Integrated Communication and Control for Intelligent Formation Management of AAV Swarms: A Deep Reinforcement Learning Approach

Thanks to the comprehensive ability of information collection, integration and processing, autonomous aerial vehicles (AAV) swarm is playing a crucial role for remote monitoring or task executing in various fields. In this letter, the formation management of AAV swarms is studied with the aid of integrated communication and control (ICAC), where the leading AAV is responsible for controlling the following AAVs wirelessly in order to achieve the desired flight states. The real-time control signals’ generation, subcarrier, allocation, and data rate optimization are obtained with the aid of the deep deterministic policy gradient (DDPG) algorithm, by avoiding the potential collisions among AAVs. Simulation results demonstrate that our scheme is able to not only ensures collision avoidance but also reduce the control mean maximum square error (MMSE) with a better convergence compared to other benchmarks.