车路协同系统下多智能体微观交通流模型  

Microscopic Traffic Flow Model Based on Multi-agent in CVIS Circumstance

作　　者：杨帆[1] 云美萍[1] 杨晓光[1] 

机构地区：[1]同济大学道路与交通工程国家重点试验室,上海201804

出　　处：《同济大学学报（自然科学版）》2012年第8期1189-1196,共8页Journal of Tongji University:Natural Science

基　　金：国家自然科学基金(60974093)

摘　　要：介绍车路协同系统(cooperative vehicles infrastructure system,CVIS)下车辆多智能体(vehicle multi-agent,VMA)的概念及其属性列表.提出传统交通环境和CVIS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.建立CVIS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车辆行为的影响.最后,数值试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.结果表明:CVIS下的车辆比传统交通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.Concepts and attributes list of vehicle multi-agent （VMAs） in cooperative vehicles infrastructure system（CVIS） are introduced at first. Decision strategies in both traditional and CVIS circumstance are compared to analyze differences in traffic status judgment and decision at intersections and links between two circumstances. Single lane traffic flow models in CVIS are established. Microscopic traffic flow models such as acceleration and deceleration are given too. Furthermore, the intersection impacts are taken into consideration in analyzing the signal influence. The numeral experiments analyze the trajectories and macroscopic parameters in both circumstances. The results show that VMAs in CVIS decline sharply in total travel time, average travel time and average delay time which indicate that vehicles in CVIS are more stable and successive than those in traditional circumstance. Besides, headway and velocity variances in CVIS are lower than those in traditional circumstance which increases the fleet stable.

关 键 词：车路协同 多智能体 微观交通流模型 决策机制 

分 类 号：U495[交通运输工程—交通运输规划与管理]