摘要
建立表面肌电信号(surface electromyography,sEMG)和关节运动量的映射模型,对康复机器人的连续运动过程控制具有重要意义。为提高预测模型的准确性和适用性,基于肌肉协调理论,采用特征提取和主成分分析提取肌肉激活度,提出遗传算法(GA)优化的支持向量回归算法(Support Vector Regression,SVR)构建肌肉激活度和膝关节角度之间的映射模型,并对健康受试者和膝关节损伤患者在不同运动状态下的sEMG进行膝关节角度预测。GA-SVR算法预测结果的均方根误差分别为1.94°和2.44°,远优于BP神经网络的18.21°和18.28°。结果表明,GA-SVR具有较高的预测精度和较强的通用性。
The model of mapping between surface electromyography(sEMG)and joint motion is very important for the continuous motion control of rehabilitation robot.In order to improve the accuracy and universality of the prediction model,based on the muscle coordination theory,feature extraction and principal component analysis were used to extract muscle activation.The support vector regression(SVR)optimized by genetic algorithm(GA)was proposed to construct the mapping model between muscle activation and knee joint angle.And the model was used to predict the knee angle of healthy subjects and patients with knee joint injury under different motion conditions.The root mean square error of GA-SVR algorithm was 1.94°and 2.44°respectively,which was much better than that of BP neural network 18.21°and 18.28°respectively.The results show that GA-SVR has high prediction accuracy and strong universality.
作者
唐晓娅
陈峰
Tang Xiaoya;Chen Feng(School of Electrical Engineering,Nantong University,Nantong 226019,Jiangsu,China)
出处
《计算机应用与软件》
北大核心
2023年第11期335-340,共6页
Computer Applications and Software
作者简介
唐晓娅,硕士生,主研领域:机器人,智能控制;陈峰,副教授。
