基于云推理的车辆跟驰模型CFCM研究  

Research on CFCM:Car Following Model Based on Cloud Model Theory

作　　者：贾顺平[1] 毛保华[1] 

机构地区：[1]北京交通大学轨道交通控制与安全国家重点实验室,北京100044

出　　处：《交通运输系统工程与信息》2007年第6期67-73,共7页Journal of Transportation Systems Engineering and Information Technology

基　　金：国家重点基础研究发展计划项目(2006CB705500);国家自然科学基金项目(70571005)

摘　　要：车辆跟驰模型中的驾驶行为是一个复杂的推理、决策、执行过程,在知识表达和推理机制上存在模糊性和随机性.针对当前车辆跟驰模型研究存在的不足,引入云模型理论,确定了跟驰模型中的决策变量及概念判断子集,设计了模型参数标定的实验方法,在实验数据的基础上,计算了相关参数的隶属云特征值,总结了基于行为规则的车辆跟驰云推理机制,构造了车辆跟驰云模型CFCM(Car Following Cloud Model).Driving decision in the car following model is a complex process of reasoning, deciding and executing. There are a lot of vagueness and randomness in knowledge description and reasoning mechanism. Through the analysis of the problem of recent research, the paper has applied cloud model theory to car following model. Based on the description of processing of different decision information and car following pattern, the paper defines the decision variables and concept judging subset and designs experimental methods of model parameter calibration. On the basis of plenty of experimental data, the paper has calculated the membership cloud eigenvalue of related parameters, summarized the reasoning mechanism of car following cloud based on the behavior rules, and finally constructed the car following cloud model （CFCM）. Key

关 键 词：驾驶行为 跟驰模型 交通流 云理论 隶属云 

分 类 号：U491[交通运输工程—交通运输规划与管理]