Secure UAV-aided Reinforcement Learning for Integrated Sensing and Communication

As communication and sensing technologies evolve, unmanned aerial vehicles (UAVs) are increasingly utilized for integrated sensing and communication (ISAC). Ensuring secure and efficient communication amidst potential eavesdroppers remains a critical challenge. This paper introduces secure UAV-aided reinforcement learning (SURE) for ISAC, a novel approach that addresses these security issues. Our method employs reinforcement learning, specifically proximal policy optimization (PPO), to optimize UAV trajectories, transmission power, and user scheduling. The goal is to enhance secure data transmission rates while mitigating interference from eavesdroppers. SURE-ISAC offers a comprehensive framework that balances secrecy rate maximization, destination reaching, and constraint satisfaction through an advanced reward structure. Extensive simulations show that SURE-ISAC significantly improves secure transmission rates compared to traditional methods, even with multiple eavesdroppers present. Our results demonstrate that this approach not only enhances UAV communication security but also effectively manages various operational constraints.