Violation Behavior Analysis Of Pedestrians And Cyclists At Multi-phase Intersection

Red light running at intersections is one of the main reasons for the high traffic accident rate.The existing violations at intersections are mainly studied for the whole red light period.Few documents divide the red light period in detail and deeply study the characteristics and key influencing factors of violations at different red light periods.Firstly,this paper divides the red light period of three-phase and four-phase intersections into straight red light,left turn red light and signal transition period,and analyzes the basic characteristics of crossing behavior in each period in detail.Secondly,builds the models of violation behavior and gap acceptance behavior in different periods and analyzes its key influencing factors.then puts forward targeted measures.The results can provide a scientific basis for the refined management and control of road traffic management departments.The main work of this paper is as follows:(1)Analysis of the basic characteristics of crossing behavior in different periods.At the four-phase intersection,the pedestrian cycle is subdivided into 8 periods: steady green man,green ligh ? near left turn,near left turn steady red man,near left turn ? straight,straight steady red man,straight ? far left turn(three-phase: straight ? green man),far left turn steady red man,far left turn ? green man.At the four-phase intersection,the pedestrian cycle is subdivided into 6 periods.By using the basic statistical analysis methods,this paper studies the differences of the distribution of sample violation behavior,gap acceptance,waiting position and waiting time in different crossing periods.The results show that pedestrians and cyclists arrive at the first 10 seconds of left turn steady red man and two transition periods before and after the green man,and waiting close to the motorway,and the gap value is within the acceptable range of 7 to 13 seconds.,When there is no conflict vehicle before crossing,violation behavior occurs more.(2)Modeling and analysis of violations in different crossing periods.Logistic regression is used to establish multi-phase violation models respectively for the subdivided periods.The results show that there are significant differences in the main influencing factors in different periods.Whether there is a car before crossing is the most important factor that affects the red light running behavior during multi-phase transition periods.,The significant factors that affect the violation during the straight steady red man are traffic flow and maximum endurance time,and the significant factor that affect the violation during the left-turn steady period is traffic mode.(3)Modeling and analysis of gap acceptance behavior in different crossing periods.Based on the boruta algorithm,the optimal feature variables are selected,and the gap acceptance models of three phases and four phases are established by logistic regression and SVM respectively.The results show that the influence factors of gap acceptance during transition period are relatively single,and the gap value is the most important one.However,the influence factors of the steady period are relatively complex.The gap acceptance behavior is mainly affected by the gap value and the crossing stage during the straight steady period,the gap acceptance behavior is mainly affected by the gap value during the near left turn steady period,and the gap acceptance behavior is mainly affected by the gap value,gap type and the number of illegal cyclists during the far left turn steady period.In all violation periods,the prediction accuracy of SVM model is significantly higher than that of logistic model,and the prediction effect of SVM model is better.