Safety Discrimination And Analysis Of Road Sections With Bus Lanes In The Intersection Area

Using bus lanes is gradually becoming a significant form of traffic organization in urban traffic.The establishment of bus lanes changed the original road structure,which may bring some new traffic safety hazards.Currently,the research of the safety evaluation of road with bus lanes and related affecting factors is still rare.Based on measured data,this paper comprehensively deduced the relationship between safety measurement of roads with bus lanes and road facilities or traffic volumes through data collection,data processing,data modeling,data analysis and other steps.In this paper,a total of 18 sections with bus lanes in Shenzhen and Chengdu were selected for investigation.The investigation was conducted by combining shooting videos with on-thespot investigations.The road traffic facilities data of the road section is obtained through onthe-spot investigation,and the captured video is processed to obtain the traffic conflict PET data and the traffic volume data of the road section.Then,based on the concept of regression level in the extreme value model,the concept of the accidents return level of millions of intersection imported road traffic volume(ARL)was proposed to measure the safety of the road sections with bus lanes.The threshold value of the acquired PET data is determined.The data exceeding the threshold is modeled according to the generalized Pareto model,and the ARL value is calculated,based on which,the safety of the road section is judged.After that,the obtained data of various influencing factors are preliminarily analyzed and used to predict ARL of road sections,and a linear model is established.At the same time,for safe and unsafe binary variables that are more concerned in safety identification,three discrimination models are proposed,which can discriminate the safety condition of roads with bus lanes.And also,their discrimination accuracy rate were evaluated respectively,and the discriminant effect of the decision tree learning model is most ideal.Finally,based on the results of data analysis,further data analysis is carried out.According to the results of data analysis and model interpretation,influencing factors such as the standard deviation of traffic volume in all directions,the distance from the bus stop to the stop line,the standard deviation of traffic volume in all direction on each lane,play an important role in the planning and safety discrimination of roads with bus lanes.In the end,this paper provides some suggestions for the safe operation of the road sections with bus lanes.