A Two-Stage RIS-Assisted Approach for Near-Field Multi-Target Localization

This letter presents a two-stage near-field multi-target localization approach assisted by a reconfigurable intelligent surface (RIS). By dynamically configuring the phase shifts, the RIS reflects sensing signals to a single-antenna receiver, constructing a multidimensional received signal. Our method first establishes a temporal-domain signal model to obtain pre-estimates using the 2D-MUSIC algorithm. These estimates are then refined by jointly optimizing the transmit and receive beamforming, with the goal of maximizing the sum sensing rate. Simulation results show that our approach achieves high distance resolution even when multiple targets are located in the same direction. In addition, the proposed beamforming optimization improves estimation accuracy by 2 to 3 orders of magnitude compared to the non-optimized case.