Secure Downlink Transmission in Massive MIMO System with Zero-Forcing Precoding

Abstract:
We analyze the ergodic secrecy rate of multi-cell downlink systems assuming that the number of base station (BS) antennas per cell, the number of mobile users, and the number of eavesdropper antennas are all growing without bound. Our analysis includes the effect of uplink channel estimation, pilot contamination, and path-loss. The BS employs zero-forcing (ZF) precoding and artificial noise (AN) generation for secrecy enhancement. Our analytical results reveal that in a secure massive multiple-input multiple-output (MIMO) system employing AN, ZF precoding outperforms matched-filter (MF) precoding, unless the signal power is very small, inter-cell interference is severe, or the effect of pilot contamination is dominant.