Study On Passenger Overall Comfort In High-speed Railway Environments Based On Eeg Signals

The key to regulating and optimizing the overall comfort of high-speed railway passengers is to analyze the mechanism of comfort deterioration and evaluate it accurately.However,the overall comfort of high-speed railway passengers is affected by vibration,noise,air pressure,light,temperature and passengers’ own emotions,etc.The interaction of the above factors brings difficulties in analyzing the deterioration mechanism and evaluating passenger comfort.As the "gold standard" of physiological signals,EEG signals have the advantage of accurately reflecting human physiological and psychological states.Therefore,this thesis takes high-speed railway passengers as the research object and carries out a study on passenger overall comfort in high-speed railway environments based on EEG signals.The main content are as follows:(1)Designed and carried out the high-speed railway occupant field test.The EEG signal data and subjective evaluations of 20 test passengers were collected in the Changsha-Guiyang section of the Shanghai-Kunming line.Correlation analysis and gender difference analysis were carried out on the comprehensive comfort level of high-speed train occupants and eight single factors influencing comfort level.(2)Revealed the potential comfort degradation mechanisms in highspeed railway environments based on EEG signals.The beta band associated with the comfort level was found.Based on the beta band,the functional brain networks of 20 subjects in the most comfortable and least comfortable states were constructed.The functional connectivity phenomenon of the brain was analyzed and the collaborative change process among the brain regions were revealed under different comfort levels.The sensitive brain regions were significantly activated in the beta band compared with those in the uncomfortable state.Based on the above neural characteristics,we found the mechanism of response to discomfort,including the perception of the surrounding environment,stimulation of negative emotions,and finally,body movement intention,which provided a neurological explanation for the overall comfort deterioration mechanism.(3)Proposed a comprehensive comfort evaluation model of highspeed train passenger.1779 features were extracted from the EEG signals after pre-processing.Various machine learning algorithms were used to construct a high-speed railway occupant comfort assessment model and to compare and choose the best,and finally the Light GBM(Light Gradient Boosting Machine,Light GBM)algorithm was used as the assessment model;To reduce data dimension and redundancy,two feature selection methods based on feature types and feature importance were proposed.The results showed that the feature selection method based on feature importance significantly improved the evaluation efficiency of the highspeed railway occupant comfort assessment model,with the RMSE(Root Mean Squared Error,RMSE)and MAE(Mean Absolute Error,MAE)values of 0.1704 and 0.1261,respectively.