Uplink Sensing and Localization in RIS-Assisted MISO-OFDM Systems via Newton-Refined MLE

Reconfigurable Intelligent Surfaces (RIS) are increasingly recognized as a promising enabler for integrated sensing and communication (ISAC) in next-generation wireless networks. This paper investigates an uplink sensing framework in a RIS-assisted MISO-OFDM system, where a multi-antenna base station (BS) receives pilot signals transmitted by a single-antenna user equipment (UE) through RIS reflections. The objective is to estimate the geometric properties of the UE-toRIS channel, including the angles of arrival (AoAs) and path delays, for UE localization and environmental sensing. A two-stage channel estimation procedure is proposed. Initially, coarse AoA estimation is performed via Simultaneous Orthogonal Matching Pursuit (SOMP). Subsequently, a refinement step is carried out using Maximum Likelihood Estimation (MLE) with Newton-based iterative updates. The UE location is estimated by utilizing the geometric properties of the RIS-UE line-of-sight (LoS) path obtained through channel estimation. Non-line-of-sight (NLoS) paths are also used for passive scatterer sensing. Simulation results demonstrate that the proposed method achieves high estimation accuracy in both angle and delay, leading to precise UE localization and environmental reconstruction.