Research And Simulation On Car-Following Model For Multi-Information Coupling In Connected Environment

With the construction of intelligent transportation infrastructure and the increasing installation rate of intelligent on-board units,the road traffic environment and the actors involved are transforming towards intelligence and interconnection.The continuous development of the new traffic environment necessitates new adaptation requirements and ideas for the study of traffic instability problems using traditional car-following models.This paper addresses the problem of the traditional car-following model’s insufficient adaptability to simulate the traffic jam law under intelligent and connected traffic environments.It extends the modeling research on the traditional car-following model by combining multiple information obtained through vehicle network technology.In addition,the paper completes a linear stability analysis and simulation validation of the model and builds a system for real vehicle acquisition to complete data acquisition and model calibration in a closed test site.The main work of the paper is as follows:(1)Based on the OVCM(Optimal Velocity Changes with Memory)model,the paper proposes VLOVCM(Velocity Limit and Optimal Velocity Changes with Memory)and VLMVIF(Velocity Limit and Multi-preceding Vehicle Information Fusion)by combining coupled driving information including speed limit information,speed difference of preceding vehicle in front and optimal speed memory of preceding vehicle.(2)The paper derives the neutral stability condition of the proposed model using Lyapunov theory and compares and analyzes the stability of the proposed model with that of the traditional OV,FVD,and OVCM models by plotting critical stability curves.The results show that the VLOVCM and VLMVIF models that introduce speed limiting factors have larger stability regions than the traditional models.The VLMVIF model coupled with multiple front vehicle status,has a smaller instability region than the VLOVCM model.That is,the introduction of multiple information such as speed limits in the networked environment improves the smoothness of traffic flow for the traditional car-following model.(3)The paper conducts numerical simulation experiments of classical scenarios,such as traffic disturbance and convoy start.The experimental results show that the average velocity fluctuation rate of the fleet can be as low as 23.76%and 9.26%after the VLOVCM and VLMVIF models are disturbed,respectively,which is much smaller than the fluctuation change of FVD and OVCM models.The overall start of the following fleet,considering the connected information during the starting process,is faster.During the stopping process,the peak acceleration of the following convoy is lower than that of the conventional model by about 0.5 m/s~2 for the VLMVIF model,considering the speed limit and multi-vehicle driving status information ahead,and the overall stopping time of the convoy is reduced by 2s.The simulation results verify the ability of the proposed model to improve the smoothness of traffic flow.(4)The paper designs and builds a millimeter-wave radar-based car-following data acquisition system,creates a real-vehicle test scheme,and conducts data acquisition tests in a closed test site.Based on the collected following data,the paper calibrates the parameters of the proposed model,such as the speed limit term coefficient r,the front speed difference sensitivity coefficient λ,and the optimal speed memory sensitivity coefficient γ.This verification of the effect of the introduction of the sensitivity factor provides a reference for the selection of the parameters of the following model.