Sensing-based Two-timescale Channel Estimation for RIS-assisted Hybrid Millimeter Wave Systems

Abstract

In this paper, we propose a sensing-based two-timescale channel estimation algorithm for reconfigurable intelligent surface (RIS) assisted multi-user hybrid millimeter wave systems. The proposed channel estimation algorithm aims to separately estimate the base station (BS)-RIS and RIS-user channels instead of estimating the cascaded BS-RIS-user channel with a limited number of radio frequency chains equipped at the BS. In particular, we first cooperatively acquire parameters of the BS-RIS channel, e.g., the line-of-sight component and the equivalent channel state information (CSI) of the BS-RIS channel, via sensing methods by both transmitting and receiving sensing signals at the BS once over the large timescale. Then, users transmit orthogonal training sequences to the BS while RIS elements are sequentially turned on for obtaining CSIs of time-varying RIS-user channels over the small timescale. Our analytical and simulation results show that the proposed channel estimation algorithm can effectively estimate RIS-user channels with hybrid architecture at the cost of a reasonable training overhead.