Research On Indoor Thermal Environment Control Method Based On EEG Thermal Comfort Discrimination

Thermal comfort is defined as the satisfaction degree of human with thermal environment.A large number of studies have reported that thermal comfort has an important influence on human health and work performance.The evaluation and measurement of thermal comfort has become one of the key links in the indoor environment control.In recent years,researches on using physiological parameters to measure thermal comfort have gradually increased.However,the current physiological parameters focus on peripheral system signals such as skin temperature and electrocardiogram.The characteristics of electroencephalogram(EEG)are little discussed,which can directly respond to changes of central nervous system activity and mental state.Therefore,this paper studies the monitoring of thermal comfort based on EEG and the control of indoor environment.First,this paper explores the feasibility of continuously determining individual thermal comfort by using EEG signals.Two indoor thermal environments were constructed respectively.The EEG signals in the two environments of 22 subjects were collected.Then the power spectral density characteristics of EEG were extracted for analysis.The ensemble learning method was used to build an individual thermal comfort discrimination model.The results showed that there were sigbificant differences in multiple frequency bands and channel locations.And when the length of detection window was 1 second,the discrimination accuracy was significantly higher than random guessing.In addition,the continuous multi-window ensemble learning method effectively improved the discrimination accuracy,and when the length of detection window was 1 minute,the average discrimination accuracy can be up to87.9%.The paper further constructs a brain-computer interface-air conditioner(BCI-AC)automatic control system.This system can continuously collect EEG and send EEG data to the computer.Then the computer analyzes EEG data and obtains the discrimination results by the discrimination model,achieving the automatic control of the air conditioner.Based on the BCI-AC automatic control system,the online thermal environment control experiments of 11 subjects were carried out.By the online control of BCI-AC system,thermal comfort level of indoor environment was obviously improved,and the average thermal comfort score was dropped from hot discomfort(2.45)to comfort(0.55).When subjects were covered(the clothing insulation increased by 1.1 clo),the BCI-AC system can automatically adjust the room temperature to a lower level for maintaining subjects' thermal comfort,showing a good adaptability.The research in this paper shows that the good feasibility of continuously determining individual thermal comfort and thermal environment automic control method based on EEG signals.It provides a certain foundation for the technical development of thermal environment control and optimization in intelligent buildings.