Hybrid-Field Channel Modeling and Spatial Correlation for RIS-Assisted THz Communications
Abstract
Reconfigurable intelligent surfaces (RISs) are regarded as a promising technology for terahertz (THz) communications. However, existing models often ignore coupled near- and far-field scattering. This letter establishes a near-field model for the RIS–user equipment (UE) link and a hybrid-field model for the base station (BS)–RIS link. Their spatial correlations are derived and used to develop a projected gradient ascent (PGA) algorithm for passive beamforming. Numerical results validate the expressions and show PGA achieves coverage close to semidefinite relaxation (SDR) and higher than geometric alignment and random phase shifts, while running approximately 593– $1222\times $ faster than SDR.