Data-driven Comfort Models for User-centric Predictive Control in Smart Buildings: Poster Abstract

Occupant comfort and energy cost are the dual, explicit or implicit, objectives of advanced control strategies implemented in smart buildings. Focusing on the former, we present a novel approach for data-driven modeling of user comfort and a path for integrating the resulting models in a generic predictive framework for optimal HVAC control. This offers significant improvement potential over the static scheduling or occupancy based temperature bounds.