Reconfigurable intelligent surface (RIS) provides a new electromagnetic response control solution, which can proactively reshape the characteristics of wireless channel environments. In RIS-assisted communication systems, the acquisition of channel state information (CSI) and the optimization of reflecting coefficients constitute major design challenges. To address these issues, codebook-based solutions have been developed recently, which, however, are mostly environment-agnostic. In this paper, a novel environment-aware codebook protocol is proposed, which can significantly reduce both pilot overhead and computational complexity, while maintaining expected communication performance. Specifically, first of all, a channel training framework is introduced to divide the training phase into several blocks. In each block, we directly estimate the composite end-to-end channel and focus only on the transmit beamforming. Second, we propose an environment-aware codebook generation scheme, which first generates a group of channels based on statistical CSI, and then obtains their corresponding RIS configuration by utilizing the alternating optimization (AO) method offline. In each online training block, the RIS is configured based on the corresponding codeword in the environment-aware codebook, and the optimal codeword resulting in the highest sum rate is adopted for assisting in the downlink data transmission. Third, we analyze the theoretical performance of the environment-aware codebook-based protocol taking into account the channel estimation errors. Finally, numerical simulations are provided to verify our theoretical analysis and the performance of the proposed scheme. In particular, the simulation results demonstrate that our protocol is more competitive than conventional environment-agnostic codebooks.

翻译：可重构智能表面（RIS）提供了一种新的电磁响应控制解决方案，能够主动重塑无线信道环境的特性。在RIS辅助的通信系统中，信道状态信息（CSI）的获取与反射系数的优化构成了主要的设计挑战。针对这些问题，近年来发展了基于码本的解决方案，然而这些方案大多与环境无关。本文提出了一种新颖的环境感知码本协议，该协议能在保持预期通信性能的同时，显著降低导频开销与计算复杂度。具体而言，首先，引入一种信道训练框架，将训练阶段划分为若干区块。在每个区块中，我们直接估计复合端到端信道，并仅专注于发射波束成形。其次，我们提出一种环境感知码本生成方案，该方案首先基于统计CSI生成一组信道，随后利用交替优化（AO）方法离线获取其对应的RIS配置。在每个在线训练区块中，RIS根据环境感知码本中的相应码字进行配置，并采用能产生最高和速率的优化码字来辅助下行数据传输。第三，我们在考虑信道估计误差的情况下，分析了基于环境感知码本协议的理论性能。最后，通过数值仿真验证了我们的理论分析及所提方案的性能。特别地，仿真结果表明，我们的协议相较于传统的环境无关码本更具竞争力。