Sensor Data Driven Modeling and Control of Personalized Thermal Comfort Using Interval Type-2 Fuzzy Sets

Abstract

As people in different rooms usually have different thermal comfort feelings or demands, it is valuable to study the modeling and control of thermal comfort to meet the personalized requirements. This paper tries to solve this issue using the data collected by the temperature and humidity sensors in the working or living time periods in the room being studied. We firstly present a statistic method based sensor data preprocessing strategy to discard noisy data and obtain the reasonable intervals for the temperature and humidity of each day. Then, we construct the Gaussian interval type-2 fuzzy set models to depict the personalized temperature and humidity comfort through measuring the uncertainty degrees of the obtained intervals. At last, we propose a control scheme to realize the personalized thermal comfort regulation. Our results show that the constructed thermal comfort models can recommend a reasonable temperature and humidity range for the demand in a specific room.