单车道车辆多智能体行为决策机制.doc

车路协同环境下基于多智能体的 单车道微观交通流模型 杨 帆,云美萍,杨晓光 (同济大学 道路与交通工程国家重点实验室，上海 201804) 摘要：车路协同环境(Cooperative Vehicles Infrastructure System, CVIS) 因信息因素的引入而不同于传统的交通环境，其环境下的交通流特性也随之发生了变化。以往的交通流模型建立在传统交通环境下，无法描述CVIS环境下的交通流微观特性。本文首先提出CVIS环境下车辆多智能体(Vehicle Multi-Agent, VMA)的概念及其属性列表。其次，提出在传统环境下和CVIS环境下的车辆决策机制，分析两者在交叉口及路段上对于交通状态判断与决策的差异。在此基础上，提出在车路协同环境下的单车道微观交通流模型，给出车辆的微观动力学模型，包括加速模型和减速模型，同时考虑交叉口的信号灯对于车辆行为的影响。最后，数据实验将分析车辆在两种不同环境中的时空轨迹图以及宏观的交通参数，如整体行程时间，平均行程时间以及平均延误等。结果表明，CVIS环境下的车辆比传统环境下的车辆总行驶时间、平均行程时间以及平均延误均有极大降低，提高了通行的效率，车队车头时距降低、速度方差减小，提高了车队行驶的稳定性。 关键词：车路协同; 多智能体; 决策机制;微观交通流模型 Single Lane Microscopic Traffic Flow Model Based on Multi-Agent in CVIS Circumstance Yang Fan, Yun Meiping, Yang Xiaoguang (Key Laboratory of Road and Traffic Engineering of the Ministry of Education, Tongji University, Shanghai 201804, China) Abstract：CVIS (Cooperative Vehicles Infrastructure System) is different from traditional traffic circumstance because of information involvement and the traffic flow characteristics are also changed simultaneously. Traditional traffic flow models based on traditional traffic circumstance are not adept to describe the traffic flow characteristics in CVIS circumstance. Concepts and attributes list of VMAs (Vehicle Multi-Agent) in CVIS circumstance are presented at first. Then, decision strategies in both traditional circumstance and CVIS circumstance are compared to analysis differences in traffic status judgment and decision at intersections and links between two circumstances. Based on those, single lane traffic flow models in CVIS circumstance are presented in this paper. Microscopic traffic flow models such as acceleration and deceleration are given in this paper. Furthermore, the intersection impacts are also considered in this paper to analyze signal influence. The numeral experiments analyze the trajectories and macroscopic parameters, such as total travel time, average travel time and average delay, in both circumstances. The results show that VMAs in CVIS circum