Leakage Rate Analysis for Artificial Noise Assisted Massive MIMO With Non-Coherent Passive Eavesdropper in Block-Fading

Massive MIMO is one of the salient techniques for achieving high spectral efficiency in the next-generation wireless networks. Recently, a combined strategy of the massive MIMO and the artificial noise (AN), namely, AN assisted massive (ANAM) MIMO has been investigated for security enhancement. However, most of the recent studies have been built upon the assumption of perfect channel state information (CSI) at the eavesdropper (ED), the results of which may be too pessimistic in terms of security because there exists no ED-friendly downlink training from the base station in the practical ANAM systems. In this paper, we provide more sophisticated analysis on the secrecy performance of the ANAM systems assuming that the CSI of the ED channels are unknown to both the base station and the ED or partially known to the ED. We measure the secrecy in terms of the leakage rate to the ED (or the secrecy rate to the legitimate users) and characterize their upper and lower bounds in the high signal-to-noise ratio regime. Our analytical result is represented by a single compact expression as a function of the number of ED antennas, the dimensionality of signal space, and the channel coherence time, and thus offers useful insights that help us fully exploit the secrecy potential of the ANAM systems. Finally, several useful observations are made from the numerical examples.