Training Signal Design for Discriminatory Channel Estimation

A training-based channel estimation scheme is proposed in the paper to enable the quality-of-service discrimination between legitimate and non-legitimate receivers in wireless networks. This method has applications ranging from user discrimination in wireless TV broadcast systems to the prevention of eavesdropping in secret communications. Specifically, by considering a network that consists of a multiple-antenna transmitter and two single-antenna receivers (i.e., the legitimate and non-legitimate receivers), we propose a multi-stage training-based channel estimation scheme that minimizes the normalized mean-squared error of the channel estimate at the legitimate receiver subject to a constraint on the estimation performance attainable by the non-legitimate receiver. The key idea is to exploit channel feedback from the legitimate receiver at the beginning of each stage to enable the use of artificial noise in the training signal, which allows us to effectively degrade the channel estimation performance at the non-legitimate receiver. The channel estimate obtained by the legitimate receiver in earlier stages are restricted due to constraints on the performance of the non-legitimate receiver, but may improve rapidly in later stages owing to the help of artificial noise and more accurate knowledge of the legitimate receiver's channel. Simulation results are presented to demonstrate the efficacy of the proposed channel estimation scheme.