Effect of Spatial Interference Correlation on the Performance of Maximum Ratio Combining

TL;DR

This paper analyzes how spatial interference correlation affects the performance of maximum ratio combining (MRC) in wireless networks, providing exact and bounded SIR distributions to improve accuracy over previous extreme assumptions.

Contribution

It derives the exact SIR distribution for two antennas and bounds for multiple antennas under spatially correlated interference, advancing understanding of MRC performance.

Findings

Full-correlation assumption overestimates outage probability by up to 30%.

No-correlation assumption underestimates outage probability significantly.

Derived bounds and exact distributions improve performance benchmarking.

Abstract

While the performance of maximum ratio combining (MRC) is well understood for a single isolated link, the same is not true in the presence of interference, which is typically correlated across antennas due to the common locations of interferers. For tractability, prior work focuses on the two extreme cases where the interference power across antennas is either assumed to be fully correlated or fully uncorrelated. In this paper, we address this shortcoming and characterize the performance of MRC in the presence of spatially-correlated interference across antennas. Modeling the interference field as a Poisson point process, we derive the exact distribution of the signal-to-interference ratio (SIR) for the case of two receive antennas, and upper and lower bounds for the general case. Using these results, we study the diversity behavior of MRC and characterize the critical density of simultaneous transmissions for a given outage constraint. The exact SIR distribution is also useful in benchmarking simpler correlation models. We show that the full-correlation assumption is considerably pessimistic (up to 30% higher outage probability for typical values) and the no-correlation assumption is significantly optimistic compared to the true performance.