On optimal artificial-noise assisted secure beamforming for the multiple-input multiple-output fading eavesdropper channel

In this paper we consider secure transmission for the multiple-input, multiple-output, single-antenna-eavesdropper systems with perfect information of the main channel and only the statistics of the eavesdropper's channel state information known at the transmitter. We adopt the celebrated artificial-noise (AN) assisted beamforming, which was studied by Goel and Negi with the limitation that the AN must be allocated in the null space of the main channel. By removing this limitation, we study the rate optimization problems over the covariance matrices of the signal and the AN. After exploring the structure of the optimal solutions, we show that the original optimization problem can be simplified as a much simpler power allocation problem. Our results show that to improve the rate performance, one may allocate the power of the AN in the directions of the right singular vectors of the main channel. Thus Goel and Negi's AN selection is strictly sub-optimal. Moreover, our simulations show that when the main channel has no null space, the rate performance of our optimized signaling outperforms that of Goel and Negi's significantly.