Robot Teaching and Learning Based on “Adult” and “Child” Robot Concept

Abstract

Recently robotic technology has been advanced rapidly. There are many robotic applications in manufacturing environments to replace human workers. However, there are many unsolved problems in robotic automation. One issue is how to teach a robot without sacrificing the safety issues; the second one is how to make a robot adapt to the changes and uncertainties in a robotic manufacturing process. To face these challenges, the concept of "adult" robot and "child" robot is presented. Role characteristics and task division are identified and analyzed in this paper. These concepts are applied in a force guided assembly task. The proposed solution involves robot navigation, mobile manipulation, robot learning and force controlled intelligent assembly. An "adult" robot with advanced sensing and decision-making capabilities will teach a "child" robot how to perform a task. The force controlled intelligent assembly strategy will enable a "child" robot to successfully perform assembly tasks with changes and uncertainties. Experimental platforms have been developed and some preliminary results demonstrate that the proposed method is very promising. The proposed methods will enable industrial robots to be more intelligent.