Research on Traffic Flow Simulation Model with Multi-Agent Based in Vehicular Sensor Network

A vehicle can be represented as an autonomous simulated agent which can be efficiently analyzed (for the vehicular behaviors and the dynamic performance of traffic flow) by using a multi-agent system. In this article, a new microscopic vehicular behavior mechanism is proposed based on a multi-agent in a vehicular sensor network (VSN) circumstance. Different from the traditional vehicle-following principles, the vehicle agent considers the behavioral decision influence by interactive communication of the signalized intersection-vehicle and vehicle-vehicle (ad hoc). The underlying idea is to establish a single-lane traffic simulation model on the basis of the influence of the VSN. According to the proposed approach, the decision of the acceleration or deceleration rate by the vehicle agent will be altered depending on the front signal phase. In this occasion, the simulations respectively verify the model on microscopic level and macroscopic level in order to analyze the time-space distribution characteristics of the vehicles by interactive information influence. The results show that the proposed mechanism and model can provide better performance than the classic vehicle following model in traditional circumstances.