Abstract
In this paper, we propose a visualization method to integrate significant amounts of information relative to human motion to facilitate convenient visual perception during motor learning. This motor learning system helps subjects acquire developed motor skills by referencing integrated information of optimized motion data using a visualized motor skill map. In the proposed method, a self-organizing map (SOM) is employed to visualize the integrated motion data. It is expected that subjects can perceive optimized motor skills involving the activation of various muscles easily by comparing the trajectories of integrated data of optimized motion and that of human subjects wearing inertial and electromyography (EMG) sensors. Here integrated information is comprised of muscle activation signals, joint angle and joint angular velocity, and was visualized with a SOM as a two-dimensional map. To search an optimal landing motion, the human body was modelled as a musculoskeletal system composed of eight rigid bodies and nineteen muscle-like actuators. We performed numerical optimization using muscle-actuated forward dynamics simulations with a multi-objective genetic algorithm. Then, we created a motor skill map by using SOM with motion data that we got by optimization and experiment. We visualized motion characteristics as a color map, and compared optimized motor skills and subject’s motor skills using the obtained SOM.
