基于谱聚类与RS-KNN的城市快速路交通状态判别

发布时间：2018-05-18 14:31

本文选题：交通工程 + 交通状态判别　；参考：《华南理工大学学报(自然科学版)》2017年06期

【摘要】：为了提高城市快速路交通状态判别的准确性,构建了一种基于谱聚类与随机子空间集成K最近邻(RS-KNN)的交通状态判别模型.以地点交通参数为基础,根据交通流运行特性并结合中国道路服务水平的4个等级,采用谱聚类算法将交通状态划分为4类;然后使用已分类的交通流数据训练RS-KNN模型.通过上海快速路的实测数据完成模型的实验验证和对比分析.实验结果表明,所提出的模型不仅能够提高交通状态判别的精度,而且具有良好的鲁棒性,其判别率比标准KNN模型、BP神经网络模型和SVM模型分别提高7.3%、4.9%和4.5%.
[Abstract]:In order to improve the accuracy of traffic state identification on urban expressway, a traffic state discriminant model based on spectral clustering and stochastic subspace integration K nearest neighbor RS-KNN was constructed. Based on the location traffic parameters, according to the traffic flow characteristics and the four levels of road service level in China, the traffic state is divided into 4 classes by spectral clustering algorithm, and then the RS-KNN model is trained with the classified traffic flow data. The experimental verification and comparative analysis of the model are completed by the measured data of Shanghai Expressway. The experimental results show that the proposed model can not only improve the accuracy of traffic condition discrimination, but also has good robustness. The discriminant rate of the proposed model is 7.3g% and 4.5% higher than that of standard KNN model / BP neural network model and SVM model, respectively.
【作者单位】：吉林大学交通学院;吉林大学吉林省道路交通重点实验室;青岛理工大学汽车与交通学院;山东高速公路股份有限公司;
【基金】：国家科技支撑计划项目(2014BAG03B03) 国家自然科学基金资助项目(51408257,51308248) 山东省省管企业科技创新项目(20122150251-1)~~
【分类号】：U491

【相似文献】