东南大学学报（自然科学版）2017，Vol.47Issue(6)：1117-1122,6.DOI:10.3969/j.issn.1001-0505.2017.06.006
基于张量线性拉普拉斯判别的肌电特征提取方法
EMG feature extraction based on tensor linear Laplacian discriminant
摘要
Abstract
To analyze the multi-dimensional characteristics in the time-frequency-space domain implied in the surface electromyogram (sEMG) signlas,a novel feature extraction method for sEMG was proposed based on tensor linear Laplacian discriminant (TLLD) First,sEMG signals were transformed into the 4-D tensor data including the information of temporal,spatial,spectral,and trials by complex morlet wavelet.Secondly,the TLLD analysis algorithm was used to obtain the projection matrix,and the training and test sets were projected into the projection matrix to obtain features with greater discrimination.Finally,the linear discriminant analysis algorithm was used to identify six forearm movements,including the wrist flexion,wrist extension,forearm pronation,forearm supination,hand close,and hand open.The experimental results show that the accuracy of the proposed method is more than 98 ％,and its recognition performance is better than that of three methods of the root mean square,autoregressive coefficient and tensor high order discriminant analysis.关键词
表面肌电信号/人机交互/特征提取/张量线性拉普拉斯判别Key words
surface electromyogram/human-robot interaction/feature extraction/tensor linear Laplacian discriminant分类
信息技术与安全科学引用本文复制引用
佘青山,马鹏刚,马玉良,孟明..基于张量线性拉普拉斯判别的肌电特征提取方法[J].东南大学学报（自然科学版）,2017,47(6):1117-1122,6.基金项目
国家自然科学基金资助项目(61201302,61372023,61671197)、浙江省自然科学基金资助项目(LY15F010009). （61201302,61372023,61671197）
