Secret key generation for reconfigurable intelligent surface-assisted MISO multi-user system

Abstract

Reconfigurable intelligent surface (RIS)-assisted physical layer key generation plays an important role in the eavesdropping problem on a typical multi-user scenario of wireless networks. Existing RIS-assisted physical key generation technology mainly solved the security issues in the single-antenna and multi-user scenario, which is based on the idealistic assumption of a spatially independent channel model at the RIS. It limits the degrees of freedom and neglects the existence of channel spatial correlation, which would possibly degrade the sum key generation rate (SKGR). To solve the above problem, this paper proposes a secret key generation scheme for RIS-assisted multiple-input single-output multi-user system. This scheme establishes a spatial correlation channel model, which comprehensively considers the influence of spatial correlation among antennas of the base station (BS), multi-user, and RIS elements on SKGR. Based on this, a key generation framework is designed to generate a more secure secret key with the maximum SKGR. Specifically, a closed-form SKGR expression and an optimization problem formulation are obtained to obtain the joint optimal coefficient of the BS and RIS. Then, the difference convex-successive convex approximation based alternating optimization algorithm is proposed to solve this optimization problem. The simulation results show that the proposed scheme has improved 10-bit/channel use compared with the randomization scheme of BS precoding and RIS reflection coefficient. It also has an improvement of 12-bit/channel use compared with the spatially independent scheme.