Energy-Efficient Joint Trajectory and Reflecting Design in IRS-Enabled UAV Edge Computing

Intelligent reflecting surface (IRS) enabled unmanned aerial vehicle (UAV) edge computing, a new communication technology, can provide sufficient capacity for edge computing system. However, due to the line-of-sight or the non line of sight of communicating environments will impact transmitting rate or delay, the IRS can be utilized to compensate the channel fading in the IRS-enabled UAV edge computing. In this article, the joint problem of IRS phase shift, UAV trajectory and power allocation in the system is investigated, aiming to maximize the energy efficient. The corresponding optimization problem, which consists of mixed integer nonlinear programming problem, is formulated. To solve the problem, the original problem is decomposed into two subproblems, and an iterative method framework based on convex optimization and deep reinforcement learning is proposed. Given the UAV trajectory and IRS phase shift, the convex optimization algorithm is used to solve the power allocation schemes. Then, given the power allocation schemes, the double deep Q network and deep deterministic policy gradient algorithms are utilized to solve the problem of optimal UAV trajectory and IRS phase shift. The simulation results demonstrate that our proposed method outperforms other schemes in terms of energy efficiency, providing significant enhancements.

I. Introduction

Recent years, utilizing unmanned aerial vehicles (UAVs) to assist network communications in edge computing systems is becoming increasingly popular [1]. UAVs can move in 3-D space and thus are not affected by obstacles on the ground [2], [3]. Due to the flexible mobility, UAVs-assisted network communications are ideally suited for network deployments in emergency scenarios, such as disaster area communications, UAV-assisted ground vehicle or marine craft communications to improve the users’ Quality of Service (QoS) [4], [5], [6], [7], [8], [9]. However, communication for UAVs will face a tremendous challenge when encountering scenarios with large amounts of obstacles blocking or stronger channel fading [10], [11], [12]. Therefore, the utilization of intelligent reflective surface (IRS) or reconfigurable intelligence surface (RIS) to assist network communication in UAV edge computing systems has caught researchers’ attention in many scenarios [13], [14], [15], [16].