Signal Classification with Linear Phase Modulation for RIS-Assisted Near-Field Localization

Abstract

Reconfigurable intelligent surface (RIS), one core element in 6G, opens a new opportunity to design near-field (NF) localization since a signal's propagation distance can be different depending on the reflected points on large RIS. For the design of NF localization to be effective, signals reflected on distinct RIS points should be profiled without interfering with the others, called signal classification (SC). In this paper, we propose a simple yet novel SC technique, called linear phase modulation (LPM), where sequences of distant RIS elements' phases are linearly modulated with different rates. Along with the conventional or-thogonal frequency division multiplexing waveform, LPM makes it possible to classify the reflected signals by distant RIS elements as well as estimate their propagation distances using a two-dimensional Fourier transform (2D-FT) technique. It exploits one more signal dimension for performance improvement than conventional one-dimensional FT (1D-FT) based SC. Through analytic and numerical studies, we verify the effectiveness of the proposed SC using LPM by comparing its localization accuracy with several benchmarks designed based on 1D-FT.