利用自组织竞争神经网络提取NEMG信号的MUAP模板

Extraction of MUAP from NEMG Signal Using Self organization Competing NN

采用自组织竞争人工神经网络 ,完成对针电极肌电信号 (NEMG)的运动单位动作电位 (MU AP)的模式分类。MU AP波形的特征取自于其自回归 (AR)模型系数 a1 ～ ap 及激励白噪的功率 εp构成的特征向量。模拟NEMG信号和真实 NEMG信号的实验结果表明 ,这种分类方法具有很高的正确率 ,从而为 NEMG信号分解研究中提取 MU

The pattern classification of motor unit action potential (MUAP) of needle electrode electromyogram (NEMG) signal by means of self organization competing Neural Network(NN) has been accomplished in this paper. The parameters and the power of excited white noise of auto regressive (AR) model are taken as the feature of MUAP. The results of simulated NEMG and real NEMG all show that this method of classification is very effective and correct, and thus it presents a new approach to the extraction of MUAP template in the study of decomposition of NEMG.