基于主成分分析和ART2网络的车辆波形分类  

Clustering of vehicle waveform based on principal component analysis and ART2 neural network

作　　者：沈艳超[1] 叶青[1] 吕望[1] 

机构地区：[1]长沙理工大学电气与信息工程学院,湖南长沙410076

出　　处：《湖南文理学院学报（自然科学版）》2009年第4期58-61,共4页Journal of Hunan University of Arts and Science(Science and Technology)

基　　金：湖南省自科基金(07JJ6123)

摘　　要：利用环形线圈车辆检测器,对不同车辆经过环形线圈时产生电磁感应数据进行采集,再对这些维数较大的车辆感应数据,运用主成分分析法对其降维处理,以减少ART2神经网络输入维数,便于对这些数据进一步分类.通过实验研究可知,这种主成分分析与ART2神经网络相结合的分类方法,使数据通过降维处理,并减少网络的输入维数,不仅大大简化了网络结构,还明显提高了分类的准确性、快速性,加快了网络的学习速率,而且从分类效果上看,也具有很好的正确率.Loop induction detector has been used to collect electromagnetic induction data which were caused by vehicles, which was crossing the toroidal coil under the road. The dimension of vehicle electromagnetic induction data were reduced by principal component analysis and a new sample space was created. The new sample space was employed as the inputs of ART2 network, so the the input dimension of ART2 network was reduced. Through the experimental study, this classification methods can get good results. The construction of ART2 network was simplified, and the convergent speed of the ART2 network was enhanced greatly due to the number of the ART2 inputs was reduced. From the Classification results, a good Accuracy can be got.

关 键 词：主成分分析法 ART2神经网络 车辆检测 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]