Dynamic Motion Phase Segmentation Using sEMG During Countermovement Jump Based on Hidden Semi-Markov Model

Abstract

Dynamic motion of human shows interesting kinematic aspects related to storing elastic energy to skeletal muscle. This results from joint stiffness modulation and as a consequence, countermovement which is opposite to intended motion is observed. We propose a segmentation algorithm utilizing probabilistic inference of dynamic motion phases from sEMG observations during countermovement jump based on hidden semi-Markov model. In addition, both left-right state transition and restriction of state duration are applied in order to reduce frequent state transition due to large variation of sEMG observation probability. Experimental result shows that the segmentation of motion phase using sEMG satisfies dividing the phases of the vertical position of torso and both the left-right transition and the restriction of state duration succeed to reduce the error rate and the transition occurrence.