用小波和神经网络相结合的方法识别人体表面肌电信号

Uniting the means of wavelet and ANN to identify the sEMG signal

以MALTAB语言作为系统设计工具,将小波分析与神经网络相结合分析人体表面肌电信号。对SEMG信号的识别分为3个步骤:数据预处理,特征的提取,设计分类器分类。首先利用小波分析进行消噪,提取特征;然后采用BP神经网路进行分类、识别;最后通过对分类结果的分析与比较,证明小波与神经网络相结合是一种有效的表面肌电信号的模式识别方法。

In this paper we use the programing language of MATLAB as system designing tools, uniting the means of Wavelet and ANN to identify the SEMG signal. To identify the sEMG signal include three processes: preprocessing dates, charactering pick-up and classification. In the first using Wavelet to denoise and pick up the characters, then using BP NN to classify and identify, at last the result was analyzed and compared. The result indicate that the union of Wavelet and ANN is an effective mode to identify means of sEMG signal.