Research On Safety Risk Analysis Of Traffic Violation At Urban Signal Intersections

Urban intersections have complex traffic environment and relatively high safety risks,and the key factor affecting intersection safety is traffic violation.Therefore,analyzing traffic violation at intersections is significant to improve urban traffic safety.With the development of urban intelligence level and the construction of intelligent traffic management platform,the automatic equipments to capture traffic violation have been gradually improved,and the traffic violation data gets more diverse.However,the current research work on traffic violations in China is on starting stage,lacking deep mining and processing to data,insufficient utilization of traffic violation data,and inadequate research on safety risk of traffic violations,thus not enough support for traffic safety management.This paper bases on traffic violation data of urban signal intersections,traffic flow data and accident data to explore the safety risk of traffic violation at urban signal intersections.Firstly,pre-processing methods such as spot coding,data filtering,data cleaning,format conversion,etc.are carried out for traffic violation data,traffic flow data and accident data in this paper.Aiming at the characteristics of urban accident report,the accident severity classification method based on semantic analysis model is proposed,and the process including classification number definition,text data preprocessing,word segmentation based on Hidden Markov Chain(HMM),feature word selection and weight setting,classifier construction based on decision tree,then getting accident severity and explain the model accuracy evaluation method.Accident severity data obtained by this method can provide basis for safety risk research of traffic violation,which is measured by accident loss expectation value.Secondly,the definition of intersections with abnormally high occurrence of traffic violation is given in this paper,then analyzing characteristics of traffic flow distribution,constructing regression model between traffic violation and and traffic flow,proposing an identification method for intersections with abnormally high occurrence of traffic violation.Thirdly,this paper clarifies the safety influence area for intersections,combines the identification results of intersections with abnormally high occurrence of traffic violation and the classification results of accident severity to connect traffic violation data and accident data,studies the relationship between specific traffic violations and accident classification by correspondence analysis,thus achieving safety risk analysis of traffic violations at intersections.This paper selects the traffic safety management data of Wujiang District of Suzhou City for example analysis,and verifies the effectiveness of the accident severity classification method based on the semantic analysis model.Then the intersections with abnormally high occurrence of three main traffic violation are identified,the safety risk of traffic violation is evaluated,and the results show that the safety risk caused by road traffic lights violation by motor vehicles is relatively high;the safety risk caused by stopping at the stop line or in the intersection when it's red light is second;the safety risk cased by locating the motor vehicles in the wrong approach at signal intersections is the lowest among three.The results can provide high-risk traffic violation and intersections with abnormally high occurrence of corresponding traffic violation,clarify the objects and areas for urban road traffic administration and indicate the direction for improving road traffic safety level effectively.