Center for Research in Energy Systems Transformation (CREST)

Personal comfort models predict an individual’s thermal comfort instead of the average response for a large population. We attempted to develop personal comfort models for car drivers using data collected from 10 cars while driving for approximately 2,000 hr. We measured conditions collected by the CAN-bus (Controller Area Network), a data acquisition system that is present in most of the modern cars. Data includes information about the in-vehicle thermal conditions, the surrounding environment, the status of the Heating, Ventilation, and Air Conditioning (HVAC) system, and the behavior of the occupant. The objective of the study is to assess the feasibility of inferring occupant’s thermal preference from the data available already available in most cars. By selecting and filtering all the available signals that are relevant for comfort, in this study we map the user actions of turning on/off their seat heating and correlate them to the vehicle indoor and outdoor conditions. The presented study provides the basis for using a machine learning automated process for thermal self-regulating HVAC system with the aim to improve comfort conditions and safety.