Research On Vehicle Following Behavior And Construction Of Vehicle Following Model On Highway Under Rainfall Environment

In the rainfall environment,highway drivers are at higher risk of traffic safety and traffic accidents during vehicle following.Although scholars have established car-following models for rainfall environments,the models must accurately reflect the effects of different rainfall intensities on car-following behaviors.Therefore,studying the effect of rainfall on drivers’ following behavior and constructing a following model for different rainfall intensities can help traffic management departments to take corresponding measures to improve the safety of driving behavior under rainfall conditions.In order to quickly obtain the basic data needed for the study,four parameters characterizing the following state were clearly defined based on the definition of the following state,the criteria for distinguishing the following state were proposed,and the following data were extracted using Python language to automatically extract the vehicle following data from the Hang shao tai smart expressway cloud control platform.Based on the mathematical statistics method,the characteristics of the vehicle following speed,acceleration,headway time distance and headway spacing in normal weather,light rain,medium rain and heavy rain were analyzed,and the results showed that the rainfall weather had a significant impact on the vehicle following behavior of the highway.Compared with normal weather,the speed of following vehicles in rainy weather showed different degrees of decrease,and the speed of following vehicles in heavy rainfall environment decreased by more than 10%,while the headway time distance showed different degrees of increase,and the headway time distance in heavy rainfall environment decreased by more than 13%.The Full Velocity Difference(FVD)model was modified by introducing the safe spacing and expected speed in the rainfall environment to obtain the car-following model in the rainfall environment.The parameters in the model were calibrated by cross-validation,and the model was validated with the measured tracking data under different rainfall intensities.The results show that the maximum root mean square percentage error between the model calculation results and the measured values does not exceed0.5,indicating that the model established in this paper can better describe the following behavior of highway drivers under different rainfall intensities.The research results can provide further evidence for an in-depth understanding of vehicle following behavior and provide a reference for refined management of highway traffic safety under rainfall conditions.