Traffic flow brake light model simulation based on driver behavior learning

Abstract

The theory of urban traffic flow has been developed and new types of meta-automata have emerged and simulate realistic traffic conditions relatively well. Among these models, the brake light model can simulate the three-phase traffic flow theory very well. However, the existing brake light model also has certain shortcomings, in that the model will change the speed of congestion propagation upward when the model is covariant, which is not realistic, and the model also lacks simulation parameters for driver behavior. In this paper, we propose a new model based on the brake light model, which can achieve a certain degree of simulation of driver behavior by adjusting the parameters, and also achieve a stabilization of the propagation speed of the congestion wave when the parameters are changed.