Managing industrial automation. How knowledge graphs can boost production

The use of automation in flexible and reconfigurable manufacturing and assembly systems asks for tools and methodologies to manage human-machine collaboration, in terms of tasks’ optimization, performances and safety requirements. The study of automation in production cells and lines shall consider dynamic working environment and variable operational settings. Mapping this knowledge requires structured approaches, such as ontology models, used as references for knowledge representation and reasoning guidance. In this working paper, we propose to merge an ontology with experimental data to construct a knowledge graph for industrial automation management. The knowledge graph representation acts as a digital twin of the process, allowing to exploit graph metrics (e.g., shortest path, degrees, centrality) and subsequently develop indicators for guiding systems' tasks’ assignment and adaptation. The proposed solution is showcased on a case study related to a full-scale lab model of a Self-Adaptive Smart Assembly Cell.