Reconfigurable intelligent surfaces (RISs) have been proposed as a key enabler to improve the coverage of the signals and mitigate the frequent blockages in millimeter wave (mmWave) multiple-input multiple-output (MIMO) communications. However, the channel state information (CSI) acquisition is one of the major challenges for the practical deployment of the RIS. The passive RIS without any baseband processing capabilities brings difficulty on the channel estimation (CE), since the individual channels or the cascaded one can be estimated only at base station (BS) via uplink training or mobile station (MS) via downlink training. In order to facilitate the CSI acquisition, we focus on the hybrid RIS architecture, where a small number of elements are active and able to receive and process the pilot signals at the RIS. The CE is performed in two stages by following the atomic norm minimization to recover the channel parameters, i.e., angles of departure (AoDs), angles of arrival (AoAs), and propagation path gains. Simulation results show that the proposed scheme can outperform the passive RIS CE under the same training overhead. Furthermore, we also study the theoretical performance limits in terms of mean square error (MSE) via Cramér-Rao lower bound (CRLB) analyses.

中文翻译：

---

混合 RIS 辅助 MIMO 系统的两阶段信道估计

已提出可重构智能表面 (RIS) 作为提高信号覆盖范围和减轻毫米波 (mmWave) 多输入多输出 (MIMO) 通信中频繁阻塞的关键推动力。然而，信道状态信息（CSI）的获取是 RIS 实际部署的主要挑战之一。没有任何基带处理能力的无源RIS给信道估计（CE）带来了困难，因为单个信道或级联信道只能在基站（BS）通过上行链路训练或移动台（MS）通过下行链路训练进行估计。为了促进 CSI 采集，我们专注于混合 RIS 架构，其中少量元素处于活动状态，能够在 RIS 接收和处理导频信号。CE 分两个阶段执行，通过遵循原子范数最小化来恢复信道参数，即出发角 (AoD)、到达角 (AoA) 和传播路径增益。仿真结果表明，在相同的训练开销下，所提出的方案可以优于被动 RIS CE。此外，我们还通过 Cramér-Rao 下限 (CRLB) 分析研究了均方误差 (MSE) 方面的理论性能限制。