Learn to Beamform in Reconfigurable Intelligent Surface Aided MISO Communications with Channel Aging

Doppler-shift-induced channel aging effect significantly erodes the system performance due to the channel mismatch that evolves with time. This paper aims to investigate the joint beamforming strategy in reconfigurable intelligent surface (RIS)-aided high-mobility communications with channel aging effect. Specifically, a novel frame structure is first proposed for alleviating the heavy signaling overhead in the RIS-aided high-mobility scenario. Furthermore, a deep reinforcement learning (DRL)-based algorithm is devised for rapidly co-designing the precoder at the base station (BS) and the passive beamforming at the RIS relying exclusively upon partial channel state information (CSI) and real-time environment feedback, instead of only employing the outdated estimated CSI. The proposed joint beamforming scheme is capable of adapting to the dynamic propagation environment by exploiting the channel correlation between consecutive instants. Finally, simulation results demonstrate that the proposed algorithm can effectively mitigate the performance degradation caused by channel aging while being computationally efficient and outperforms several benchmark schemes.