摘要
在高速公路交通动态数据融合需求分析的基础上,采用环形线圈和微波检测器建立多检测器动态数据的现场试验站,通过多检测器组合方式构建数据检测方案;应用基于自适应加权和改进BP神经网络的数据融合方法,建立交通动态数据融合模型,研究高速公路同一时间、相同断面的多检测器的数据融合。现场试验与检测数据分析表明:基于改进BP神经网络融合方法所获得试验数据的平均相对误差较微波和环形线圈各自的精度提高了10%-20%。
Based on the requirement analysis of traffic dynamic data fusion of expressway, the field experiment station was set up by combining microwave with loop detectors. The data detection scheme was formed by the multi- detection mode. The traffic dynamic data fusion model was established by using the data fusion methods based on adaptive weighting and improved BP neural network to study the traffic data fusion of multi-detector at the same time and section. The results indicate that the average relative error of test data acquired by the data fusion methods of im- proved BP neural network can be increased by a range between 10% and 20% comparing with the detection preci- sion of microwave or loop detection.
出处
《公路交通科技》
CAS
CSCD
北大核心
2009年第3期130-134,共5页
Journal of Highway and Transportation Research and Development
基金
河南省科技攻关计划项目(072102360060)
江苏省交通重大计划项目(7621006024)
关键词
智能运输系统
多检测器数据融合
BP神经网络
高速公路
Intelligent Transport Systems
data fusion of multi-detector
BP neural network
expressway