Using Bluetooth Low Energy (BLE) Signal Strength Estimation to Facilitate Contact Tracing for COVID-19

TL;DR

This paper presents a BLE-based method for digital contact tracing to identify close contacts within 6 feet for at least 15 minutes, enhancing COVID-19 mitigation efforts.

Contribution

It introduces a data-driven approach for estimating proximity using BLE signal strength, improving detection accuracy with more samples and auxiliary data.

Findings

Larger sample sizes improve detection accuracy.

Auxiliary information enhances proximity detection.

BLE signal analysis can support contact tracing efforts.

Abstract

The process of contact tracing to reduce the spread of highly infectious and life-threatening diseases has traditionally been a primarily manual process managed by public health entities. This process becomes challenged when faced with a pandemic of the proportions of SARS-CoV2. Digital contact tracing has been proposed as way to augment manual contact tracing and lends itself to widely proliferated devices such as cell phones and wearables. This paper describes a method and analysis of determining whether two cell phones, carried by humans, were in persistent contact of no more than 6 feet over 15 minutes using Bluetooth Low Energy signals. The paper describes the approach to detecting these signals, as well as a data-driven performance analysis showing that larger numbers of samples coupled with privacy preserving auxiliary information improves detection performance.