Secure Communication in Relay-Assisted Massive MIMO Downlink

摘要

In this paper, secure transmission strategies for the relay-assisted massive multiple-input multiple-output (MIMO) downlink are investigated by using artificial noise (AN) generation and by exploiting the excess degrees-of-freedom through the random and null-space based precoders. The achievable rate expressions are derived for the estimated channel state information, and hence, the detrimental effects of channel estimation errors in designing precoders are quantified. Specifically, the achievable secrecy rate at the relayed user nodes is derived in closed-form, and thereby, the performance gap between the random and null-space based precoders is investigated. Our performance analysis reveals that the AN generation by using random precoders can be employed to design secure physical layer transmission strategies for the relay-assisted massive MIMO downlink systems. The AN generation can be useful in the finite antenna regime for guaranteeing physical layer security in the base-station-to-relay hop of the relay-assisted massive MIMO downlink. Although the AN generated at the base-station propagates to the relay-to-user hop via amplify-and-forward operation at the relay, sophisticated security provisions are needed to secure the end-to-end transmission when the relay is equipped with finitely many antennas. Consequently, the performance gap between the random and null-space based precoders gradually diminishes in the limit of infinitely many base-station antennas. Nevertheless, the achievable secrecy rate steadily vanishes with an increasing number of antennas at the passive eavesdropper.