---

The project will improve building energy efficiency and occupant comfort by providing technical contributions in four aspects:

1. Occupancy sensing and estimation: The project will use a depth sensor to detect number of occupants, their body structure, and other relevant parameters to determine their comfort level. It will consider many realistic scenarios in a commercial space, e.g., door opening and closing, people moving with backpacks/lunch boxes, cleaning workers moving with large drum of cleaning equipment, food caterers moving with box of foods, people carrying laptops in hands, people wearing caps, etc. It will investigate how to deal with noise in the depth data and how to co-ordinate among different depth sensors in order to cover larger entrance/exit areas. It will investigate what is the minimum frame rate and resolution to detect entrance and exit events while maintaining high accuracy, which will guide building a cheaper depth sensor for this application.




2. Collecting large volume of data from commercial spaces: Almost all of the people counting modeling and prediction techniques rely on simulation data instead of actual people count. This project will collect real data regarding number of occupants from one Bosch office and two CMU campus buildings for over a year by integrating it with existing infrastructure, e.g. Volttron, which will enable developing and testing of novel techniques for modeling and prediction of occupancy pattern.




3. Constructing models of occupancy patterns and occupant comfort: By collecting large volume of real-data, this project aims to build models for capturing and predicting occupancy patterns. It will also collect data to determine how body shapes can be useful for determining occupant comfort. In order to obtain the ground truth of occupant comfort, it will get subjective opinion from users (via smartphones) as well as use a thermal imager to capture the actual body temperature of the subjects. The end goal is to build individual and group comfort models that provide more occupant comfort than traditional PMV based techniques.





4. Developing novel control solution: This project will develop novel control techniques by incorporating the occupancy pattern and occupant comfort models into it. The control algorithm will be used to control HVAC system for at least 6 months in at least one building in order to show the true potential of saving energy and providing occupant comfort.