RESOURCE ALLOCATION FOR NON-ORTHOGONAL MULTIPLE ACCESS NETWORKS

Abstract

With the rapid growth of multimedia applications, spectrum scarcity remains one of the most important challenges in sixth-generation (6G) wireless communication networks. Due to its superiority over orthogonal multiple access (OMA) in terms of improved spectral efficiency, Non-orthogonal multiple access (NOMA) has been extensively studied by academia and industry in recent years. The key idea of NOMA techniques is multiple mobile users are served at the same time/frequency/code with diverse power levels, where the spectral efficiency gain is obtained by opportunistically exploiting the users' dynamic channel conditions or heterogeneous quality of service (QoS) requirements. Resource allocation, as one of the most important aspects of NOMA, is investigated for various NOMA-enabled networks in this thesis. In particular, the resource allocation for the unmanned aerial vehicle (UAV)-reconfigurable intelligent surfaces (RIS)-NOMA networks, wireless power transfer (WPT) in NOMA networks and backscattering communication-NOMA (BAC-NOMA) are formulated as different optimization problems and then are solved by convex optimization or machine learning (ML) based methods. First, this thesis investigates the combination of UAV downlink networks and RIS-assisted NOMA. A novel UAV-RIS-NOMA scheme is proposed, in which the strong user's data rate is maximized while guaranteeing the weak user's QoS after the UAV location is pre-optimised. The beamforming vectors of the base station and the phase shift of the RIS are jointly optimized, where the optimal beamforming vectors are derived with closed-form expressions and the phase shift optimization is performed by applying two algorithms (i.e., semi-definite relaxation (SDR) and successive convex approximation (SCA)). Second, if multiple users are taken into account in UAV downlink network with RIS-assisted NOMA, jointly optimizing beamforming, phase shift and UAV horizontal position becomes a more difficult problem. In this study, downlink users' sum rate is maximized by using a deep reinforcement learning (DRL) method, namely, deep deterministic policy gradient (DDPG), where the constraints adaptive problem has been analysed and solved in a proper way. Third, the application of wireless power transfer (WPT) to the NOMA downlink network is investigated. A new energy and spectrum cooperation scheme among WPT-deployed devices and NOMA downlink users is proposed. The energy efficiency of WPT devices is maximized while guaranteeing NOMA downlink users' QoS. Time switching coefficient and beamforming vectors are alternatively optimized by the proposed algorithm based on Dinkelbach method and the quadratic transform, respectively. The DDPG-based algorithm is also performed to solve the optimization problem, and compared with the proposed algorithm. To further demonstrate their respective strengths, robust beamforming designing is also realized by using the proposed algorithm and the DDPG-based algorithm. Fourth, a novel energy and spectrum cooperation BAC-NOMA transmission with hybrid successive interference cancellation (SIC) is proposed. The scenario of backscattering devices transmission to a data fusion centre is viewed as NOMA uplink and hence hybrid SIC strategy can be used to enhance its performance further. The closed-form optimal backscattering coefficients for one of the decoding cases is provided, and then the beamforming vector is optimized by applying SDR. For the other decoding case, backscattering coefficients and beamforming are jointly optimized by the proposed alternating algorithm based on SDR and SCA.

Date of Award	1 Aug 2023
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Emad Alsusa (Supervisor) & Zhiguo Ding (Supervisor)