基于表面肌电信号的步态切换意图识别及控制研究  

Research on Intention Recognition and Control of Gait Switching Based on Surface Electromyographic Signal

作　　者：尚长沛[1,2] 程志 SHANG Chang-pei;CHENG Zhi(School of Automotive Engineering,Henan Polytechnic Institute,Nanyang 473009,China;Henan Material Forming Equipment Intelligent Technology Engineering Research Center,Henan Polytechnic Institute,Nanyang 473009,China)

机构地区：[1]河南工业职业技术学院汽车工程学院,河南南阳473009 [2]河南工业职业技术学院河南省材料成形装备智能技术工程研究中心,河南南阳473009

出　　处：《控制工程》2021年第8期1662-1668,共7页Control Engineering of China

基　　金：2019年河南省科技攻关项目(192102210165)。

摘　　要：为了帮助老年人、残疾人及康复中后期的下肢功能损伤患者进行自我康复训练和日常活动,提出了基于表面肌电信号的关节角度预测方法和主动控制外骨骼触发步态切换策略。采用基于BP神经网络的控制器预测关节角度,采集腿部6块肌肉的表面肌电信号,利用均方根对其进行特征提取。然后将信号数字归一化后定向到BP神经网络模型作为输入,关节角度作为输出。根据输出角度识别人体步态切换意图并主动控制外骨骼触发相应的步态。受试者穿戴外骨骼进行在线实验。实验结果表明,所提方法具有较好的实时性和准确性,在自我康复训练和辅助行走等方面具有广阔的应用前景。To help the elderly,the disabled and the patients with lower limb function injury in the middle or later stage of rehabilitation with rehabilitation training and daily activity,the joint angle prediction method based on surface electromyographic(EMG)signal and strategy of actively controlling the exoskeleton to trigger gait switching are proposed in this paper.The controller based on BP neural network is used to predict joint angles.The surface EMG signals of six leg muscles are collected,and the root-mean-square method is used to extract their features.Then,the surface EMG signals are number normalied and directed to the BP neural network model as input,and the joint angles as output.Human gait switching intention is recognized according to the output angles,then the exoskeleton is actively controlled to trigger relevant gait.The subjects wear exoskeleton for online experiments.The experimental results reveal that the proposed control method has good real-time performance and accuracy,and has a broad application prospect in self-rehabilitation training and walking assistance.

关 键 词：表面肌电信号 下肢外骨骼 BP神经网络 步态切换 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]