Simultaneously Transmitting and Reflecting Reconfigurable Intelligent Surface (STAR-RIS) Assisted UAV Communications

A novel air-to-ground communication paradigm is conceived, where an unmanned aerial vehicle (UAV)-mounted base station (BS) equipped with multiple antennas sends information to ground users (GUs) the aid of a simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS). In contrast conventional RIS whose main function reflect incident signals, STAR-RIS capable both impinging signals from either side surface, thereby leading full-space 360 degree coverage. However, transmissive reflective capabilities require more complex transmission/reflection coefficient design. Therefore, in this work, sum-rate maximization problem formulated for joint optimization UAV’s trajectory, active beamforming at UAV, passive STAR-RIS. This cutting-edge also subject flight safety, maximum duration constraint, as well GUs’ minimum data rate requirements. Given unknown locations obstacles prior flight, we provide online decision making framework employing reinforcement learning (RL) adjust trajectory beamformer. To enhance system’s robustness against associated uncertainties caused by limited sampling environment, “distributionally-robust” RL (DRRL) algorithm proposed offering adequate worst-case performance guarantee. Our numerical results unveil that: 1) assisted UAV communications benefit significant gain over reflecting-only RIS; 2) DRRL achieves stable robust than state-of-the-art algorithms.