Development of Algorithm for Abstraction and Teaching of Human Skills by Using Robot

点击翻译按钮获取中文摘要

First, by using an exoskeletal robot, physical parameters such as inertia, friction, and mass of humans can be identified without large equipment such as MRIs. Therefore, the identification is easy and low in cost.Second, hitting skills of humans using hammers are extracted by subtracting the effect of the dynamics from the hitting motion. Then, the skills are unified to dynamics of a robot and the skills are transferred. Therefore, complicated motion that explicitly exploits interference of joints is obtained without solving difficult equations. As a result, the robot can produce faster end-effector speed than its maximum actuator speed.Third, the skills are transferred to other subjects by using an exoskeletal robot. The skills are unified to the dynamics of the subjects, and arms of the subjects are guided to appropriate motion. Hence, the subjects can learn the skills. The way to teach the skills is very similar to the case of skill teaching of sports for the beginners.