ABSTRACT
AutomationML (AML) can be seen as a partial knowledge-based solution for manufacturing and automation domains since it permits integrating different engineering data format, and also contains information about physical and logical structures of production systems, using basic concepts as resources, process, and products, in semantic structures. However, it is not a complete knowledge-based solution because it does not have mechanisms for querying and reasoning procedures, which are basic functions for semantic inferences. Additionally, AutomationML does not deal with aspects of sensor fusion naturally. In this sense, we propose an ontology to describe those sensors' fusion elements, including procedures for runtime processing, and also elements that can turn AutomationML into a complete knowledge-based solution. The approach was applied in a case study with two different industrial processes with some sensors under fusion. The results obtained demonstrate that the ontology allows describing sensors that are under fusion and deal with the occurrence of data divergence. In a broader view, the results show how to apply AutomationML description for runtime processing of data generated from different sensors of a manufacturing system using an ontology to complement the AML description, where AutomationML concentrates knowledge about a specific production system and the ontology describes a general and reusable knowledge about sensor fusion.
