A Study On Driver Cognitive Distraction Detection Based On The Driving Performance Measures

The Auto has played a huge positive role in the development of the society,but the frequent occurrences of traffic accidents have also done harm to the society.As a major participant in the traffic system.To study driver driver behaviour and driving style,this paper mainly completed the following works:(1)In this study,the questionnaire of private drivers is conducted via popular instant messengers such as We Chat and QQ.Exploratory factor analysis(EFA)based on a four-factor structure and the oblimin rotation is used to evaluate the feasibility of the driver behavioral questionnaire(DBQ)in China.First of all,the paper studies which behaviors would have an effect on traffic safety and then the influence of demographic variables(e.g.driving years).In addition,the study also o evaluate whether the EFA with target rotations was appropriate for China.Furthermore,the factor structure that is appropriated for China is put forward.The results show that some behaviors(e.g.dropping litter,aggressive character,drunk driving,inattention,etc.)and demographic variables(e.g.age,annual driving mileage,driving years)are heavily related to traffic safety.Meanwhile,the analysis results also show that the four-factor structure is suitable in China by EFA with oblimin rotation.(2)At the same time,we use multiple linear regression methods to explore whether gender,age,driving experience and annual mileage can be used to predict DBQ scale scores.It is found that gender,driving experience,annual mileage and age can be used as predictors to predict ordinary violations.Operational errors can only be predicted by annual driving mileage,etc.Gender can be used to predict aggressive violations.And lapses can be predicted jointly by annual driving mileage and age.(3)The principal component analysis method was used to reduce the dimensions of driver's style variable,it is found that the eigenvalue of the first principal component is 1.863,and the eigenvalues of the other six principal components are less than 1.In addition,the first principal component can explain the variance contribution of 44.26%.Therefore,thePCA score of first principal component of can be used to represent the overall score of the driving style.Secondly,a logistic regression model was used to explore the driving style and the demographic demographic variables can predict the driver's deduction of traffic rules in the previous year.We found that age,annual driving mileage,and driver style composite scores can be used as predictors to predict the driver's deductions for traffic rules in the previous year.More specifically,aggressive drivers report more traffic violations than mild drivers.