Optimal resource allocation for unmanned aerial vehicle- assisted wireless communications

Abstract

Unmanned aerial vehicles (UAVs) have had an impressive number of real-world applications and will continue to play a significant role in the future. Yet, UAV-assisted communication is constrained by scarce resources - a common issue in wireless communication. Optimal resource allocation is thus of critical importance for UAVs to operate and fulfil their missions. Although resource allocation in UAV-assisted communication is not a new topic, many challenges exist. Resource allocation is a non-trivial task due to the constraints of UAVs (such as flight time, deployment strategy, cache storage) and the many constraints of the wireless network supported by UAVs (such as power of the base station, quality-of-service), amid the presence of numerous users and devices. Moreover, optimisation problems in this context are often highly non-convex and difficult to solve.

Inspired by the aforementioned discussion, this thesis proposes optimal resource allocation strategies in UAV-assisted wireless communication, taking into account resources such as spectrum, power, and cache in specific UAV use cases. In particular, Chapter 3 looks at a spectrum-sharing cognitive radio network where the UAVs are deployed as flying base stations to provide network coverage to the secondary network in a disaster area. A learning-aided optimisation scheme is designed to allocate radio resources under the constraints of maximum tolerable interference. Chapter 4 considers integrating reconfigurable intelligent surfaces onboard the UAVs to extend network coverage in a massive multiple-input multiple-output system. The joint problem of optimal power allocation and phase-shift is solved, subject to deployment strategy and minimum data throughput. Finally, in Chapter 5, the UAVs assist in content caching in an integrated terrestrial-non terrestrial network. The joint optimisation problem of user clustering, cache placement, and power allocation is solved efficiently by using a distributed approach. In all these cases, low-complexity algorithms are proposed and their usefulness is confirmed through simulation.

Date of Award	Jul 2024
Original language	English
Awarding Institution	Queen's University Belfast
Sponsors	Northern Ireland Department for the Economy
Supervisor	Francisco Garcia-Palacios (Supervisor) & Thai Son Mai (Supervisor)