Correlation-Based Multi-Stage Channel Estimation for RIS-Assisted Multi-User mmWave Systems

This paper proposes a correlation-based multi-stage channel estimation strategy for reconfigurable intelligent surface (RIS)-assisted multi-user (MU) millimeter wave (mmWave) systems, in which the base station (BS), the RIS and the users are all equipped with a uniform planar array (UPA). Specifically, by exploiting the correlation of angle information during multiple consecutive channel coherence blocks, we estimate the full channel state information in the first coherence block (FCB) and only update the gain information in each subsequent block. Then, based on the ambiguity property, a low-pilot-overhead two-stage method is adopted in the FCB. In Stage 1, all users jointly transmit pilot to estimate the angles of arrival at the BS and the correlation factors between different paths of the common BS-RIS channel via introducing a set of matching matrices and a designed RIS phase shift matrix, which achieves a significant MU diversity gain. By exploiting the mmWave cascaded channels' ambiguity property, an ambiguous common BS-RIS channel is constructed with the estimated correlation factors. In Stage 2, each user independently transmits a limited number of pilots to estimate their individual ambiguous RIS-user channel, leading to a comprehensive cascaded channel estimation. For the remaining blocks, we adopt a straightforward least-squares method for rapid gain estimation. Simulation results confirm the method's effectiveness in providing accurate estimation with reduced pilot overhead, outperforming existing methods.