Study On The Choice Of Street Crossing Forms Based On The Traffic Conflict Between Right-turning Motor Vehicles And Pedestrians

At urban road intersections,right-turning motor vehicles are generally not restricted by signal control and are prone to traffic conflict with pedestrians.When the traffic conflict develops to a certain degree,it will evolve into a traffic accident.Therefore,it is necessary to study the traffic conflict between right-turning motor vehicles and pedestrians.Research on the traffic conflicts between right-turning motor vehicles and pedestrians in different right-turning areas,and explore the impact of different right-turning areas and traffic flow factors on the traffic conflicts between right-turning motor vehicles and pedestrians,and then give the choice of crossing form,to reduce traffic conflicts between right-turning motor vehicles and pedestrians,thereby to improve traffic safety in the right-turning area at the intersection.Taking the traffic conflict between right-turning motor vehicles and pedestrians as the research object,selecting three intersections with different forms of crossing the street,using a combination of drone video surveys and manual surveys for data collection,and tracking right-turning motor vehicles and pedestrians through Tracker The trajectory is extracted,and the system analyzes the traffic characteristics of right-turning motor vehicles and pedestrians.Use the mean filter algorithm to smooth the track information extracted by Tracker,select TTC as the traffic conflict discrimination index,and use the traffic conflict algorithm to further discriminate and analyze the smoothed trajectory.Using the 85% digit cumulative frequency curve method to classify the severity of traffic conflicts.When the value range of TTC is 0～1.1s,it is regarded as a serious conflict,and when the value range of TTC is 1.1～2.6s,it is regarded as a general conflict.When the value range of is above 2.6s,it is regarded as traffic interference,and the traffic interference is regarded as not constituting a conflict.And analyze the impact of traffic conflicts between right-turning motor vehicles and pedestrians from different right-turning areas,such as the form of street crossing and traffic flow.Considering the heterogeneity of three different street crossing forms,based on the Bayesian method,a random-effect Poisson log-normal regression model and a random-effect negative binomial regression model were established.The model was calibration and goodness-of-fit test.The results showed that the random-effect Poisson log-normal distribution regression The model fitting effect is better.Combining the posterior distribution of the model,quantitatively analyze the impact of different right-turning areas and traffic flow factors on traffic conflicts.Taking the traffic conflict rate as the evaluation index,the Gaussian mixture model is used to divide the safety level of the right turn area at the intersection,and the safety level is divided into five levels.In addition,an intersection is selected for example analysis,and the selection of the crossing form is given based on the calculation result of the traffic conflict model and the safety level of the right-turn area at the intersection,and some reasonable suggestions are put forward for the right-turn area at the intersection.