Direct teaching and error recovery method for assembly task based on a transition process of a constraint condition

Abstract

A method for simplification of teaching in robotic assembly is presented. Generally, an assembly task is a process in which a constraint condition between two workpieces changes. There has been much research on teaching an assembly task to a robot system by a model-based approach. However, it takes a lot of time to revise information of models that depend on the task environment when every new task is taught to a robot in these model-based approaches. We propose a method to simplify teaching by taking an approach where a human teaches a task to a robot without giving models of the task environment and task processes. Additionally, we also propose a method to realize robotic assembly that is robust to errors by executing a task as a transition process of a constraint condition. Experimental results show the effectiveness of our method.