Skip to content
Open access

3D Trajectory and Resource Optimization for UAV-IRS-Assisted Covert Communications via Deep Reinforcement Learning

2026 · IEEE Transactions on Communications · Vol 74, pp. 15124-15141 · 0 citations · 57 references
Computer Science

Abstract

This paper investigates a uncrewed aerial vehicle (UAV)-intelligent reflecting surface (IRS)-assisted covert communication system under heterogeneous eavesdropper channel knowledge, where channel state information (CSI) is available for the line-of-sight (LoS)-dominated UAV–Willie link, while only channel distribution information (CDI) is available for the Alice–Willie link due to blockage and rich scattering. Under the minimum detection error probability (DEP)-based covertness criterion, the objective is to maximize the average covert communication rate (ACCR) through the joint optimization of the UAV three dimensional (3D) trajectory, IRS phase shifts, and transmit power. The resulting problem is highly coupled and non-convex due to heterogeneous channel knowledge, covert constraints, and UAV mobility. To address this challenge, we reformulate the problem as a Markov decision process and propose a deep deterministic policy gradient with reward shaping (DDPG-RS) framework. A constraint-aware reward shaping mechanism provides informative feedback to alleviate sparse-reward learning and improve convergence stability. Simulation results demonstrate that the proposed scheme achieves higher ACCR and faster convergence than benchmark methods, providing an effective approximate solution to the formulated constrained optimization problem.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.