A robot learning from demonstration framework for skillful small parts assembly

Increasing demand for higher production flexibility and smaller batch size pushes the development of manufacturing industry towards robotic solutions with fast setup reprogram capability. Aiming to facilitate assembly lines robots, learning from demonstration (LfD) paradigm has attracted attention. A robot LfD framework designed skillful small parts applications is developed, which takes position, orientation wrench data into consideration. In view constraints in industrial applications, two cascaded polices are learned separated avoid potential under-fitting problem. With proposed policies, reference trajectories generated as well coupled position data. Effectiveness validated by a printed circuit board experiment torque-controlled robot.