Research On Merging And Lane Changing Risks In Highway Exit Areas Under Heterogeneous Traffic Flow Environment

In recent years,China has issued multiple policies to support the development of intelligent driving technology and the construction of smart highways.Currently,cars with intelligent driving functions are already on the road,and smart highways have been built and put into use in many cities across the country.In the future,intelligent driving cars will gradually enter highspeed highways,and the heterogeneous traffic flow composed of intelligent and traditional driving vehicles will inevitably coexist on the highways for a long time.The traffic environment in the exit area of the highway is complex,and the traffic volume and operating speed on the mainline are higher than on the ramp.Lane changing behavior in the exit area of the highway is more frequent and complex than on the mainline.Therefore,it is necessary to further study the risk of lane changing in the heterogeneous traffic flow at different proportions of intelligent vehicles merging out.This article is based on the research conducted under the National Natural Science Foundation of China(52002033): " Mechanism and Optimization Method of Driving Behavior Chain at Agglomeration Area between nearby On and Off Ramps of Freeway".The study focuses on the risk of lane-changing at the outlet of heterogeneous traffic flow with different proportions of intelligent vehicles in a direct single-lane exit area.This abstract has been written in accordance with the rules of capitalization and word usage.This article begins by investigating the distribution of highway exit types in Xi’an,and selecting the direct single-lane exit at Qujiang Interchange for field research.Video analysis software is used to statistically analyze traffic flow,speed,headway,and lane-changing time characteristics in the exit area.Based on the quantitative analysis of traditional driving vehicles in the direct single-lane exit area,the Krauss and LC2013 models are selected as car-following and lane-changing models,and the ACC/CACC model is selected as a car-following model based on qualitative analysis of intelligent vehicle operating characteristics.A game model is constructed as a lane-changing model,and model parameters are calibrated using average speed as an evaluation index through multiple simulations.Based on simulation data,the speed,acceleration,lane-changing time,longitudinal distance,headway,and speed difference of vehicles in highway exit areas are analyzed,and the TTC and SDI are selected as microscopic risk evaluation indexes in the time and space dimensions,respectively.Macro-level risk exposure levels(REL)and risk severity levels(RSL)are constructed to further quantify the risks of exit lane changing.Finally,a risk comprehensive evaluation model is built based on the fault tree theory,and highway exit lanechanging safety assessment is conducted using different proportions of heterogeneous traffic flow simulation results as data.The risk conflict exposure level and risk conflict severity level are statistically analyzed,and the K-Means clustering method is used to cluster the comprehensive risk index of direct single-lane exit lane-changing into three categories to determine different levels of exit lane-changing risk indexes.The percentage of exit lanechanging risk levels under different proportions of intelligent driving vehicles is analyzed.This abstract has been written in accordance with the rules of capitalization and word usage.The study found that increasing the proportion of intelligent driving vehicles improves exit lanechanging efficiency.However,the risk of exit lane-changing in heterogeneous traffic flow shows a trend of increasing and then decreasing,reaching its maximum when the proportion of intelligent vehicles is 50%.This abstract has been written in accordance with the rules of capitalization and word usage.