Joint Trajectory and Beamforming Optimization for Federated DRL-Aided Space-Aerial-Terrestrial Relay Networks With RIS and RSMA

To overcome the long transmission distances and limited spectrum resources issues, both the space-aerial-terrestrial relay networks (SATRNs) and hybrid-free space optical/radio frequency (FSO/RF) mode have attracted significant attentions. Specifically, high-altitude platform (HAP) and unmanned aerial vehicle (UAV) are employed in this paper to enhance the transmission reliability and improve the resource utilization along with the reconfigurable intelligent surface (RIS) and rate splitting multiple access (RSMA) techniques. Besides, we propose a novel access-free federated deep reinforcement learning (DRL) framework, which exploits the privacy-preserving security features of federated learning (FL) and DRL, to optimize active beamforming vectors, RIS reflection coefficients, UAV trajectory, and power splitting ratio. The learning process of the algorithm is performed locally which significantly reduces the computational overhead compared to traditional algorithms. Simulation results demonstrate that the proposed federated DRL-aided framework achieves higher energy efficiency compared to the reference schemes.