借鉴了主成分分析算法（PCA），基于负熵最大的独立分量分析，含噪脉冲信号进行相关检测,数据模型归一化，模态振动，通过反复训练模板能有较高的识别率。- It draws on principal component analysis algorithm (PCA), Based on negative entropy largest independent component analysis, Noisy pulse correlation detection signal, Normalized data model, modal vibration, Through repeated training wPdMTyilate have higher recognition rate.