电子测量与仪器学报

裴晓敏,宋佳强,曹江涛,刘洪海.基于MEMD和TK能量算子的肌电信号手势识别[J].电子测量与仪器学报,2021,35(1):82-87

基于MEMD和TK能量算子的肌电信号手势识别

Surface EMG signal hand motion recognition based on MEMD and TK energy operators

DOI：

中文关键词: 表面肌电信号多元经验模态分解 Teager Kaiser能量多维尺度分析线性判别分类器

英文关键词:surface EMG signal multivariate empirical mode decomposition Teager Kaiser energy multidimensional scaling analysis linear discriminant classifier

基金项目:国家自然科学基金（61873259）、辽宁省自然科学基金计划（2019ZD0066）资助项目

作者	单位
裴晓敏	辽宁石油化工大学信息与控制工程学院抚顺113001
宋佳强	辽宁石油化工大学信息与控制工程学院抚顺113001
曹江涛	辽宁石油化工大学信息与控制工程学院抚顺113001
刘洪海	朴茨茅斯大学智能系统与生物医学机器人实验室朴茨茅斯PO1 3QL

Author	Institution
Pei Xiaomin	School of Information and Control Engineering, Liaoning Shihua University, Fushun 113001, China
Song Jiaqiang	School of Information and Control Engineering, Liaoning Shihua University, Fushun 113001, China
Cao Jiangtao	School of Information and Control Engineering, Liaoning Shihua University, Fushun 113001, China
Liu Honghai	Intelligent Systems and Biomedical Robotics Group, University of Portsmouth, PO1 3QL Portsmouth,UK

摘要点击次数: 840

全文下载次数: 3

中文摘要:

为提高肌电信号手势识别的准确率，提出基于时频域分析的肌电信号特征提取方法。该方法利用无线肌电信号采集装置获得肌电信号，采用基于多元经验模态分解（multivariate empirical mode decomposition，MEMD）和TK（Teager Kaiser）能量算子的肌电信号特征提取方法，利用多维尺度分析（multi dimensional scaling，MDS）对多通道特征降维，采用线性判别分类器（linear discriminant analysis，LDA）对手势特征分类识别。将该算法应用于UCI数据库，手势识别准确率达9896%, 应用于自主采集数据库准确率达9937%，同时F1 score 具有明显提升。实验结果表明，与典型方法相比，所提出的肌电信号特征提取方法对手势识别的准确率更高。

英文摘要:

To enhance the accuracy of gesture recognition using electromyogram(EMG) signals, we present an EMG signal feature extraction method based on time frequence domain analysis. Firstly, a wireless EMG signal acquisition device is designed. Secondly, a gesture recognition method based on multivariate empirical mode decomposition (MEMD) and Teager Kaiser (TK) energy operator is proposed. Multi dimensional scaling (MDS) method is used to reduce the dimensionality of multi channel features. then, linear discriminative classifier (LDA) is used to classify and recognize gesture features. The accuracy of this algorithm for UCI database can reach 9896%. The recognition accuracy for self collected database can reach 9937%. Meanwhile, F1 score also enhances significantly. The experiments verify that the method we proposed can reach a higher accuracy recognition results than other typical methods.

查看全文查看/发表评论下载PDF阅读器