Full Space Source Sensing Method Assisted by STAR-RIS

Existing reconfigurable intelligent surfaces (RIS) assisted sensing methods fail to adequately consider mixed near-field (NF) and far-field (FF) environments. Consequently, this leads to a diminishment in sensing accuracy, particularly in large-scale RIS deployment scenarios. This paper proposes a full space source sensing method assisted by simultaneously transmitting and reflecting (STAR)-RIS. Specifically, a STAR-RIS-assisted unified NF and FF model is established. Subsequently, we reduce the model’s dimensionality by decoupling the transmission and reflection coefficients of STAR-RIS. Finally, a two-dimensional (2D) model-applicable newtonized orthogonal matching pursuit (NOMP) algorithm for STAR-RIS tailored to the mixed NF and FF scenario is proposed. Simulation results show that the proposed method can achieve full space source position sensing estimation with an accuracy of 1 mm.