Study On Micro Car-following Model And Its Characteristics Based On Multiple Ahead Vehicle & Velocity Difference

As the basis of the traffic flow theory branch, Car-following model is one of the most widely used and most in-depth study category of traffic flow theoretical models. Abroad research in this area has been going on for more than half a century. But there still exists some shortages such as model stability is not so ideal, difference between simulation acceleration and actual situation still exists, etc. On the other hand, the existing research results still can not fully describe the various observed traffic flow phenomenon. How to optimize micro car-following model and to explain the changes in actual traffic flow, needs further study and discussion.On the basis of reviewing all existing micro car-following model of traffic flow, this paper focuses on analyzing the problems of some classic car-following models, which are based on optimal velocity function, e.g., Newell model, OV model, GF model and FVD model. And taking OV model for instance, this paper verifies the model defects via numerical simulation.Based on shortages of existing car-following models, considering multi-expecting optimal velocity and multi-velocity difference information, this paper proposes Multiple Ahead & Velocity Difference model (MAVD). And paper completes the theoretical analysis on linear stability and gets the stability conditions of MAVD model.Comparison& analysis with previous models found that, in MAVD model, the critical value of sensitivity factor for free flow stability is smaller, stable region increases, and model performance improves. Simulation & comparison results show that, the relationship among multi-ahead vehicles, considered by MAVD model, has very important influence on model performance. Therefore, MAVD model can more effectively describe the actual states of traffic flow, and enhance the stability of traffic flow and restrain traffic jam during the process of simulation.For further discussing on MAVD model performance, this paper also analyzes the model parameters' influences on traffic flow simulation and studies on the relationship between parameter changes and hysteresis phenomenon. Besides, this paper studies on the phase transition from free flow to synchronized flow, carrying out the simulation with MAVD model. Analysis indicates that phase transition is continuous if the velocity limit region is long and not continuous if the velocity limit region is short.