Joint UAV Trajectory and Beamforming Design for RIS-aided Integrated Sensing and Communication System

Integrated sensing and communication (ISAC) is developing rapidly owing to the advantages of energy saving and hardware cost reduction. With the aim of pursuing sensing and communication performance improvement, we investigate to deploy reconfigurable intelligent surface (RIS) and Unmanned aerial Vehicles (UAV) to aid the ISAC system serving multiple user equipments (UE) and targets, and focus on both the trajectory design of the UAV and decoupled uplink (UL)-downlink (DL) user association for such networks. Accordingly, We formulate a joint optimization problem to maximize the sum of sensing rate and UE communication rate. To address the non-convexity ISAC design problem above, we propose a distributed multi-agent deep reinforcement learning (MADRL) based design solution to alternately optimize the UAV's transmit beamforming and the RIS's matrix phase in a low-complexity manner. Simulation results illustrate the advantages of deploying UAV and RIS in ISAC systems and show that the proposed algorithm outperforms the baseline algorithms. .