Research On Distracted Driving At Intersections Based On Observation

With the popularization of motor vehicles and the development of mobile communications,more and more factors induce the occurrence of distracted driving accidents.Therefore,distracted driving gradually attracts the attention of domestic and foreign research institutions and scholars.This article mainly uses the on-site observation method to collect the distracted driving behavior information of urban signalized intersections,and establishes a distracted driving regression and discriminant model to provide supporting materials for distracted driving research and future improvement of traffic safety at intersections.This article first conducts a domestic and international research review from the four aspects of distracted driving performance,research methods,the effects of distracted accidents,and distracted driving management,and then describes the driver’s distracted driving behavior theory,explains the causes of distracted driving,and lists common distracted driving behavior and its category,as the research object of this field observation.First,select Chengdu as the observation location,collect data on drivers,vehicles,intersections and traffic organization environment at signal intersections in different areas,and calculate the ratio of distracted driving behavior at the intersection is 15.3%,of which the highest frequency behavior is in order of talking with passengers(7.7%),smoking in the car(2.3%)and mobile phone conversation(1.5%),and the~2 test results indicate the driver’s age,presence or absence of passengers,vehicle type,operation status,and intersection entrance factors such as the number of lanes,the speed limit of the intersection,the time period and the traffic conditions have a significant impact on the distracted driving.Then build a Logistic regression model to perform regression analysis on the distracted driving of intersection drivers.The results show that:young drivers,no passengers in the car,parking at intersections,morning peaks and traffic congestion will increase the possibility of distracted driving.The random forest model was used to discriminate the distracted driving at the intersection.The results showed that 80%of the samples were used as the training and set 500 decision trees were set to have the highest discriminating accuracy(96.97%).Finally,the left-turn and straight-through behavior of the intersection under distracted driving are studied from the aspects of headway time and start-up loss time.It is found that an increase in the ratio of distracted driving in the vehicle queue will lead to an increase in headway time and start-up loss time,resulting in traffic delay and reducing the capacity of intersections.And from the four aspects of laws and regulations,drivers,intersection management and in-vehicle technology,it puts forward suggestions and measures for improving the driver’s distracted driving at the intersection to improve driving safety.