Abstract: In order to increase the fidelity and credibility of the driving simulation system，this paper establishes the intelligent driving behavior model of virtual autonomous vehicle，on the basis of micro-macro external road environment. Also，a simulation method based on local information perception and stochastic decision making has been proposed to improve the probability of independent intelligence vehicle and real-random driving behavior. Finally，the result illustrates the driving behavior of virtual autonomous vehicle can be realized by using these methods，which simulate real traffic environment and also enhance authenticity and real-time capability of system.

摘要： 为了提高驾驶模拟系统的逼真度和可信度，以自主车驾驶行为作为研究对象，在结合宏观与微观信息进行虚拟道路环境建模的基础上，提出一种基于局部信息感知和知识随机决策的概率模拟方法，目的是提升虚拟车的自主智能性和实现驾驶行为的真实随机性。实验与应用结果表明，提出的方法能够较真实地再现车辆的自主驾驶行为，模拟现实中交通车辆环境，并提高了模拟系统的真实性、形象性和实时性。