Joint Training and Beamforming Design for Performance Discrimination Using Artificial Noise

Abstract

Recently, in multi-antenna wireless systems, the use of artificial noise (AN) in training and data transmission phases has been respectively proposed to achieve performance discrimination between a legitimate receiver (LR) and an unauthorized receiver (UR). For data transmission, an AN-aided beamforming (ANBF) scheme has been proposed where the message is sent towards LR using beamforming while AN is imposed in the null space of LR's channel to disrupt UR's reception. For channel estimation, the so-called discriminatory channel estimation (DCE) scheme has been proposed where a multi-stage training scheme is employed and AN is imposed in the null space of the estimated LR's channel obtained in previous stages to degrade the channel estimation performance of UR. In this work, the optimal power allocation between DCE and ANBF (as well as AN in both phases) is derived with the goal of maximizing the receive signal-to-interference-plus-noise ratio (SINR) of LR subject to a constraint on the maximum achievable SINR of UR. The simulation results show that, with the joint power allocation of DCE and ANBF, the SINR at LR and UR can be effectively discriminated even when UR is equipped with more antennas than the transmitter. Moreover, it is observed that the proposed joint DCE and ANBF scheme would allocate more power to the channel estimation phase compared with that using conventional channel estimation (without considering URs) in the training phase.