Dynamic motion phase segmentation using electromyogram

Musculoskeletal system of human modulates elasticity of skeletal muscle so that it produces a great speed by storing a large amount of elastic energy and releasing it over a short period of time. As a consequence, dynamic motion can be divided into several phases related with storing and releasing elastic energy including a countermovement phase which is in the opposite direction to the intended motion. We propose a segmentation algorithm based on a hidden semi-Markov model, which can segment the phases of motion of dynamic tasks based on electromyogram. Two types of experiments, i.e., overhand pitching and countermovement jump were performed to validate the proposed segmentation algorithm.