基于PCA和SVM的轨道不平顺状态识别

Track Irregularity State Recognition Based on PCA and SVM

为快速识别轨道不平顺中存在的短波不平顺类型,提出基于主成分分析(PCA)和支持向量机(SVM)进行轨道不平顺状态识别的方法。首先提取轴箱加速度的特征参数,并采用主成分分析法对特征参数进行降维处理,提取出轨道不平顺的主元特征;然后构建支持向量机多分类器,以不同不平顺类型下轴箱加速度数据来验证模型的准确性;最后对实测数据进行轨道不平顺识别。通过对不同轨道不平顺下轴箱加速度的分析,结果表明该方法能够有效地实现一定区段内轨道不平顺类型的识别。

In order to identify the type of short wave irregularity in track irregularity,a method based on principal component analysis and support vector machine is proposed for the identification of track irregularity.Firstly,extracting the characteristic parameters of the axle box vibration acceleration,and then the main elements of the track irregularity are extracted by using principal component analysis to analyze the feature parameters.Secondly,the support vector machine multi-classifier is constructed to verify the accuracy of the model by the use of different types of axle box acceleration data.At last,the measured data are analyzed to identify the track irregularity.Results indicate that the proposed method can effectively realize the identification of the track irregularity in a certain range according to the analysis of the acceleration of the axle box with different track irregularity.