Continuous Knee Joint Angle Prediction with Surface EMG

Jiyuan Yang,Zhiguo Lu,Siwei Chen,Chong Liu,Haibin Zhao
DOI: https://doi.org/10.1016/j.bspc.2024.106354
IF: 5.1
2024-01-01
Biomedical Signal Processing and Control
Abstract:Surface Electromyography (sEMG) is an important source for obtaining human motion information, the joint angle of the human body can be predicted. Therefore, this paper proposed a method for predicting the knee joint angle of the lower limb based on sEMG signals. Firstly, the muscles with the strongest correlation during the walking state were selected by using muscle synergy. Then the root mean square (RMS) of the selected muscles and their gradients were used as features for predicting the knee joint angle. Finally, regression models were used to predict the knee joint angle. We chose to use publicly available dataset from the Exoskeleton and Prosthetic Intelligent Control (EPIC) laboratory to validate the proposed method in this paper, and the predicted angle can provide real-time reference for the control process of the lower limb exoskeleton and other devices.
What problem does this paper attempt to address?