Upper Limb Prosthesis Control: A Hybrid EEG-EMG Scheme for Motion Estimation in Transhumeral Subjects

This study described the use of Kernel Least Square Tracker based estimation for 3-dimensional shoulder, elbow motion kinematics from surface Electromyogram (EMG) and a two-stage multiclass Support Vector Machine based classification of different wrist, grip and finger motions from Electroencephalogram (EEG). The advantage of employing hybrid EEG-EMG strategy for upper limb motion estimation was demonstrated for a transhumeral subject. The method utilized EMG from upper arm muscles for elbow motion (and shoulder motion in case of higher degree amputation scenario) and used EEG for discerning basic wrist, grip and finger motions. The results showed that the hybrid scheme could estimate shoulder, elbow motion with more than 90% accuracy and wrist, grip and finger motion with 65%-70% accuracy. This strategy of using hybrid EEG-EMG motion estimation, thus, could be employed in developing a more intuitive upper limb prosthesis controller with multiple degrees of freedom.

[1]  Jun Lv,et al.  Decoding hand movement velocities from EEG signals during a continuous drawing task , 2010, 2010 Seventh International Conference on Fuzzy Systems and Knowledge Discovery.

[2]  Ignacio Santamaria,et al.  A comparative study of kernel adaptive filtering algorithms , 2013, 2013 IEEE Digital Signal Processing and Signal Processing Education Meeting (DSP/SPE).

[3]  Erik J. Scheme,et al.  Support Vector Regression for Improved Real-Time, Simultaneous Myoelectric Control , 2014, IEEE Transactions on Neural Systems and Rehabilitation Engineering.

[4]  D. Farina,et al.  Optimization of a set of wavelets for classification of imaginary movement-related cortical potentials , 2008, 2008 First International Symposium on Applied Sciences on Biomedical and Communication Technologies.

[5]  Chih-Jen Lin,et al.  LIBSVM: A library for support vector machines , 2011, TIST.

[6]  Seong-Whan Lee,et al.  Reconstruction of hand movements from EEG signals based on non-linear regression , 2014, 2014 International Winter Workshop on Brain-Computer Interface (BCI).

[7]  Cuntai Guan,et al.  Hand Movement Trajectory Reconstruction from EEG for Brain-Computer Interface Systems , 2013, 2013 IEEE International Conference on Systems, Man, and Cybernetics.

[8]  D. Farina,et al.  Linear and Nonlinear Regression Techniques for Simultaneous and Proportional Myoelectric Control , 2014, IEEE Transactions on Neural Systems and Rehabilitation Engineering.

[9]  Sethu Vijayakumar,et al.  Evaluation of regression methods for the continuous decoding of finger movement from surface EMG and accelerometry , 2015, 2015 7th International IEEE/EMBS Conference on Neural Engineering (NER).

[10]  Erik J. Scheme,et al.  Bagged regression trees for simultaneous myoelectric force estimation , 2014, 2014 22nd Iranian Conference on Electrical Engineering (ICEE).

[11]  Koushik Bakshi,et al.  Estimation of three degrees of freedom of wrist movement from electromyogram using Kernel Ridge Regression , 2017, 2017 8th International IEEE/EMBS Conference on Neural Engineering (NER).

[12]  A. Phinyomark,et al.  Evaluation of EMG feature extraction for hand movement recognition based on Euclidean distance and standard deviation , 2010, ECTI-CON2010: The 2010 ECTI International Confernce on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology.

[13]  Wenwei Yu,et al.  Hand motion reconstruction using EEG and EMG , 2016, 2016 4th International Winter Conference on Brain-Computer Interface (BCI).