Hierarchical Multi-Agent Reinforcement Learning-Based UAV Control for Wireless Covert Communications

Abstract

In this study, we consider wireless covert communication within unmanned aerial vehicle (UAV) environments. Here, the UAV functions as a covert transmitter, sending data to predetermined ground receivers while avoiding detection by ground-based detectors. We aim to maximize the UAVs' through-put and the detector's minimum detection error probability by optimizing the UAV's transmission power and positioning through Q-learning. We utilize reinforcement learning to de-termine UAVs' optimal transmission power and location in complex environments, ensuring effective problem-solving even in challenging scenarios.