基于sEMG信号的下肢康复动作识别算法  

作　　者：谭智峰 张磊 苗娟[1] 刘剑伟 陈永源 

机构地区：[1]山东信息职业技术学院,山东潍坊261000 [2]山东泽普医疗科技有限公司,山东潍坊261000

出　　处：《科技创新与应用》2023年第34期37-40,共4页Technology Innovation and Application

基　　金：潍坊市科技局项目(2021TSGC1260)。

摘　　要：表面肌电信号(sEMG)是一种神经电信号,其能反映神经、肌肉的状态及功能。近年来,越来越多的研究人员对基于sEMG信号的下肢康复动作的问题进行探讨并取得巨大突破。随着深度学习模型的日益成熟,各种模型对sEMG信号下肢康复动作的准确率有明显提升。该文使用基于CNN网络的sEMG信号下肢康复动作识别算法,首先数据预处理过程,选用截止点为20 Hz(低)和450 Hz(高)的有限脉冲响应带通滤波器对sEMG信号进行处理,之后进行特征提取,满足实验数据的需求。最后介绍基于CNN网络的sEMG信号下肢康复动作识别,说明CNN网络的结构以及参数设置。结果显示该文采用的CNN算法具有良好的识别效果。Surface electromyography(sEMG)is a kind of neuroelectric signal,which can reflect the state and function of nerves and muscles.In recent years,more and more researchers have discussed the problem of lower limb rehabilitation based on sEMG signal and made a great breakthrough.With the increasing maturity of deep learning models,the accuracy of various models for lower limb rehabilitation movements with sEMG signals has been significantly improved.In this paper,the lower limb rehabilitation action recognition algorithm of sEMG signal based on CNN network is used.In the process of data preprocessing,the finite impulse response bandpass filter with cutoff points of 20 Hz(low)and 450 Hz(high)is selected to process the sEMG signal,and then the features are extracted to meet the needs of the experimental data.Finally,the lower limb rehabilitation action recognition of sEMG signal based on CNN network is introduced,and the structure and parameter setting of CNN network are explained.The results show that the CNN algorithm used in this paper has a good recognition effect.

关 键 词：CNN模型 深度学习 SEMG信号 运动肌肉群 康复动作 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]