Resource Allocation Techniques for Non-Orthogonal Multiple Access in Beyond 5G

To support the wide range of envisioned applications, including autonomous vehicles, augmented reality, holographic communication, and Internet of Everything (IoE), future wireless networks must meet demanding requirements for higher spectral and energy efficiency, lower end-to-end latency and massive connectivity. This requires a vast upgrade in the technologies of the sixth-generation (6G) wireless networks. Non-orthogonal multiple access (NOMA) has been advocated as a prospective effective multiple access technique for future wireless networks due to the wide range of its potential benefits, including superior spectral efficiency (SE), energy efficiency (EE), compatibility, user fairness, and flexibility. To exploit additional degrees of freedom and address the computational complexity with massive connectivity, NOMA has been recently combined with different types of multiple access techniques and appropriate
optimization designs. Hence, this thesis attempts to utilize the combination of NOMA with different key technologies, including multiple antenna techniques, conventional OMA techniques, and intelligent reflecting surface (IRS). In particular, different resource allocation techniques have been developed for such integrated NOMA systems, from the downlink (DL) single-input single-output (SISO)-NOMA system, to DL multiple-input single-output (MISO)-NOMA system, as well as the IRS-assisted NOMA system.

Firstly, a hybrid time division multiple access (TDMA)-NOMA system is considered, where both the available time slots and the available transmit power are jointly allocated to maximize the global EE. To further exploit the promising advantages of this hybrid system, the SE-EE trade-off based design and max-min fairness based design are presented in this thesis. By utilizing different convex relaxation and approximation techniques, the non-convexity of the formulated optimization problems are transformed into convex problems. Finally, this thesis investigates a worst-case robust design for an IRS-assisted NOMA multi-user MISO system to maximize the EE with a set of quality of service (QoS) constraints. In particular, an iterative algorithm based on alternating optimization (AO) is proposed to design the transmit beamforming vectors at the base station (BS) and reflection coefficient matrix for IRS. The effectiveness advantages of all the proposed schemes are demonstrated through numerical simulation results.