Local Diversity and Ultra-Reliable Antenna Arrays

TL;DR

This paper investigates the limitations of traditional diversity metrics in ultra-reliable low-latency communication systems with large antenna arrays, revealing that classic diversity overestimates or underestimates reliability in different regimes.

Contribution

It introduces the concept of local diversity to accurately characterize outage probabilities in large-scale antenna systems under ultra-reliable conditions.

Findings

Classic diversity overestimates the slope for weak deterministic components.

Classic diversity underestimates the slope for strong deterministic components.

Local diversity provides a more accurate asymptotic characterization.

Abstract

Ultra-reliable low-latency communication enables new use cases for mobile radio networks. The ultra-reliability (UR) regime covers outage probabilities between $10^{-9}$ and $10^{-5}$, obtained under stringent latency requirements. Characterisation of the UR-relevant statistics is difficult due to the rare nature of outage events, but diversity defines the asymptotic behaviour of the small-scale fading distributions' lower tail. The UR-relevant regime in large-scale antenna systems behaves differently from the tail. The generalising local diversity at a certain outage probability shows this difference clearly. For more than four independent antenna elements, the classic diversity overestimates and underestimates the slope of the cumulative density function for weak and strong deterministic channel components, respectively.