Two-Timescale Channel Estimation for RIS-Aided Near-Field Communications

In this paper, we investigate the channel estimation problem in reconfigurable intelligent surface (RIS)-aided near-field communication systems, where the extremely large number of RIS elements imposes considerable pilot overhead and computational complexity. To address this, we employ a two-timescale channel estimation strategy that exploits the asymmetric coherence times of the RIS-base station (BS) channel and the User-RIS channel. We derive a time-scaling property indicating that for any two effective channels within the longer coherence time, one effective channel can be represented as the product of a vector, referred to as the small-timescale effective channel, and the other effective channel. By utilizing the estimated effective channel along with processed observations from our piecewise beam training, we present an efficient method for estimating subsequent small-timescale effective channels. We theoretically establish the efficacy of the proposed RIS design and demonstrate, through simulations, the superiority of our channel estimation method in terms of pilot overhead and computational complexity compared to existing methods across various realistic channel models.