Channel Estimation for Near-Field XL-RIS-Aided mmWave Hybrid Beamforming Architectures

Abstract

Extremely large-scale reconfigurable intelligent surface (XL-RIS) has been a prospective capacity-enhancing technique. Faced with XL-RIS-aided millimeter wave (mmWave) communications, we propose a low-complexity channel estimation strategy with the near-field spherical wavefront assumption. It separates the whole channel estimation procedure into two phases to decrease pilot use and computational complexity: 1) estimation of angular-domain parameters of the base station (BS) and users equipped with the hybrid beamforming architecture, and 2) estimation of cascaded angular- and polar-domain parameters of the XL-RIS. Particularly, in the first and second phases, the three-dimensional multiple measurement vector compressive sensing (3D-M-CS) framework and the 3D distributed CS (3D-D-CS) framework are proposed for enhanced parameter estimation, respectively. Then the variants of Look Ahead Orthogonal Match Pursuit (LAOMP), i.e., 3D-M-LAOMP and 3D-D-LAOMP, are developed for solving our proposed CS frameworks.