Achievable Rates and Training Overheads for a Measured LOS Massive MIMO Channel

TL;DR

This paper evaluates the achievable uplink sumrates and training overheads in a measured LOS massive MIMO scenario, highlighting the trade-offs between spatial multiplexing, channel de-coherence, and high-speed train environments.

Contribution

It provides measured data-driven predictions of sumrates and analyzes training overheads in high-speed LOS massive MIMO systems, including an OFDM case study.

Findings

25% sumrate loss with 4x increase in coherence interval

OFDM training overhead as low as 2% at 90m/s

Trade-off between multiplexing performance and channel de-coherence

Abstract

This paper presents achievable uplink (UL) sumrate predictions for a measured line-of-sight (LOS) massive multiple-input, multiple-output (MIMO) (MMIMO) scenario and illustrates the trade-off between spatial multiplexing performance and channel de-coherence rate for an increasing number of base station (BS) antennas. In addition, an orthogonal frequency division multiplexing (OFDM) case study is formed which considers the 90% coherence time to evaluate the impact of MMIMO channel training overheads in high-speed LOS scenarios. It is shown that whilst 25% of the achievable zero-forcing (ZF) sumrate is lost when the resounding interval is increased by a factor of 4, the OFDM training overheads for a 100-antenna MMIMO BS using an LTE-like physical layer could be as low as 2% for a terminal speed of 90m/s.