基于高斯混合模型聚类算法的交通状态划分  

Urban road traffic state identification based on Gaussian mixture model clustering algorithm

作　　者：黄艳国 张升升 刘红军 HUANG Yanguo;ZHANG Shengsheng;LIU Hongjun(College of Electrical Engineering and Automation,Jiangxi University of Science and Technology,Ganzhou 341000,China)

机构地区：[1]江西理工大学电气工程与自动化学院,江西赣州341000

出　　处：《现代电子技术》2022年第7期168-173,共6页Modern Electronics Technique

基　　金：国家自然科学基金项目:基于多源信息融合的城市路网交通拥堵形成与演化机理研究(72061016);留学基金委资助项目(2019年75号文201908360225);江西省教育厅科技项目(GJJ160608)。

摘　　要：为了能够对道路实时交通状态进行有效判别,提出一种利用高斯混合模型聚类算法判别城市道路交通状态的方法。通过同时对高德地图数据平台实时监控和现场数据采集获得道路的速度、流量等判别指标以及交通状态,并且在流量的基础上提出宽裕度判别指标。为防止GMM算法陷入局部最优,运用改进的均值-标准差算法分析预处理数据获得聚类中心,将实验所得聚类中心值代入高斯混合模型进行聚类分析实验,通过匹配矩阵将各算法实验分类结果与实际交通状态值进行对比获得算法的分类精度。结果表明,改进的GMM算法精确率达到98.8%,较未改进的GMM算法精度提高了0.2%,改进GMM算法在精确度和稳定性方面均优于其他几种算法,验证了高斯混合模型对交通数据进行状态判别的有效性。In order to effectively identify the real-time traffic state of the road,a method of identifying urban road traffic state with Gaussian mixture model(GMM)clustering algorithm is proposed.By real-time monitoring and field data collection of Gaode map data platform at the same time,the road speed,traffic flow and other discriminant indexes as well as traffic status are obtained,and the margin evaluation indicators are proposed on the basis of traffic flow.In order to prevent GMM algorithm from falling into local optimum,the improved mean-standard deviation algorithm is used to analyze the preprocessed data to obtain the clustering center.The clustering center value obtained in the experiment is brought into the GMM for the experiment of clustering analysis.The matching matrix is adopted to compare the experimental classification results of each algorithm with the actual traffic state value to obtain classification accuracy of the algorithm.The results show that the accuracy rate of the improved GMM algorithm reaches 98.8%,which is 0.2%higher than that of the unmodified one.In terms of the accuracy and stability,the improved GMM algorithm is superior to the other algorithms,which verifies the effectiveness of the GMM for traffic data state discrimination.

关 键 词：交通状态判别 高斯混合模型 聚类分析 数据预处理 聚类中心 匹配矩阵 判别指标 

分 类 号：TN911.1-34[电子电信—通信与信息系统] U491.17[电子电信—信息与通信工程]