Channel Estimation for Reconfigurable Intelligent Surface-Assisted Cell-Free Communications

Abstract

Recent advancements in reconfigurable intelligent surface (RIS)-assisted cell-free systems have primarily focused on improving coverage and reducing network costs. However, much of the existing literature assumes perfect knowledge of channel state information (CSI), which poses significant challenges in practical implementations. This study investigates the channel estimation problem in RIS-assisted cell-free systems, highlighting two key observations: (1) a shared channel exists between the base station (BS) and the RIS across all users, and (2) a similar common channel exists between the RIS and the users across all BSs. Building on these insights, the paper addresses the challenges of cascaded and two-timescale channel estimation. Specifically, two novel methods are introduced: (1) a 3D-MMV-based compressive sensing technique for efficient cascaded channel estimation, and (2) a pilot reduction strategy that leverages multi-BS cooperation to enhance channel estimation performance. These methods aim to improve the accuracy and efficiency of channel estimation in RIS-assisted cell-free systems while minimizing pilot overhead.