Discriminating sensor activation in activity recognition within multi-occupancy environments based on nearby interaction

TL;DR

This paper introduces a model that distinguishes sensor activations for individual activity recognition in multi-occupancy settings by analyzing proximity interactions, enhancing the accuracy of human activity detection.

Contribution

It proposes a novel approach to discriminate sensor activations per inhabitant using proximity interaction data, simplifying multi-occupancy activity recognition.

Findings

Effective discrimination of sensor activation per individual

Reduction in complexity of multi-occupancy activity recognition

Successful case study with UWB and binary sensors

Abstract

This work presents a computer model to discriminate sensor activation in multi-occupancy environments based on proximity interaction. Current proximity-based and indoor location methods allow the estimation of the positions or areas where inhabitants carry out their daily human activities. The spatial-temporal relation between location and sensor activations is described in this work to generate a sensor interaction matrix for each inhabitant. This enables the use of classical HAR models to reduce the complexity of the multi-occupancy problem. A case study deployed with UWB and binary sensors is presented.