基于相空间重构的光纤周界信号识别算法研究

Research on vibration signal recognition of optical fiber perimeter based on phase space reconstruction

文章将相空间重构和复小波包变换引入入侵信号类型的识别,对原始信号进行相空间重构,以便更深地反映光纤入侵振动数据混沌特性的内在动力性属性。以相空间重构嵌入维数作为复小波包变换数据输入长度,避免输入信号长度的随意性。采用复小波包提取重构信号的能量分布特征构成入侵信号识别的特征集,以主成分分析对原始特征集降维,通过网格参数寻优算法得到支持向量机(support vector machine,SVM)回归模型的最优参数,以最优参数进行SVM入侵类型识别。实验结果表明,该方法能正确监测入侵事件且误报率与漏报率低。

The phase space reconstruction and complex wavelet packet transform are introduced to the intrusion signal type identification. The phase space reconstruction of the original signal is made, thus to further reflect the inherent dynamic properties of the fiber intrusion time series with chaotic charac- teristic. Embedding dimension of phase space reconstruction is used as the input dimension of the complex wavelet packet transform which can avoid the randomness of the length of input signals. The energy distribution features of the reconstruction signal, which are extracted by using complex wave- let packet, constitute the feature sets of intrusion signal recognition. Principal components analysis （PCA） is carried out to reduce the dimensionality of original feature sets. Grid parameter optimization algorithm is used to obtain the optimal parameters of support vector machine（SVM）, and the intrusion type is identified by SVM with the optimal parameters. The experimental results show that this method can correctly monitor intrusion events with low false positive rate and false negative rate.