Remote Detection of Indoor Human Proximity using Bluetooth Low Energy Beacons

Abstract

The way people interact in daily life is a challenging phenomenon to capture and to study without altering the natural rhythm of interactions. Our work investigates the possibility of automatically detecting proximity among people, the first mandatory condition before a dyad starts interacting. We present Remote Detection of Human Proximity (ReD-HuP), an algorithm based on the analysis of Bluetooth Low Energy beacons emitted by commercial wearable tags. We validate ReD-HuP with real-world indoor settings and we compare its performance with respect to detailed ground truth data collected from a number of volunteers. Experimental results show an accuracy and F-Score metric up to 95%.