Trajectory Design in multi-UAV-assisted RSMA Downlink Communication

Abstract

In this study, we investigate the multi-UAV trajec-tory design problem for downlink rate splitting multiple (RSMA) access, and design the movement function for the UAVs with its corresponding constraint. Furthermore, RSMA physical layer is the promising technique, which is believe to be able to enhance the robustness to imperfect channel state information (CSI) and massive machine type communication (MTC). In particular, we consider the sum-rate maximization objective, in which the scheduling matrix, variables for proposed moving function, precoding matrix, common rate vector are jointly optimized. Since the objective function with the corresponding constraints are non-concave, we proposed the multi-agent-deep- reinforcement-learning (DRL)-based deep deterministic policy gradient (DDPG) scheme without knowing a priori knowledge of the dynamic environment. Additionally, mapping function for output actions are also proposed. Simulation results are conducted and demonstrated the effectiveness of our proposed method.