Vision Transformer-Based Passive Beamforming for RIS-Assisted Multi-User Channels

In this paper, a geometrical cluster-based channel model is proposed for reconfigurable intelligent surface (RIS)-assisted multi-user communications, where line-of-sight (LoS) and non-LoS components are considered in BS-RIS and RIS-Rx links. Then, a novel solution method of RIS phase is presented based on vision transformer (ViT) network, which intuitively solves RIS phase by position of multiple users and reduces solution complexity. Based on the proposed model, the impacts of RIS on CIR magnitude and coverage probability are simulated and analyzed. The results show that RIS phase generated by ViT network can achieve fairly well solution and increase channel condition, which can provide a foundation for deploying and designing RIS-assisted multi-user systems.