A Simulation of Car-Following Driving Based on Coordinated Model

The different car-following indexes in different traffic flow states are extracted first. And then a model is advanced to imitate the whole course from collecting information, making decisions to controlling the states of vehicles by using a new method named Information Fusion. One neuron is set, as a perception, in order to show the car-following behavior is discontinues. And different decision-constrain in different traffic flow is set in order to reveal the difference of driver´s decision. The results in off-line exam show that the model could reflect the current of vehicles´ states, and its outputs are accurate. So the ideal and method used in the model is feasible.