Summary: Through the use of imitation learning we are able to capture and store the movements of highly skilled humans when operating our robotic hands on our CA.  The information stored on the CA cloud, will be used to develop a skill fusion technology that can be used to allow even beginners to work at the same level of an experts with years of experience.

We use a high performance hand robot with haptics that is attached to manipulators that provide a degree of freedom similar to the human arm.
Thanks to the ability to capture advanced skills on the CA, not only will we assist beginners by supplementing an expert's skills. But through continued CA cloud development we are able to create extraordinary models that will combine and supplement the skills of multiple highly skilled experts.

Goal: By recording the skills of various types of experts, we will develop an imitation learning model to predict the movement and actions of the human hand by recording the joint angles, haptics readings, torque values and other similar parameters. By recording and capturing various experts actions, we will change the way that beginners are evaluated, as through the CA cloud they can perform as highly skilled experts, while learning these applied skills at a fraction of the time.

Since FY2023, Group 3 "Modeling of Expert-Skills and Development of Skills Transfer Technology using CA" has been integrated into Group 4 "Research and Operation of the Next-Generation CA Cloud and Promotion of International Standardization".