Towards Ontology-based Handling of Uncertainties in Robotic Assembly

abstract

In the context of Industry 4.0, robot systems need to handle new and more complex tasks to produce highly customized products at small lot sizes. Small and medium-sized enterprises, in particular, lack the expert knowledge to parameterize such systems and to take relevant uncertainties into account. Ontologies provide functionality to explicitly encode knowledge using a common vocabulary. In this work, ideas towards the ontology-based representation of uncertainties and associated handling strategies are presented. We define an uncertainty taxonomy and combine it with knowledge about products, manufacturing processes, and resources, following the PPR modeling paradigm. The concept is implemented and tested using a robotic assembly task of an electronic component. As a result, the integration of different types of knowledge enables the automatic adjustment of robot processes based on the consideration of involved uncertainties. This may lead to an easier adaption of robot programs for new products and a more robust operation.