THERMOSTAT

Developing Statistical Methods for Analyzing Time-Series Thermal Physiological Data to Increase Accuracy in Occupant Behavior and Thermal Comfort Modeling

Principal Investigators 

Prof. Thomas Auer,
Chair of Building Technology and Climate Responsive Design, TUM School of Engineering and Design

Prof. Mathias Drton,
Chair of Mathematical Statistics, Department of Mathematics, TUM School of Computation, Information and Technology

Abstract The THERMOSTAT project aims to establish a new standard for modeling thermal comfort and occupant behavior in buildings by integrating physiological experimentation with advanced causal inference methods. We will develop statistical tools capable of analyzing complex, multivariate, and time-series physiological data, uncovering causal relationships between thermal environments and human responses. The project will span building science and mathematical statistics via a collaboration between the chairs of building technology and climate responsive design (Auer) and mathematical statistics (Drton). Over four years, the team will conduct targeted thermal physiology experiments, apply and refine causal modeling frameworks, and create FAIR-compliant data management guidelines to ensure reproducibility. By developing models that capture interindividual variability, adaptation, and temporal dynamics, THERMOSTAT will enhance the precision of thermal comfort simulations and support the design of more energy-efficient and occupant-responsive buildings.