基于最佳小波包的表面肌电信号分类方法

The Approach to Surface EMG Signals Classification Based on Optimal Wavelet Packet

针对表面肌电信号的分类问题,采用最佳小波包分解构造最能体现分类能力的小波包基,用Fisher线性判别分析对肌电信号各个子空间的相对能量特征进行降维处理,然后利用BP神经网络进行分类识别。实验表明该方法能够有效地从伸肌和屈肌采集的两道肌电信号中识别前臂内旋、前臂外旋、握拳和展拳四种运动模式,是一种稳定、有效的特征提取方法,为非平稳生理信号的分析提供了新的手段。

This paper presents an efficient approach to surface electromyographic （SEMG） signals classification. Specifically, SEMG signals were decomposed into a great deal of subspaces by optimal wavelet packet that has maximum discriminant power. Relative energy attached to each subspace was calculated as eigenvalue and feature dimension reduction was conducted by Fisher linear discriminant analysis. The reduced dimensional feature vectors were then used as inputs to a neural network classifier for assessing the classification results. Experimental results show that this approach can identify four types of forearm movement and has great potential in analyzing other nonstationary physiological signals.