Deployment Issues for Massive MIMO Systems

TL;DR

This paper investigates practical deployment challenges in massive MIMO systems, highlighting how spatial correlation, LOS, and antenna clustering affect interference and performance, and compares precoding techniques under imperfect CSI.

Contribution

It provides insights into how antenna clustering and precoding choices impact interference and system performance in real-world massive MIMO deployments.

Findings

Antenna clustering reduces spatial correlation and improves performance.

Zero forcing precoding outperforms matched filter in interference mitigation.

Imperfect CSI significantly increases noise in ZF precoding.

Abstract

In this paper we examine a number of deployment issues which arise from practical considerations in massive multiple-input-multiple-output (MIMO) systems. We show both spatial correlation and line-of-sight (LOS) introduce an interference component to the system which causes non-orthogonality between user channels. Distributing the antennas into multiple clusters is shown to reduce spatial correlation and improve performance. Furthermore, due to its ability to minimize interference, zero forcing (ZF) precoding performs well in massive MIMO systems compared to matched filter (MF) precoding which suffers large penalties. However, the noise component in the ZF signal-to-noise-ratio (SNR) increases significantly in the case of imperfect transmit channel state information (CSI).