Fatigue Driving Diagnosis Based On ECG Signal

It is found that fatigue driving is one of the most important factors causing major road traffic accidents.By identifying the fatigue driving behavior of driver and early warning,the accident rate can be effectively reduced.Based on the analysis of the driver’s long-term driving ECG data,the paper obtains the driver’s ECG time-frequency domain index,and based on the multi-index fusion theory,the fatigue driving diagnosis algorithm is proposed to diagnose the driver’s fatigue.The main contents of this paper are as follows:This paper first introduces the significance and research status of fatigue driving diagnosis,organize and analyze the advantages and disadvantages of the research methods,theories and equipment of fatigue driving diagnosis at this stage.According to the research needs of this topic,carry out experimental program design,and introduces the experimental platform and experimental process;establishes the road network through UC-winroad for driving simulation experiments,uses the bracelet,Ergolab and other equipment for data collection,and based on the Stanford sleepiness scale Subjective division of fatigue driving levels.Secondly,based on the ECG data collected by the smart bracelet,the ECG signal processing algorithm and flow are proposed.Use wavelet transformation to remove the noise of ECG raw data,and then extrac the feature point by adaptive threshold method to obtain the initial RR interval.In addition,the excessive RR interval should also be repaired based on the mean value.In order to obtain the equal time interval RR interval,the patched data needs to be linearly interpolated,and finally to calculate the heart rate variability index recalculation standard.Then the heart rate variability time-frequency domain index is selected to analyze the individual differences of ECG indicators,and the fatigue driving diagnosis algorithm is proposed.The diagnostic algorithm uses principal component analysis to perform multi-indicator fusion and obtain a comprehensive indicator T;and use BP neural network to supplement the missing T value;the threshold of each fatigue state is obtained,the driver fatigue state is classified in the driving process.Experimental data shows that BP neural network supplementation results are consistent with data fluctuation trends.Finally,compare and analyze the bracelet,questionnaire and Ergolab data to clarify the effectiveness of ECG signal processing algorithm and fatigue driving diagnosis algorithm,Combine the video data of the subjects for comparative analysis,the fatigue diagnosis algorithm can effectively identify the awake and severe fatigue states.