Resource Allocation and 3D Trajectory Design for Power-Efficient IRS-Assisted UAV-NOMA Communications

TL;DR

This paper proposes a joint resource allocation, 3D UAV trajectory, and IRS phase control strategy to minimize energy consumption in IRS-assisted UAV-NOMA systems, leveraging deep neural networks for channel approximation.

Contribution

It introduces a novel joint optimization framework for UAV trajectory, IRS phase control, and resource allocation, incorporating deep learning to handle altitude-dependent channel variations.

Findings

IRS enhances UAV trajectory flexibility and energy efficiency.

Passive beamforming with IRS reduces power requirements significantly.

Optimizing UAV altitude improves outage performance and power savings.

Abstract

In this paper, an intelligent reflecting surface (IRS) is introduced to assist an unmanned aerial vehicle (UAV) communication system based on non-orthogonal multiple access (NOMA) for serving multiple ground users. We aim to minimize the average total system energy consumption by jointly designing the resource allocation strategy, the three dimensional (3D) trajectory of the UAV, as well as the phase control at the IRS. The design is formulated as a non-convex optimization problem taking into account the maximum tolerable outage probability constraint and the individual minimum data rate requirement. To circumvent the intractability of the design problem due to the altitude-dependent Rician fading in UAV-to-user links, we adopt the deep neural network (DNN) approach to accurately approximate the corresponding effective channel gains, which facilitates the development of a low-complexity suboptimal iterative algorithm via dividing the formulated problem into two subproblems and address them alternatingly. Numerical results demonstrate that the proposed algorithm can converge to an effective solution within a small number of iterations and illustrate some interesting insights: (1) IRS enables a highly flexible UAV's 3D trajectory design via recycling the dissipated radio signal for improving the achievable system data rate and reducing the flight power consumption of the UAV; (2) IRS provides a rich array gain through passive beamforming in the reflection link, which can substantially reduce the required communication power for guaranteeing the required quality-of-service (QoS); (3) Optimizing the altitude of UAV's trajectory can effectively exploit the outage-guaranteed effective channel gain to save the total required communication power enabling power-efficient UAV communications.