Attention-based Phase Shift Control for Multi-RIS-aided Systems

To support extremely high data rates in 6G wireless networks, reconfigurable intelligent surface (RIS) assisted communications have gained much attention in recent years. By controlling the phase shift of each reflecting element, the RIS can proactively modify the wireless propagation environment, thereby improving the signal quality for mmWave/THz systems. One important problem of multi-RIS-aided systems is that accurate channel estimation is difficult due to the heavy pilot overhead. In this paper, we propose a deep learning (DL)-based multi-RIS codeword selection scheme that selects multi-RIS codewords maximizing the sum rate of multi-RIS-aided systems. Specifically, we exploit the attention technique to calculate the impact of each RIS on the UE and emphasize the important reflected channels for multi-RIS codeword selection. From the simulation results, we demonstrate that the proposed scheme outperforms the benchmark schemes by a large margin.