UAV-IRS-Assisted ISAC Secure Transmission Design

Integrated Sensing and Communication (ISAC), as a core pillar of 6G, provides critical enablement for innovative applications across industries by sharing hardware and spectrum resources. However, in complex and dynamic environments, blockage and malicious eavesdropping pose serious challenges to secure transmission and accurate sensing. In this paper, an intelligent reflecting surface (IRS) aided ISAC system for physical-layer security (PLS) is considered, in which a co-design of base station (BS) beamforming and IRS phase shift matrix is proposed to enhance the security performance. The optimization seeks to maximize the long-term average secrecy rate of legitimate users (LUs), while a minimum echo SNR and the BS transmit power limitation are imposed as constraints. To address the inherent nonconvexity of the optimization, we employ two deep reinforcement learning (DRL) algorithms. Simulation validated the advantages of the two DRL approaches in terms of efficiency and scalability. The introduction of IRS within the ISAC system delivers significant performance gains, further demonstrating the promising future of IRS-enabled ISACs in 6G.