摘要
为了提高高速公路的交通运行效率,需要实时预测各路段交通流参数状况,通过对高速公路宏观动态交通流模型的分析,以及对SMO支持向量机参数选择的研究,提出了高速公路动态交通流支持向量机预测模型.以西安-宝鸡高速公路交通流信息采集数据对模型进行训练、测试和仿真,预测平均相对误差小于3.84%,表明了模型的有效性.
In order to improve freeway transportation efficiency, Traffic flow parameters of the road situation need forecasting in real time. By analyzing the freeway macroscopic dynamic traffic flow model and carrying out detail research on selecting parameter of SMO support vector machines, a model of freeway dynamic traffic flow forecasting that bases on SMO support vector machines is proposed. The real data collected from traffic flow at Xi＇ an-Baoji freeway are used to train, test and simulate the model. The average relative error of forecasting is less than 3. 84% . The result is satisfied.
出处
《西安工业大学学报》
CAS
2009年第3期280-284,共5页
Journal of Xi’an Technological University
关键词
高速公路
交通流参数
支持向量机
预测
freeway
traffic flow parameters
support vector machines
forecast
