Robust secrecy beamforming for full-duplex two-way relay networks under imperfect channel state information

Consider a two-way amplify-and-forward (AF) relay network where two legitimate nodes communicate through a relay in the presence of an eavesdropper. Assuming full duplexity at the legitimate nodes and the relay, a robust artificial noise (AN)-aided AF scheme is proposed to maximize the worst-case sum secrecy rate under imperfect channel state information (CSI) of the eavesdropper. This robust sum secrecy rate maximization (SSRM) problem is formulated as a max-min semi-infinite problem and is tackled by the semidefinite relaxation (SDR) method. In particular, we first convert the max-min semi-infinite problem into a maximization problem with a finite number of constraints. Then, an efficient two-block alternating difference-of-concave (DC) programming approach is proposed to iteratively solve the SDR problem, with one of the blocks computed in closed form. In addition, a specific robust rank-one solution construction procedure is presented to extract a feasible solution for the original robust SSRM problem from the SDR solution. The efficacy of the proposed method is demonstrated by numerical simulations.