Intelligent Reflecting Surface-based Smart Resource Management in Secure Wireless Communications
Abstract or Summary
Recently, the intelligent reflecting surface (IRS) has been considered a promising solution to address the channel blockage in wireless communication systems. However, the reflecting coefficients of IRS elements are unchanged over all communication channels, it requires a smart management mechanism to can securely support multiple users simultaneously. In this paper, we study the joint multi-IRS control and resource management to enhance the user secrecy rate. Our design aims to optimize the IRSs’ coefficients, the transmit powers, and channel allocation to minimize the maximum weighted secrecy rate subject to practical constraints on the required communication rate and orthogonal transmission. To tackle the mixed-integer non-linear programming (MINLP), we propose an alternating algorithm for determining the supoptimal of the underlying problem. In particular, the IRSs’ coefficients-related sub-problem and resource allocation sub-problem are iteratively solved until convergence. Furthermore, we also propose a deep neural network (DNN)-based frame-work to learn the initial point of the channel assignment in the optimization algorithm. Numerical studies confirm that the proposed design can significantly reduce the leakage ratio up to 0.1278.
