Positioning Error Bound Analysis for RIS-Aided Sensing Systems With Hardware Impairments

Abstract

Reconfigurable intelligent surfaces (RISs) can enhance wireless sensing and localization capabilities by reconfiguring the wireless propagation environment. However, most existing localization frameworks rely on idealized RIS models and fail to fully capture practical hardware non-idealities. This paper considers a semi-passive RIS-aided target localization system with a single-antenna transmitter and a RIS with reflecting and sensing elements. The considered system jointly incorporates transmitter hardware distortion, sensing-side hardware impairments, finite RIS reflection efficiency, and element-level random phase perturbations. Under the resulting Gaussian observation model, we derive the Fisher information matrix (FIM) via a reference-position covariance approximation, leading to matrixform expressions for the Cramér-Rao lower bound (CRLB) and the positioning error bound (PEB). Furthermore, we derive a semi-explicit PEB to analytically reveal the effects of key system parameters on positioning accuracy. Numerical results validate the accuracy of the adopted approximation and demonstrate the effects of transmit power, RIS reflection-array size, sensing-array size, and hardware impairments on localization performance.