基于FFT盲辨识的肌电信号建模及模式识别

Modeling and Classifying of sEMG Based on FFT Blind Identification

针对表面肌电信号(Electromyographic signal,sEMG)产生原理复杂、易受人体自身及外界因素影响的特点,采用基于快速傅里叶变换(Fast Fourier transform,FFT)的盲辨识方法建立肌电信号模型.该方法通过计算即可确定信道阶次,无需人为凭借经验设定,且计算简单、易于实现、运算速度快.其利用输出信道间的相互关系特性,实现信号的频域盲辨识,建立数学模型.此方法适用于小样本信号建模,非常适合易受肌肉疲劳影响的表面肌电信号.将模型系数作为改进的BP神经网络的输入,实现多运动模式识别,与其他盲辨识方法比较,此方法识别效果较好.

In this paper, the FFT-based blind identification method is used to establish surface electromyographic signal （sEMC） in order to overcome the disadvantage of sEMG, which is susceptible to muscle fatigue and external factors. With no assumption on the precise knowledge of channel order, the FFT （fast Fourier transform）-based method is able to estimate the channel parameters as well as determine channel order. It extends the cross-relation principle to the frequency domain via the discrete Fourier transform, and performs better in small sample signal modeling, which is suitable for sEMG. The parameters of sEMG model are used as the input of the improved BP neural network to classify different movement patterns and a better recognition result is achieved compared with other blind identification methods.