Traffic Safety Evaluation Of Expressway Merging Area With Mixed Traffic Of Autonomous And Manual Vehicles

Intellectualization is the trend of the road traffic system development,and the automatic driving vehicle is considered as a feasible way to solve the city traffic congestion problem and improve the efficiency of road traffic.In the future,manual driving vehicles may be gradually replaced,but in this process,road traffic will be in the state of mixed traffic of autonomous and manual vehicles.Expressway serves the city’s medium and short distance travel,and bears a large number of travel demands.As the channel connecting ordinary roads and expressways,the traffic condition here is extremely complex,which can easily lead to traffic accidents.Therefore,in the case that the real accident data can’t be obtained,based on the microscopic simulation and traffic conflict technology,this paper carried out a research on the traffic safety evaluation of expressway merging area with mixed traffic of autonomous and manual vehicles.Firstly,this paper chose a single-lane ramp merging area of Huanshan Road as the research object,and the drone aerial photography was used to conduct data surveys.Traffic flow data such as traffic volume,traffic composition and vehicle speed were extracted to further analyze the current traffic operation characteristics.Based on the principle of traffic conflict,both subjective and objective methods was used to judge traffic conflicts.The metric indicators TTC(Time to Collision)and PET(Post Encroachment Time)of rear-end collision and lane-changing conflict were extracted,and their 30% and 85% cumulative frequency values were used as the threshold to divide the severity of the conflict.Besides,calculate the spatial distribution characteristics of conflicts by lanes and regions and identify the conflict influence factors,so as to provide a basis for constructing conflict prediction models under mixed traffic.Secondly,the simulation scene of merging area in VISSIM software were built based on the collected date.Chose travel time and number of collisions as correction indexes,and the two-stage calibration method was used to calibrate the manual driving vehicle’s driving behavior parameters.A meta regression analysis was carried out on the impact of autonomous driving on travel time,and then the driving behavior parameters of autonomous vehicles were calibrated based on the results.Finally,in order to obtain traffic conflict samples,the orthogonal experiment method was used to design mixed traffic simulation experiments under different traffic flow.Based on the negative binomial regression method,this paper built the traffic conflict prediction model in mixed traffic flow.On this basis,use traffic conflict rate to construct the safety rating index system,and then the grey fixed weight clustering model was used to evaluate the safety level of merging area under different mixed traffic flow conditions.According to the distribution characteristics of indicators in evaluation results,the recommended classification thresholds of different traffic safety levels were determined,besides that,this paper gave the overall process and application examples of the traffic safety evaluation research.The results show that,the relative error between predicted and actual number of conflicts is less than 10%.The safety assessment results of the current confluence area are grade C in peak hours and grade A in peak hours.The proposed traffic safety evaluation process can be used to quantify the safety status of mixed traffic in the future.This paper is an effective attempt to pre evaluate the traffic safety situation of mixed traffic,and has great significance to promote the popularization of automatic driving vehicles.