Research On Intersection Safety Assessment Method Based On Conflict Between Pedestrian And Right-turning Vehicle

In recent years,the accelerated urbanization and the increased proportion of private car trips lead to a rising number of traffic accidents year by year.In traffic accidents,pedestrians are the vulnerable group,whose accidents mainly occur at road intersections.In order to improve the safety of pedestrian crossing at intersections,this paper,taking traffic conflict technology as theoretical support,based on the survey data,studies and analyses the conflict between pedestrian and right-turning vehicle to evaluate the safety of pedestrian crossing at intersections.Firstly,the investigation plan and the data extraction plan for pedestrian and right-turning vehicle conflict at intersections are analyzed.Tracker software is applied to extract the trajectories of pedestrians and vehicles based on the observation videos.The camera calibration is completed by homogeneous coordinate transformation method to convert the image coordinates into actual coordinates.The characteristics of conflict between pedestrian and right-turning vehicle are summarized through data analysis.The critical gap between pedestrians and conflict vehicles is analyzed,and the probability distribution is fitted by Weibull function.Secondly,the applicability and validity of TTC,PET and DST in judging the severity of pedestrian-vehicle conflicts are analyzed,and calculation models are established to solve the index values of each sample.Based on the fuzzy C-means clustering algorithm,a hierarchical partition model of the severity of pedestrian and right-turning vehicle conflicts at intersections is established,which divides the conflict into serious conflict,general conflict and minor conflict.The critical value of each grade is calculated.Next,the observational variables are increased,and a Multi-ordinal regression analysis model is established to explore variables which have influence on the severity of conflict between pedestrians and right-turning vehicles.It is found that seven observational variables,namely,conflict vehicle type,traffic flow,car-following situation,vehicles' velocity variation,the number of conflict pedestrians,pedestrians'velocity variation and conflict type,are significant influencing factors.The influence equations are established to analyze the influence degree of each factor.Finally,an improved grey clustering method is used to establish an evaluation model of pedestrian crossing safety at intersection with the impact of right-turning vehicle based on traffic conflict technology.Taking the conflict areas as the evaluation objects,the model establishes an evaluation index system including the number of conflicts,the number of serious conflicts,TC/MPCU and the serious conflict rate,and constructs an improved grey clustering model based on the whitenization weight fun.ction of sinusoidal curve to evaluate the safety situation of survey samples.And the accuracy of the model is verified.In this paper,Tracker software is used to automatically track the calibrated conflict objects and extract their trajectories,and the critical gap between the conflict vehicles and pedestrians is studied.And the accuracy of safety evaluation has increased by refining the evaluation objects,establishing a conflict index system,and improving the clustering model.