【摘要】 针对表面肌电信号(SEMG)的非平稳特性,提出了一种以复杂性测度和支持向量机(SVM)相结合的肌电信号模式识别新方法。肌电信号的复杂度作为一种新的肌电信号特征,算法简单。支持向量机是一种新的机器学习机制。通过对采集的四通道表面肌电信号进行分析,提取其复杂性测度信息构建特征矢量,利用“一对一”分类策略和二叉树构建的多类支持向量机分类器,很好地实现了对前臂的八种动作表面肌电信号的模式分类。实验表明,由支持向量机对肌电信号的复杂度特征进行分类,具有很好的稳定性和准确率,为肌电信号及其它非平稳生理电信号的模式分类提供了一种新思路。更多还原

【Abstract】 To overcome the obstacle of Surface Electromyography(SEMG) nonstationary characteristic,a novel SEMG pattern classification method,which is based on the complexity and the Support Vector Machine(SVM),is proposed.The complexity of SEMG is a new characteristic and its algorithm is simple,while SVM is a new mechanism of machine learning.From raw four channel SEMG signals of corresponding muscles,their complexity are extracted and signal characteristics are constructed.A new multi-class support vector machine classifier is designed in "one versus one"classification strategy and binary tree.Eight forearm movement patterns are successfully identified.Experiments show that the SVM can accurately sort out eight movement patterns and the recognition result is robust.It offers an alternate method for SEMG pattern classification,which can be straightforwardly expanded to other nonstationary bioelectric signals pattern classification study.更多还原