Smart Jamming Using Reconfigurable Intelligent Surface: Asymptotic Analysis and Optimization

We take the viewpoint of wireless attackers, and investigate the use of the reconfigurable intelligent surface (RIS) in degrading the communications performance of a time-division duplex system, in which a multiple-antenna base station (BS) transmits independent data streams to multiple user terminals (UTs). Each channel coherent time block consists of a channel training (CT) phase followed by a data transmission (DT) phase. During the CT phase, the UTs broadcast pilots to enable the BS to estimate the wireless channels, and the RIS manipulates the wireless environment so that the channel estimations obtained by the BS are incorrect. During the DT phase, the BS generates beamforming vectors, which are based on the channel estimations obtained during the CT phase, to transmit data to the UTs, and the RIS adopts a randomly time-varying reflection pattern to distort the signals received by the UTs. The mean square errors (MSEs) of the UTs are used as the performance metric, for which, analytical expressions are derived in large system limit. Based on the theoretic results, an efficient method is proposed to optimize the time-varying reflection pattern of the RIS to enhance the attack performance. Numerical simulations are presented to validate our theoretical results and to demonstrate the superiority of the proposed attack scheme over an existing attack scheme wherein the reflection pattern is time-invariant.