Artificial-noise aided secure transmission over visible light communication system under coexistent passive and active eavesdroppers

Abstract

In this paper, secure transmission over multiple-input single-output visible light communication system under coexistent passive and active eavesdroppers (Eves) is studied. To enhance the achievable secrecy rate of this system given statistical channel state information (CSI) error model for the passive Eves-related channels, a robust artificial-noise (AN) based transmit strategy is proposed and a secrecy rate maximization problem subject to secrecy-outage-probability constraint, sum power constraint, and peak amplitude constraint is formulated. To solve this non-convex problem, a conservative approximation is presented to replace the probabilistic constraint and unbounded CSI error with worst-case secrecy constraints and spherically-bounded CSI errors, respectively. Then, semi-definite relaxation, S -procedure, and a Golden search-based algorithm are proposed to solve the approximated problem with fast convergence and low complexity. Simulation results show that the proposed method outperforms the other AN-aided and non-AN method for the coexistent active and passive Eves case, especially when the power budget is high.