Space-Constrained Arrays for Massive MIMO

TL;DR

This paper compares different antenna topologies for massive MIMO systems under space constraints, deriving closed-form interference metrics and validating them through SINR simulations, highlighting the importance of angular resolution in constrained environments.

Contribution

It provides new closed-form normalized mean interference equations for various antenna topologies in space-constrained massive MIMO systems and validates these with system SINR results.

Findings

Wider azimuth footprints improve performance in unconstrained setups.

Angular resolution is critical for performance in space-constrained environments.

Behavioral patterns predicted by interference metrics are confirmed by SINR simulations.

Abstract

We analyse the behaviour of a massive multi-user MIMO (MU-MIMO) system comprising a base station (BS) equipped with one of five different antenna topologies for which the spatial aperture is either unconstrained, or space-constrained. We derive the normalized mean interference (NMI) with a ray-based channel model, as a metric for topology comparison in each of the two cases. Based on the derivation for a horizontal uniform rectangular array (HURA) in [1], we provide closed-form NMI equations for the uniform linear array (ULA) and uniform circular array (UCirA). We then derive the same for a vertical URA (VURA) and uniform cylindrical array (UCylA). Results for the commonly-considered unconstrained case confirm the prior understanding that topologies with wider azimuth footprints aid performance. However, in the space-constrained case performance is dictated by the angular resolution afforded by the topology, particularly in elevation. We confirm the behavioural patterns predicted by the NMI by observing the same patterns in the system SINR with minimum mean-squared error (MMSE) processing.