CAJ | 학술논문

为提高表面肌电信号分解的正确率以及完整性，提出一种表面肌电信号的自动分解算法。考虑到其较低的信噪比，先是采用小波降噪法对信号进行降噪处理，并计算信号的非线性能量算子以加强信号波峰值。之后采用低频小波系数和高频小波系数相结合的特征值来表征运动单元动作电位，最后使用小波神经网络完成对活动段的分类。同时，为了实现表面肌电信号的分解完整性，采用递归模版算法对所提取的叠加波形加以分解。实验结果表明，该分解算法能够成功地提取到中低收缩水平下表面肌电信号中的运动单元动作电位的发放信息，同时也能够有效地对叠加波形进行分解。
위제고표면기전신호분해적정학솔이급완정성，제출일충표면기전신호적자동분해산법。고필도기교저적신조비，선시채용소파강조법대신호진행강조처리，병계산신호적비선성능량산자이가강신호파봉치。지후채용저빈소파계수화고빈소파계수상결합적특정치래표정운동단원동작전위，최후사용소파신경망락완성대활동단적분류。동시，위료실현표면기전신호적분해완정성，채용체귀모판산법대소제취적첩가파형가이분해。실험결과표명，해분해산법능구성공지제취도중저수축수평하표면기전신호중적운동단원동작전위적발방신식，동시야능구유효지대첩가파형진행분해。
In order to improve the accuracy and integrality of the decomposition rate of surface electromyography (sEMG),we propose an automatic decomposition algorithm for sEMG.Considering its lower signal-to-noise ratio,we first use wavelet de-noising method to reduce the noise of the signal,and calculate the nonlinear energy operator of the signal to strengthen the peak value of the signal.After that we use the eigenvalue combined by the high frequency and low frequency wavelet coefficients of the sEMG signal to express the motor unit action potential (MUAP).Finally,we use wavelet neutral network (WNN)to complete the classification of the active segment.Meanwhile,in order to implement the integrality in the decomposition of sEMG,we decompose the superimposed waveforms based on the recursive template alignment technique.Experimental results demonstrate that this decomposition algorithm can successfully extract the issuance information of MUAP in sEMG in medium and low contraction level,at the same time it is able to effectively decompose the superimposed waveforms as well.