Modeling And Analysis Of Heterogeneous Traffic Dynamics In A Connected And Autonomous Environment

The introduction of connected and autonomous vehicles will bring changes to the current traffic systems by enhancing safety,improving mobility,and reducing energy consumption,etc.With the increase of market penetration rate of connected and autonomous vehicles,the traffic flow will be composed of human-driven vehicles(HVs),connected human-driven vehicles(CHVs),autonomous vehicles(AVs),and connected autonomous vehicles(CAVs).Due to changes in traffic scenarios,traditional traffic models and analysis methods cannot be used to evaluate the impact of connected and autonomous vehicles on the traffic system.To this end,this paper establishes a dynamic model of heterogeneous traffic flow under an intelligent connected environment and derives an analytical method for its corresponding characteristics.The main contributions are as follows:(1)A new model to capture the car-following behavior of connected vehicles is proposed.The actual car-following data of connected vehicles are collected by carrying out field experiments.The gray correlation analysis method is used to quantitatively analyze the specific changes in driving behavior after the driver receives the vehicle-to-vehicle(V2V)warning messages.According to the data analysis results,a modified intelligent driver model is proposed to evaluate the impact of V2 V messages on traffic flow evolution.Then,the model parameters are calibrated to verify the ability to reproduce the car-following behavior of connected vehicles.Besides,the dynamic characteristic of a connected traffic system is studied by linear stability analysis and simulation.The results show that a connected vehicle can mitigate cascading braking events if the V2 V messages are properly responded to.(2)A general analytical method for nonlinear stability analysis that can be applied to multiclass car-following models with time delays is proposed.Firstly,a generic car-following model with multiple delay structures is established.The linear stability condition of this generic model is then derived.The nonlinear density wave of traffic flow is derived based on the analytical solution of the modified Korteweg–de Vries equation.The results of the analytical analysis are verified by numerical simulation.The results show that the delays of the sensing gap and the velocity have a significant impact on linear stability.The sensing delay of velocity difference does not affect the linear stability,but a slight effect on the nonlinear stability.(3)A general Hopf bifurcation analysis method is derived,which can be applied to multiclass homogeneous traffic systems.The delayed optimal velocity(OV)model,delayed full velocity difference(FVD)model,delayed intelligent driver(ID)model,and delayed cooperative adaptive cruise control(CACC)model are respectively taken as case studies.The results of bifurcation analysis are verified by traffic simulation experiments under periodic boundary conditions.The results show that the delay in sensing the vehicle velocity cancels out the negative impact of the delay in sensing the gap.(4)A general Hopf bifurcation analysis method for multiclass heterogeneous traffic systems is proposed.The bifurcation characteristics of traffic flow composed of human-driven vehicles,adaptive cruise control(ACC)vehicles,and CACC vehicles are analyzed.The bifurcation analysis results are verified by numerical simulation.The results show that popularizing CACC vehicles can gradually improve the stability of traffic flow.Finally,the framework of the heterogeneous traffic flow models and the generic bifurcation analysis methods are applied synthetically.The bifurcation characteristics of the mixed traffic flow composed of HVs,CHVs,AVs,and CAVs are studied.The results show that AVs are not conducive to improving the stability of traffic flow,while CHVs and CAVs can alleviate the formation and propagation of traffic oscillations.The car-following model presented in this dissertation can be applied to upgrade the existing traffic simulation software so that it can have the function of evaluating the impact of connected vehicles on traffic flow.The findings of model characteristics analysis results imply theoretical significance for the design of the CACC or ACC algorithm to reduce the negative impact of human/machine delays in traffic flow.