ScholarWorks@중앙대학교

초록

In this work, we investigate the performance of cognitive satellite-aerial networks (CSAN), where the satellite operates as the primary network (PN) and the unmanned aerial vehicle (UAV) acts as an aerial base station in the secondary network (SN), employing rate-splitting multiple access (RSMA) to serve user equipment (UEs). We address the problem of joint beamforming, rate allocation, UAV trajectory design, and reconfigurable intelligent surface (RIS) phase shift optimization in a CSAN. To maximize the sum rate while satisfying power and interference constraints, we propose a deep reinforcement learning (DRL) framework based on the twin delayed deep deterministic policy gradient (TD3) algorithm. Simulation results show that the proposed TD3-based scheme outperforms other DRL-based benchmarks.

키워드