EMG estimation from EEG for constructing a power assistance system

Abstract

In this paper, aiming to estimate force/torque information from brain activity to help and support the daily lives of human beings, we estimate the human muscular activity from EEG (Electroencephalogram) by PCA (Principal Component Analysis) and RLS (Recursive Least Squares). EEG and EMG (Electromyogram) are measured when a subject is flexing and extending his arm, and their linear model is established by PCA. Then, this linear model between EEG and EMG is updated by the angle, the angular velocity and the angular acceleration of the robot arm. Finally, EMG is estimated from EEG using the updated model. The results show that the estimation of EMG from EEG is possible, and using EEG support human's activities has a great potential.

Keywords

• brain-machine interface
• eeg
• emg
• pca
• rls
• estimation