Secure beamforming and artificial noise design in interference networks with imperfect ECSI

This paper investigates the joint robust beamforming and artificial noise design in two-user interference networks, where one of the two pairs is wiretapped by an external eavesdropper, and only imperfect eavesdropper´s channel state information (ECSI) is available for the two pairs. Under individual transmission power constraint, our goal is to maximize the worst-case achievable secrecy rate of the wiretapped node and guarantee the signal-to-interference-plus-noise ratio (SINR) requirement at the wiretap-free node. In particular, considering the channel vectors and channel covariance matrices errors, we first model the imperfect ECSI in the sense of ellipsoidal uncertainties. In the sequel, our design goal is transformed into solving a non-convex optimization problem which can be further reformulated as a two-stage problem. With the help of rank relaxation and one dimensional line search, the considered two-stage problem is solved efficiently. Simulation results demonstrate the effectiveness of the proposed design in different scenarios.