Digital interior: Intelligent microclimate systems for energy-efficient thermal comfort

Authors

M. Fisch - Gentherm GmbH

Summary

Humans regulate their core body temperatures within a narrow range, accomplished by autonomic and behavioural thermoregulation. Behavioural thermoregulation occurs when humans are dissatisfied with their current thermal environment. We investigated if we could detect a physiological signal that could predict when an occupant would initiate a change to vehicle temperature (i.e., behaviourally thermoregulate). Eleven participants underwent environmental temperature manipulations ranging from 40°C to 10°C, while having their heart rate (HR), forehead skin blood flow, and electrodermal activity (EDA) measured. We found that time domain HR variability (HRV) measures, EDA, and skin blood flow data were able to predict thermal preference correctly from 30-70% of the time in a controlled environment. Frequency domain HRV measures did not increase machine learning algorithm performance but combining all time domain HRV measures did. Further, including all time domain HRV measures, with skin blood flow and EDA resulted in the greatest performance. Therefore, we were successfully able to predict thermal preference in a controlled laboratory environment. This has positive implications for the vehicle cabin of the future where physiological variables are able to determine when an occupant is thermally uncomfortable to enact a change to the environment prior to the need for the occupant to behaviorally thermoregulate.