Research On Fatigue Driving Detection Based On Time Domain And Time Interval

Road traffic accidents have become the leading cause of casualties,more than 40% traffic accidents have to do with fatigue driving.Fatigue driving detection is the main way to reduce traffic accidents caused by fatigue driving,which has very important significance to the traffic safety.This paper drives the experimental data of fatigue driving rules from the time and time interval of vehicle.The change of driving behavior characteristics is analyzed objectively by quantitative model,combing with the observer's subjective judgment of the driver's facial features to identify the driving state.The main work of this paper is as follows:First,descriptive statistical analysis of driving behavior data is conducted.In this paper,driving behavior index were analyzed using the descriptive statistics(the accessory monitors real-time subjective judgment)under the two states of normal and fatigue,floating in a sequential backward selection algorithm bent for screening the characteristic structure of representative samples,for fatigue driving behavior recognition research.Secondly,based on the information fusion method and grey interval clustering method,the fatigue driving behavior identification of real vehicle driving data is studied from the perspective of time domain.In view of the real vehicle driving data using weighted voting K nearest neighbor(KNN)method for single feature fusion after fatigue driving detection,decision makers and naive bayesian method is used for multiple driving behavior characteristics of feature fusion layer after fatigue driving detection.For some abnormal data or missing sample,interval grey number is put forward,using the grey interval clustering method for fatigue driving detection.Analysis shows that recognition effect of the bayesian method is slightly better than the weighted voting KNN algorithm,after the interval grey number filling grey interval clustering method recognition accuracy is high,missing data information can be used effectively.Finally,from the angle of time interval,based on the Grey-LogACD coupling model to study the relationship between the speed change duration and the level of fatigue.In this paper,the speed change duration of the concept to depict the speed change state,using the classical autoregressive conditional duration(ACD)model to fit fluctuation of the speed change duration.The fatigue factor is introduced to LogACD model.Considering the lasting period of speed change gray feature,using fractional order grey model to simulate the model residual error correction,Grey-LogACD coupling model is set up to study the relationship between speed change duration and the level of fatigue.The results show that the Grey-LogACD coupling model has better effect on the fluctuation fitting of the speed change duration than LogACD model,as the fatigue level of ascension,the mean of speed change duration will also become bigger and bigger.