforcement Learning-aided Transmission Design for Energy-efficient Link Optimization in Vehicular Communications

Joint Deep Reinforcement Learning and Unsupervised Learning for Channel Selection and Power Control in D2D Networks

Distributed Dynamic Spectrum Access for D2D Communications Underlying Cellular Networks Using Deep Reinforcement Learning

Energy-Efficient Resource Allocation Based on Deep Q-Network in V2V Communications

Distributed Deep Reinforcement Learning for Resource Allocation in Digital Twin Networks

Application of deep neural network and deep reinforcement learning in wireless communication

Distributed Two-tier DRL Framework for Cell-Free Network: Association, Beamforming and Power Allocation.

Secure and Reliable Downlink Transmission for Energy-Efficient User-Centric Ultra-Dense Networks: An Accelerated DRL Approach

Data-Rate Driven Transmission Strategies for Deep Learning-Based Communication Systems

Advanced Technologies for Device-to-device Communications Underlaying Cellular Networks

Parallel Digital Twin-driven Deep Reinforcement Learning for User Association and Load Balancing in Dynamic Wireless Networks