Research On Traffic Safety Of Vehicles’ Lane-changing Behaviors At Signalized Intersection Entrance Based On Traffic Conflicts

The increasing number of motor vehicles has brought about more and more traffic accidents.Urban signalized intersection is the throat of urban road traffic,which can make the traffic flow in various driving directions freely turn and cross,but this powerful function also makes it a frequent place for accidents.Through on-site investigations,although vehicles’ lane-changing behaviors rarely cause serious accidents,they are prone to traffic conflicts that affects the safety and efficiency of other vehicles.Therefore,it is of great significance to study the lane-changing behavior of vehicles at the signalized intersection entrance.First of all,this paper determines the scope of the signalized intersection entrance,the type of lane-changing behaviors and the type of conflicts,and designs a data collection plan.Based on the Tracker software,the trajectories of the vehicles’ lane-changing and related traffic data are extracted,and the conflict recognition algorithm is used to obtain the number of possible conflicts for each lane-changing behavior type,which provides basic data for the following study on the characteristics of traffic conflicts.Secondly,by using 85% bit cumulative frequency curve,the severity thresholds of two traffic conflicts of two lane-changing types are obtained,and the final number of traffic conflicts is determined by eliminating the relevant traffic interference data.The conflict data is studied from traffic characteristics,location characteristics and time characteristics,and the mechanism of traffic conflicts is analyzed,which provides a basis for safety improvement in the following article.After that,the partial correlation analysis of the possible influencing factors of traffic conflict is carried out to prepare for the construction of the conflict prediction model.Thirdly,based on the negative binomial model,using Bayesian method to establish the prediction models of destination type of lane-changing behavior-rear end collision,destination type of lane-changing behavior-side collision,efficiency type of lane-changing behavior-rear end collision and efficiency type of lane-changing behavior-side collision,and the influence degree of each variable on traffic conflict is analyzed by using the elastic coefficient index.And the elastic coefficient index is used to analyze the impact of each variable on the traffic conflict.After the validity test of the model,the results show that the conflict prediction model can predict the number of traffic conflicts at the intersection entrance more accurately.Finally,this paper constructs a comprehensive traffic conflict measurement index.Taking the traffic conflict rate as the evaluation index,the safety of the intersection entrance is divided into four levels by using the grey clustering method.Combined with the conflict prediction model,the safety evaluation process of the entrance lane is established.Through the analysis of the example,the corresponding improvement measures are put forward for the safety problem of the entrance road,and the improvement results are verified by the traffic simulation software.This paper studies the characteristics of vehicles’ lane-changing behaviors,the mechanism of traffic conflicts,traffic conflict prediction,safety evaluation,and safety improvement measures at intersection entrance.The research results can provide reference and basis for determining and improving the safety of the intersection entrance with a large number of lane-changing behaviors.