Integrating occupants' voluntary thermal preference responses into personalized thermal control in office buildings

Abstract

A Bayesian modeling approach which allows incorporating voluntary feedback data (comfort-related responses), collected via participatory interfaces, along with requested feedback data, into a thermal preference learning framework. This is achieved by explicitly considering occupant participation, a type of behavior, in the model. Experiments with human subjects were conducted to collect thermal preference datasets, with both participatory and requested setups, which were used to train personalized thermal preference models. The proposed approach allows using the participatory setup without distorting the thermal preference predictive probabilities. In addition, we propose a concept of smart occupant feedback request algorithm, that determines whether and when to request feedback based on the quantified value of the request. This work will lead to smarter, user-interactive comfort delivery systems that will be continuously updated through interactions with their occupants, and will provide customized indoor environments tailored to individual preferences.