Two-Stage Channel Estimation for Hybrid RIS Assisted MIMO Systems

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