Cell-Edge-Aware Precoding for Downlink Massive MIMO Cellular Networks

TL;DR

This paper introduces a cell-edge-aware precoding technique for massive MIMO networks that significantly improves user data rates and coverage by effectively suppressing inter-cell interference, especially at the network edges.

Contribution

The paper proposes a novel cell-edge-aware zero forcing precoder that leverages excess antennas to enhance performance over conventional methods in massive MIMO systems.

Findings

CEA-ZF outperforms CEU-ZF in data rate and coverage.

Outage probability decays faster with CEA-ZF as number of antennas increases.

Proper scheduling and pilot contamination control are crucial for optimal performance.

Abstract

We propose a cell-edge-aware (CEA) zero forcing (ZF) precoder that exploits the excess spatial degrees of freedom provided by a large number of base station (BS) antennas to suppress inter-cell interference at the most vulnerable user equipments (UEs). We evaluate the downlink performance of CEA-ZF, as well as that of a conventional cell-edge-unaware (CEU) ZF precoder in a network with random base station topology. Our analysis and simulations show that the proposed CEA-ZF precoder outperforms CEU-ZF precoding in terms of (i) aggregate per-cell data rate, (ii) coverage probability, and (iii) 95%-likely, or edge user, rate. In particular, when both perfect channel state information and a large number of antennas N are available at the BSs, we demonstrate that the outage probability under CEA-ZF and CEU-ZF decay as 1/N^2 and 1/N, respectively. This result identifies CEA-ZF as a more effective precoding scheme for massive MIMO cellular networks. Our framework also reveals the importance of scheduling the optimal number of UEs per BS, and confirms the necessity to control the amount of pilot contamination received during the channel estimation phase.