Research Of Automatic Driving Lane Deployment Based On Mixed Traffic Equilibrium

Connected and Automatic Vehicles(CAV)have huge potential benefits in improving traffic efficiency and traffic safety.However,the popularization of CAV does not happen overnight.The mixed traffic mode of CAV and human-driven vehicles(HDV)will continue for a long time.The characteristics of mixed traffic flow are still unclear at present.The impact of large-scale mixed traffic on the transportation network needs to be explored urgently.Therefore,this paper researches on the optimal deployment of CAL lanes(CAVL)based on mixed traffic equilibrium under mixed traffic to improve the traffic efficiency and ensure safe by providing transportation infrastructure priority for CAV.A travel cost calculation function was proposed by considering the multi-dimensional benefits of CAV in capacity,time value,energy consumption,etc.,Then,a mixed traffic equilibrium model was innovatively proposed in the CAVL environment.Multi-granularity factors that affect the benefits of CAVL was decoupled through numerical analysis.Numerical analysis results show that:CAVL is most suitable during penetration rate is 0.5 to 0.8;CAVL’s design speed has an important influence on the calculation of travel time cost.The CAVL design on road network needs to consider the benefits during the design period.However,while CAV development forecasting methods is failing in China.In order to fill this gap,the CAV penetration rate diffusion model was invented based on innovative theory.The model parameters were calibrated using historical sales data of electric vehicles in China.The results show that 100%coverage of CAV can be achieved in about 35 years under natural diffusion.When the infrastructure construction rate is 0.5,it is 8 years earlier than under natural diffusion.Considering the CAVL periodic benefits,a bilevel programming model for CAVL optimal deployment is proposed.The upper model is the system cost optimization problem,which is to minimize the overall system cost within the planning period.The lower level is the mixed traffic equilibrium problem.Combine genetic algorithm and method of successive average to solve the bi-level programming model.A test is done to prove he feasibility of the model and algorithm through the Nguyen-Dupuis network,and compare other methods to analyze the superiority of the proposed models.The case analysis results show that traffic demand is the prerequisite for CAVL deployment,followed by penetration rate;the CAVL distribution tends to be concentrated to achieve the system’s optimal CAV flow.This research expands the mixed traffic flow distribution method in the intelligent network environment,fill the gaps in the domestic CAV permeability diffusion research,and proposed a CAVL optimal deployment model.It provides important theoretical basis and technical support for traffic flow management.