Near-field RIS-aided Localization Under Channel Non-Stationarity: A Mismatched Model Approach

Abstract

With the increasing carrier frequencies and reconfigurable intelligent surface (RIS) apertures in future sixth-generation (6G) communication systems, the near-field region expands, causing more pronounced wavefront curvature and channel non-stationarity. By exploiting wavefront curvature-enabled localization with passive RIS, we propose a virtual sub-RIS-based channel model to approximate the true RIS channel. Based on this mismatched model, we propose and evaluate an iterative solution for localizing a user using narrowband pilots. Our numerical results at 28 GHz demonstrate clearly improved performance compared to the state-of-the-art reference method.