Abstract: The traffic simulation is the important research means to Evaluate and optimize the traffic control and the traffic management. The characteristic of the multi-information stimulation, the task concentration of the driver, and the cooperative reaction behavior are not considered synthetically in the traditional microscopic traffic simulation model, especially for the car-following model which is used to portray the driver’s behavior. In this paper, the fuzzy integral and Bayes methods are used to describe the process of the multi-information fusion of the driver in the complex running environment. At the same time, these methods are also applied to the driver’s behavior after the task concentration. The model test shows that the two algorithms are effective to simulate the task concentration and the cooperative reaction before the car-following model is implemented.

Key words: driver’s behavior, multi-information fusion, task concentration, cooperative simulation

摘要： 交通仿真是交通控制与管理方案评价和优化的重要研究手段.传统的微观交通仿真模型，特别是刻画驾驶员行为的车辆跟驰模型，未能综合考虑交通环境中信息刺激的多源性和驾驶员任务集聚、协调反应的行为过程.本文利用Bayes方法和模糊积分方法描述驾驶员在复杂行驶环境中多源信息的融合过程，确定驾驶员任务集聚后对车辆应采取的驾驶行为.模型验证表明:交通仿真过程中，在车辆跟驰模型实施之前，利用Bayes算法和模糊积分算法模拟驾驶员在多源信息刺激下任务集聚、协同反应的过程是行之有效的.

关键词: 驾驶员行为, 多源信息融合, 任务集聚, 协同仿真