Thermal discomfort in the workplace: measurement through the combined use of wearable sensors and machine learning algorithms

Abstract

This study aims at evaluating the use of wearable sensors in the Industry 4.0 context to measure and assess the worker's thermal comfort, which impacts on the general well-being status and, consequently, on productivity and attention level conditions. An experimental protocol based on controlled environment was developed and tested on 14 volunteers using wearable sensors for the acquisition of multimodal physiological signals under different thermal conditions. Results show that the combined use of wearable sensors and Machine Learning (ML) algorithms allow to reach satisfying performance (prediction accuracy up to ≈ 76%) in classification between comfort/discomfort conditions, thus enabling to promptly intervene to optimize the subject's working conditions without interfering with working activities.