基于肌电信号的下肢关节连续运动预测

Continuous kinematics prediction of lower limb joints driven by electromyography

为研究不同运动模式下基于肌电信号的下肢多关节连续运动预测,通过支持向量机对肌电-运动的映射关系进行训练,实现对下肢髋、膝和踝3个关节矢状面内的连续运动预测.由10位健康受试者的运动预测和统计分析可知:在适速行走过程中,髋、膝和踝关节的关节角度预测均方根误差分别9.36°,10.82°和6.87°;在不同运动模式下,关节的运动预测值与测量值之间均表现出一定的相关性,其中,膝和髋关节的预测值与测量值之间相关系数均大于0.72,表现出比较明显的相关性.实验结果表明:基于肌电信号进行下肢多关节连续运动预测,尤其是在适速行走时对膝和髋关节的运动预测是可行的.

In order to evaluate electromyography（EMG）-driven simultaneous prediction of lower limb kinematics under different locomotion patterns,support vector machine（SVM）was used for the EMG-kinematics correlation training and the kinematics prediction of hip,knee and ankle joint in sagittal plane was achieved.From the statistical results and analysis of ten able-bodied subjects,it is found that the RMS error of hip,knee and ankle joint angle prediction is respectively 9.36°,10.82°and 6.87°during normal walking.The prediction of the joint kinematics represented consistent correlation with their measurement under different locomotion modes.The hip and knee joint angle prediction was especially correlated to the measured angle with correlation coefficient greater than 0.72.Experiment results show that the EMG can be applied to predict the lower limb joint kinematics,especially for hip and knee joint angle prediction during normal walking.