Research Article
DRL-Based Jointly Trajectory Planning and Beamforming Optimization in STAR-RIS Assisted UAV-NOMA System
@INPROCEEDINGS{10.1007/978-3-032-33662-0_20, author={Xingyuan Lv and Qian Xu and Xin Yang and Haochen Liu and Yingtai Shao and Ling Wang}, title={DRL-Based Jointly Trajectory Planning and Beamforming Optimization in STAR-RIS Assisted UAV-NOMA System}, proceedings={Communications and Networking. 20th EAI International Conference, ChinaCom 2025, Xi’an, China, November 8--9, 2025, Proceedings}, proceedings_a={CHINACOM}, year={2026}, month={10}, keywords={Non-orthogonal multiple access (NOMA) Unmanned aerial vehicle (UAV) simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) Deep reinforcement learning (DRL)}, doi={10.1007/978-3-032-33662-0_20} }- Xingyuan Lv
Qian Xu
Xin Yang
Haochen Liu
Yingtai Shao
Ling Wang
Year: 2026
DRL-Based Jointly Trajectory Planning and Beamforming Optimization in STAR-RIS Assisted UAV-NOMA System
CHINACOM
Springer
DOI: 10.1007/978-3-032-33662-0_20
Abstract
This paper investigates the non-orthogonal multiple access (NOMA) communication system of unmanned aerial vehicle (UAV) assisted by simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) in the presence of imperfect channel state information (CSI). Specifically, the UAV serves as an aerial base stations (BS) to provide communication services for multiple ground users employing NOAM. Through the STAR-RIS, the transmitted signal can be transmitted to users, who are randomly distributed on both sides of the surface, providing full spatial coverage. In order to maximize the communication sum-rate, the UAV’s trajectory, beamforming, as well as the STAR-RIS coefficients are jointly optimized. At same time, the flight safety of the UAV, maximum flight time, and minimum communication rate for each user were also considered. Deep reinforcement learning (DRL) based online decision-making framework is proposed for dynamic scenes with imperfect CSI. The simulation results have verified that the algorithm outperforms other baseline schemes in terms of sum rate.