Prediction of Rare Channel Conditions using Bayesian Statistics and Extreme Value Theory

TL;DR

This paper introduces a Bayesian and Extreme Value Theory-based framework for efficiently predicting rare wireless channel conditions, validated with both simulated and real data, improving throughput and reducing measurement needs.

Contribution

It presents a novel, sample-efficient approach combining radio maps, non-parametric models, and EVT for predicting rare channel events in wireless communication.

Findings

Bayesian formulation outperforms baselines in throughput.

EVT-based models are more accurate for rare events.

Framework reduces measurement requirements for reliability.

Abstract

Estimating the probability of rare channel conditions is a central challenge in ultra-reliable wireless communication, where random events, such as deep fades, can cause sudden variations in the channel quality. This paper proposes a sample-efficient framework for predicting the statistics of such events by utilizing spatial dependency between channel measurements acquired from various locations. The proposed framework combines radio maps with non-parametric models and extreme value theory (EVT) to estimate rare-event channel statistics under a Bayesian formulation. The framework can be applied to a wide range of problems in wireless communication and is exemplified by rate selection in ultra-reliable communications. Notably, besides simulated data, the proposed framework is also validated with experimental measurements. The results in both cases show that the Bayesian formulation provides significantly better results in terms of throughput compared to baselines that do not leverage measurements from surrounding locations. It is also observed that the models based on EVT are generally more accurate in predicting rare-event statistics than non-parametric models, especially when only a limited number of channel samples are available. Overall, the proposed methods can significantly reduce the number of measurements required to predict rare channel conditions and guarantee reliability.