Enhancing PHY-Layer Authentication in RIS-Assisted IoT Systems With Cascaded Channel Features

This paper investigates the use of physical (PHY)-layer characteristics of wireless channels for identity authentication in Reconfigurable Intelligent Surface (RIS)-assisted wireless communication systems. We propose a generalized PHY-layer authentication scheme that involves both direct and cascaded channels for RIS-assisted IoT systems. The channels in each coherence block of the authentication phase are estimated using least squares estimation, and the estimation error covariance matrices are determined for performance evaluation. Leveraging the extracted features of the direct and cascaded channels, we establish authentication verification as a problem of differentiating two classes of changes. Auxiliary information in the form of second-order statistics assists the authentication verification procedure. Through statistical derivation using tools from statistical signal processing, matrix analysis, and composite hypothesis testing, we characterize false alarm and detection probabilities of the proposed scheme. Finally, we provide extensive numerical results to evaluate the performance and demonstrate the efficiency of the proposed authentication scheme across various system parameters.