文献类型：期刊文献

英文题名：Research on the Prediction of Elbow Joint Angle Based on Mechanomyography

作者：Yao, Jiaxu[1]; Xia, Chunming[1,2]; Zhang, Hanyang[1]; Zhu, Chengjie[1]

机构：[1] School of Mechanical and Power Engineering, East China University of Science and Technology, Shanghai, China; [2] School of Mechanical and Automotive Engineering, Shanghai University of Engineering Science, Shanghai, China

年份：2022

卷号：2022-June

起止页码：547

外文期刊名：IEEE Joint International Information Technology and Artificial Intelligence Conference (ITAIC)

收录：EI(收录号：20223512639965)

语种：英文

外文关键词：Exoskeleton (Robotics) - Joints (anatomy) - Patient rehabilitation - Pattern recognition - Robots

摘要：Aiming at the problem that the current upper limb rehabilitation exoskeleton robot is based on pattern recognition to give assistance to patients, which influences the continuous motion of patients, this study proposed a model for predicting the angle of the elbow joint movement based on mechanomyography (MMG). First, the features of the MMG from different angles of the elbow joint under static conditions were extracted. After processing the acquired signal, it is found that there is a certain correlation between the feature of MMG signal and the angle values. Then the MMG signal of the elbow joint during exercises were collected. The Extreme Gradient Boosting (XGBoost) algorithm with the huber loss function was used to predict the joint angle values at the current time instant and 50-200 ms of the next time instant. The results show that the proposed model can accurately predict the future elbow angle. The prediction accuracy gradually decreases with the increase of the prediction time. Therefore, 50 ms was selected to achieve better real-time performance and higher prediction accuracy. This research has the potential to be applied to exoskeleton-assisted robots based on MMG, which can not only meet the requirement of control accuracy, but also realize the prediction of joint angles, with good real-time performance. ? 2022 IEEE.