Productivity prediction via human physiological signals for an optimum thermal environment

Abstract

This study aims to understand the relationship between indoor temperature, physiological signals, thermal sensation, and productivity and to estimate the occupant’s productivity. A series of human experiments were conducted with 48 participants, and local skin temperatures, heart rate, and thermal sensation data were collected in 6 temperature conditions. OSPAN (Operation Span Task) was used to measure the occupant’s productivity and the LightGBM algorithm was used to generate a predictive model. The result verified that there is a significant correlation between certain local body skin temperatures and the occupant’s productivity, and the overall thermal sensation between high and low performing groups was significantly different by gender and BMI groups. The result suggested gender, BMI, and two local skin temperatures as effective factors to predict the occupant’s productivity.