Joint Beamforming Design for RIS-Assisted High-Speed Railway ISAC System: A Deep Reinforcement Learning Method

In the era of extensive deployment of high-speed railway (HSR) and rapid development of wireless communication networks, it is of vital importance to provide higher quality of service (Qos) for the communication users in the HSR network. Among the two most promising technologies in 6G networks: integrated sensing and communication (ISAC) and reconfigurable intelligent surface (RIS), we have considered an RIS-assisted HSR ISAC system. This system takes advantage of the RIS that it can provide virtual line-of-sight links for transmission signals, thereby significantly enhancing the transmission rate for users in the HSR ISAC system. We jointly optimized the beamforming vector of the ISAC base station (BS) and the reflection of RIS to maximize the sum rate of the HSR communication users while meeting power constraints and sensing quality. To address this non-convex optimization problem, we develop a new algorithm based on deep reinforcement learning (DRL) to solve it. Finally, the feasibility and efficiency of the algorithm in this system were verified through simulation.