Detecting Human Thermal Discomfort via Physiological Signals

Abstract

This paper provides a new approach to the automatic detection of thermal discomfort. We see this research as a step toward the development of an intelligent climate control system that does not require any explicit input from the users. We introduce a novel dataset that simulates different thermal comfort/discomfort levels and we provide a complete analysis of different physiological signals and their capability of discriminating between these levels. Our approach is successful in detecting the thermal sensation of human subjects and it is expected to enable innovative adaptive control scenarios for enclosed environments as well as a significant reduction in energy consumption.