基于聚类分析的交叉口状态识别方法  

Research on the intersection traffic state identification based on clustering analysis

作　　者：郝斌斌 吕斌[1] 张浩玮[1] Hao Binbin;Lu Bin;Zhang Haowei(School of Traffic and Transportation,Lanzhou Jiaotong University,Lanzhou 730070,China)

机构地区：[1]兰州交通大学交通运输学院,兰州730070

出　　处：《青海交通科技》2022年第6期25-31,43,共8页Qinghai Transportation Science and Technology

基　　金：甘肃省人文社会科学项目(项目编号:5)。

摘　　要：针对城市交叉口交通状态有效识别的问题,建立了三种交叉口交通状态识别方法。结合实际交通流数据,利用模糊聚类传递闭包法、K均值聚类算法和模糊C均值聚类算法对交叉口交通流周期平均速度、排队长、平均车头时距等向量进行聚类分析,给出适合城市交叉口交通流特点的交通状态划分方法,能够实时、准确、全面地识别城市交叉口的运行状态,为城市交通管理部门交通控制、交通诱导、交通指挥等提供数据基础。定性定量分析的影响,所得结论对城市道路路内停车设施优化具有重要的参考作用。For the problem of effective identification of traffic state at urban intersections,this paper establishes three intersection traffic state identification methodsCombined with the actual traffic flow data,the fuzzy clustering transfer closure method,K-means clustering algorithm and fuzzy C-means clustering algorithm are used to cluster the average traffic velocity,queue length and average headway of the intersectionThe traffic state division method suitable for the traffic flow characteristics of urban intersections is givenThe method can identify the running status of urban intersections in real time,accurately and comprehensively,and provide data foundation for traffic control,traffic guidance and traffic command of urban traffic management departments.

关 键 词：交通工程 城市交通 交通流状态 K均值 模糊C均值 

分 类 号：U491[交通运输工程—交通运输规划与管理]