ISAC-Enabled Multi-UAV Cooperative Perception and Trajectory Optimization

In recent years, unmanned aerial vehicles (UAVs) have experienced rapid development and have been widely used in many fields. Equipped with both communication modules and sensing modules, UAVs are capable of conducting integrated communication and target detection, thus greatly improving spectrum efficiency and system performance. In this article, we consider a scenario where multiple UAVs collaborate to detect targets and transmit the collected data to a central UAV. Addressing the problem of communication and perception scheduling, we first analyze the target detection and communication performance, and then formulate the joint communication and perception scheduling problem as two optimization problems, with the objectives being maximizing the average utility function (MAUF) and minimizing the completion time (MCT), respectively. To solve the formulated problems, we first consider the dynamic characteristics of the environment, and model the problems as two Markov decision processes. Regarding the UAVs as multiple agents, we then propose a multiagent double deep Q-network (DDQN)-based MAUF algorithm and a multiagent DDQN-based MCT algorithm to determine the communication and perception scheduling strategies of the UAVs. Simulation results demonstrate the effectiveness and superiority of the proposed algorithms.