Adaptive Thermal Comfort with RL and Biometric Sensing

Conventional Heating, Ventilation, and Air Conditioning (HVAC) systems typically rely on static temperature setpoints and generalized comfort models, which often fail to account for individual physiological differences. This leads to suboptimal thermal comfort for occupants and inefficient energy use. This research introduces a novel, two-stage approach to achieve personalized thermal comfort within buildings. The proposed system integrates multimodal sensing, combining physiological data from occupants with environmental data from the building. This rich data stream feeds into a reinforcement learning (RL) based decision-making engine. The RL agent learns individual thermal preferences over time, moving beyond population-level averages. By continuously adapting to an occupant's unique physiological responses and the surrounding environment, the system can provide more responsive and personalized thermal interventions. This approach promises to enhance occupant well-being while potentially improving the energy efficiency of building control strategies.