Application of Artificial Intelligence to Thermal Comfort Sensor App for New-Generation Smart HVAC Systems
Project Description
Most HVAC systems only use temperature sensors to control actuators like fan coil units in classrooms and offices at HKUST. However, human thermal comfort depends on many parameters, such as temperature, humidity, air speed, heat radiation, human activity level, and clothing.

The Predicted Mean Vote (PMV) is a common thermal comfort index used to predict statistical thermal comfort responses based on a seven-point thermal sensation scale according to ASHRAE Standard 55. The PMV model was established through experiments involving over 1,000 human subjects in the United States and Europe. The goals of this project are to: Conduct human subject experiments to collect thermal sensation votes from occupants at HKUST, Measure relevant environmental parameters (temperature, humidity, air speed) during the experiments, Analyze the data and compare with the standard PMV model
Propose modifications to the PMV model based on the local experimental data, Implement the modified PMV model in an HVAC controller to improve thermal comfort at HKUST, The outcomes will be a customized PMV model that more accurately predicts thermal comfort for occupants at HKUST, as well as an HVAC controller that uses the modified PMV model to achieve better thermal comfort performance.
LEE Yi-Kuen
Course type
Applicant's Roles
This project is to apply the open-source deep learning algorithm and the CMOS image sensor of smartphone to monitoring the user’s activity/metabolic rate which is one important parameter to enhance the accuracy of thermal comfort sensing. The App will be connected to an existing IoT server at HKUST for statistical analysis of human thermal comfort to improve the PMV model for Hong Kong’s environment.
Applicant's Learning Objectives
Develop basic understanding of commercial HVAC systems, human thermal comfort, PMV sensors

Hands-on experience to develop a smartphone App

Collaboration with HKUST-MIT Smart EeB project team working on the same project.
Complexity of the project