Extraction of motor primitive in consideration of arm posture, movement direction and velocity using Hidden Markov Model

In this paper, we propose a method to extract motor primitives from electromyography(EMG) signals on reaching movements of human arm. EMG signals reflect the motor commands from the central nervous system (CNS). Especially, we extract the motor primitives in consideration of arm posture, movement direction and velocity using only EMG signals. As an experimental task, we performed two kinds of experiments on a horizontal plane, measured ten EMG signals and the hand trajectories during movement. Specially, we extracted motor primitive from the EMG signals during movement by using hidden Markov model. Finally, in order to verify how accurately our proposed method divides the motor primitives, we compared the boundary points between the extracted two motor primitives with via-points that were estimated by using forward and inverse dynamics models