Research On Car-Following Model Based On Multi-Vehicles Information Fusion And Its Simulation System Design

The traditional car-following models usually perceive the forward and backward vehicles restrictively,hence drivers can only judge the surrounding driving environment according to the visual information in certain errors and limitations.With the development of internet technology,connected vehicles can obtain the operating state of multiple forward and backward vehicles in real-time.In order to solve the traffic flow instability caused by the unidirectional and bidirectional speed fluctuations of the front vehicle,the beyond visual information from the forward and backward connected vehicles was fully utilized.Firstly,the state change features of the multiple forward and backward vehicles were introduced to construct a microscopic traffic flow model.Secondly,the stability analysis was performed by the micro perturbation method and the reductive perturbation method,respectively.Thirdly,the sensitivity of key model parameters and the fluctuation characteristics of parameters in different scenarios were derived from the numerical simulations.Finally,a car-following model simulation software based on MATLAB and SUMO was designed and implemented,and the mixed-flow simulation was used as a test case to further verify the feasibility of the software function and the validity of the proposed model.The main works are as follows:(1)A car-following model for connected vehicles based on Multiple Vehicles State Change Features(MVSCF)was proposed by considering the influence of state information of multiple forward and backward vehicles on vehicle following behaviors.The model of MVCM was improved by introducing the state change features,such as acceleration differences of multiple preceding vehicles,speed difference expectations and optimal speed differences.The critical stability conditions were obtained by utilizing the micro perturbation method and the approximate regression method,respectively.In addition,multiple sensitivity analysis were performed on the parameters in MVSCF.Compared with MVCM and other models,the linear stability analysis showed that MVSCF can expand the stability area by introducing the optimal velocities of the multiple preceding and following vehicles and the acceleration differences of the multiple forwarding vehicles.Further,the nonlinear stability analysis concluded that MVSCF can effectively suppress the occurrence of traffic congestion and help improve the smoothness of traffic flow.(2)The MVSCF model was numerically simulated and analyzed by combining the roundabout road and straight road simulation scenarios,respectively.Regarding the multiple preceding vehicles,the relative optimal values of those parameters,such as the acceleration difference coefficient of k,the quantity of q and the optimal speed weight of ε,were derived in the circular road scenario.And the traffic flow stabilization of MVSCF was analyzed under the non-stationary velocity variations of the forwarding vehicles in the straight road scenario.Simulation results showed that MVSCF was more effective than OVCM,ITVDM and MVCM in absorbing fleet disturbances under the simulation of unidirectional and bidirectional velocity abrupt.The fluctuations of velocity and acceleration can be reduced by12.53% and 16.18%,respectively.Hence,MVSCF can facilitate the stationary recovery of the fleet state,and can effectively improve the traffic stability.(3)A car-following model simulation system was designed based on the App Designer and the SUMO in order to further analyze the proposed car-following model.Firstly,the function requirements of the simulation software were outlined,mainly including parameter setting,numerical simulation and joint simulation.Secondly,the design view and code view in App Designer were utilized for component layout and code writing,respectively.And the self-built callback functions and the attribute access functions were used to realize the front and back interaction and data sharing.Thirdly,the information generated by MATLAB,such as vehicle state,was transmitted to SUMO through the Tra CI interface to update the traffic scenarios and vehicle states.Finally,the mixed flow simulation was set up as a test case to verify the feasibility of the developed simulation software and the effectiveness of applying MVSCF to the mixed flow scenario.The test results showed that the developed simulation system can realize the joint simulation based on MATLAB and SUMO,and MVSCF can also recover to the stationary state in fast and smooth under the perturbations,even if the queue topological structure are different and the distributions of connected vehicles are either discrete or continuous.In this paper,MVSCF can effectively absorb the disturbances of the preceding vehicle and facilitate to improve the stability of traffic flow.Meanwhile,the developed simulation system can better understand and analyze the changing characteristics of traffic flow.The results of this work can provide a theoretical basis for researching the evolution of traffic flow and ameliorating the efficiency of traffic operation.