Hybrid Reconfigurable Intelligent Surfaces-Assisted Near-Field Localization

Reconfigurable intelligent metasurface (RIS) assisted user localization has been a hot topic along with the development of the sixth generation wireless communications. The employment of large-scale antennas goes far beyond the improvement of spectral efficiency, resulting in near-field communications between users and the RIS. In this letter, we consider to employ the hybrid RIS (HRIS) which is capable of simultaneous reflection and reception in near-field user localization for the first time. With such a new architecture, we derive the Cramer-Rao lower bound (CRLB) for assessing its localization performance, and subsequently propose an automatic differentiation-based gradient descent approach to optimize the HRIS configuration. The theoretical analysis is further supported by numerical simulations, indicating the superiority of the HRIS over the conventional purely reflective RIS in near-field user localization.