Design And Implementation Of A Single-channel EEG-based Fatigue Driving Monitoring System

According to the Ministry of Transportation statistics,fatigue driving is one of the main causes of serious traffic accidents,and traffic accidents caused by fatigue driving account for 20%-30% of all traffic accidents,and possibly even more.With the rapid development of the global trade and logistics industry,long-distance transportation is inevitable,and drivers specializing in logistics are prone to fatigue during long-term monotonous driving behavior,on the other hand,the growth in the number of private cars has also exacerbated such situations,and fatigue driving has caused irreparable damage to people’s lives and property.Among the many physiological indicators that can be used to assess driver fatigue levels,EEG signals have proven to be one of the most predictive and reliable indicators,but the disadvantages of conventional EEG devices,which are bulky and complex to operate,limit their application;therefore,the development of a wireless,portable,wearable fatigue driving recognition system is a key and urgent research topic.The main work of this paper is as follows.(1)First,this paper designs a multi-person,long-duration driving simulation experiment based on the Carolingian Sleepiness Scale(KSS)and the Fatigue Scale(FS-14)to collect EEG data from subjects in fatigue and alert driving states.The collected EEG data were labeled as fatigue or alertness based on the results of the KSS and FS-14 measurement scales before and after the experiment.(2)Based on the driving simulation experimental data,this paper uses concentration,relaxation and the power spectrum features of five EEG bands of theta,low alpha,high alpha,low beta and high beta as the training and testing data of the algorithm,and analyzes the performance of K-nearest neighbor,logistic regression and CART-based random forest,the comprehensive comparison results show that CART-based random forest is more effective for fatigue recognition.(3)Based on the results of driving simulation experiments,this paper designs a portable fatigue driving monitoring system based on TGAT EEG chip,which has the functions of localization,fatigue reminder and remote monitoring.Firstly,the system has a single-channel EEG acquisition module for collecting the forehead EEG signals of drivers;secondly,when the system detects fatigue driving,it uses the voice module and vibration module to give drivers a reminder,and also adds the global positioning function to realize remote positioning monitoring;then,the system uses the cellular communication module to transfer EEG data and positioning data to the cloud server for storage,and uses the deployed machine The system then uses a cellular communication module to transfer EEG data and positioning data to the cloud server for storage,uses the deployed machine learning algorithm to recognize fatigue on the uploaded data,and stores the recognition results for calling from the WeChat Mini Program and hardware side.Finally,the lightweight WeChat Mini Program is used as the remote monitoring end to provide fatigue prompting,device online detection and positioning information display functions.The comprehensive results of several experiments show that the fatigue driving recognition rate reaches 83.4%,and this fatigue driving monitoring system has a certain degree of practicality and reliability,realizing the monitoring and early warning of fatigue driving behavior.