Research And Simulation Verification Of Efficiency Evaluation Of Cooperative Vehicle-Infranstructure System On Highways Based On Multi-Source Data

As an innovative means of traffic control,the cooperative vehicle-infrastructure system can significantly improve road traffic capacity and reduce traffic accident rate,which is of great significance for improving the efficiency of the current traffic system.However,the cooperative vehicle-infrastructure system is characterized by high complexity,high interaction,and advanced technology,and its effectiveness,feasibility,and reliability lack effective verification.Therefore,it is necessary to build an efficient and accurate cooperative vehicle-infrastructure evaluation system,which not only helps to evaluate the performance,safety and economic benefits of the cooperative vehicle-infrastructure system,but also helps to promote technological innovation and improve public trust.Based on the background,this paper conducts research on the evaluation of the cooperative vehicle-infrastructure system,constructs an objective evaluation system of the cooperative vehicle-infrastructure system,and uses the actual measurement data of ETC and commercial vehicles to carry out statistical verification of the evaluation indicators.In order to obtain data containing comprehensive traffic flow information,the extraction of the full track tracking data across the pole is completed,and it is applied to the calibration research of the car-following model to improve the accuracy of traffic flow simulation.Finally,the simulation technology is used to complete the comparative evaluation before and after the deployment of the cooperative vehicle-infrastructure system.The specific research of this paper is as follows:(1)Establish an evaluation system for the cooperative vehicle-infrastructure system.Firstly,the evaluation subject,namely highway vehicle-road coordination technology,is introduced in three aspects: system architecture,technical classification and typical scenarios.Then,based on this system,an evaluation system is constructed,which is divided into three parts: evaluation index system,evaluation method and evaluation data acquisition.This paper construct an objective evaluation index system for the cooperative vehicle-infrastructure system,which provides an objective dimension of evaluation support for the evaluation of the cooperative vehicle-infrastructure system.Then,various evaluation methods are summarized to evaluate different aspects of the cooperative vehicle-infrastructure system.Finally,based on the existing traffic data collection technology,it is proposed to use ETC data,commercial vehicle data and the full track tracking data across poles to evaluate the cooperative vehicleinfrastructure system.(2)The evaluation is based on the actual measurement data of ETC and the actual measurement data of commercial vehicles.Firstly,the format analysis and characteristic analysis are carried out based on the measured data of ETC and commercial vehicles.Secondly,based on the objective evaluation index system in this paper,the evaluation indicators of the vehicle-infrastructure system suitable for ETC data and commercial vehicle data are respectively designed.Finally,the actual measurement and verification are carried out using the ETC actual measurement data and the actual measurement data of the commercial vehicle under accident and non-accident scenarios on a section of Shandong Highway.The results show that the objective evaluation index proposed in this paper can reflect the state change of traffic flow,so it can be effectively used in the evaluation of the cooperative vehicle-infrastructure system.(3)Based on the re-identification technology,the full track tracking data across poles is extracted.In order to obtain the evaluation data containing comprehensive traffic flow information,this paper proposes the full track tracking data across poles and uses the data set of Chang’an University to carry out data extraction example verification.First,the Faster RCNN network is used for multi-target detection of a single camera.Secondly,the Deepsort is used for trajectory tracking of a single camera.Then,the trajectory re-identification technology is used to complete the cross-camera vehicle trajectory recognition,and the pixel data of the full track tracking data across poles is obtained.Finally,the Hook technology is used to calibrate the pixel position,the pixel data is converted into actual data,and the final full track tracking data across poles is obtained.(4)The car-following model is calibrated based on the full track tracking data across poles.Evaluation of the cooperative vehicle-infrastructure system based on software simulation is a mainstream evaluation method.In order to select a better car-following model for accurate traffic flow simulation,this paper uses the full track tracking data across poles to conduct a comparative study of the car-following model.Firstly,the GHR model,IDM model,and S3 microscopic model were calibrated using the full track tracking data across poles obtained from the Chang’an University dataset.Secondly,the car-following model was compared by combining the calibration results of the car-following model with the American NGSIM dataset and the German High D dataset.Finally,it is concluded that the S3 micro-model performs best in simulating the driver’s car-following behavior,which provides model support for traffic flow simulation.(5)Build a cooperative vehicle-infrastructure system based on software simulation.In order to test the application effect of the evaluation system in the actual situation,and to verify the comprehensiveness of the full track tracking data across poles in the evaluation of the highway vehicle-road coordination system,a cooperative vehicle-infrastructure system based on a software simulation platform was built.Firstly,build the traffic simulation environment before and after the deployment of the cooperative vehicle-infrastructure system,use the SUMO software to build the environment before the deployment of the cooperative vehicleinfrastructure system,and use the SUMO-OMNe T software to build the environment after the cooperative vehicle-infrastructure system;Secondly,according to the actual measurement data of Shandong Highway,the traffic accident scene of the highway is simulated,and for the environment after the deployment of the cooperative vehicle-infrastructure system,the variable speed limit strategy is adopted to implement cooperative control on the traffic accident section of the highway;Finally,the cooperative vehicle-infrastructure system of the highway is objectively evaluated by using the full track tracking simulation data across poles,and the influence of the deployment of the vehicle-road coordination system on the highway efficiency is explored.The results show that the implementation of the cooperative vehicle-infrastructure system can effectively improve the efficiency of highways,which has an important impact on the development of highways in my country.