Augmenting Model-Based Systems Engineering with Knowledge

This article presents a general approach for the integration of Knowledge Bases into Model-Based Systems Engineering tools. In existing tools, domain-specific modeling languages are well supported. However when it comes to enforcing design constraints, existing approaches are verbose, it is difficult to be complete and consistent, and the reuse of knowledge is only possible in a limited way (mainly through model libraries). Furthermore, current tools usually lack or have limited capability to detect semantic errors, ability to evaluate the models with respect to formal expert knowledge, and the ability to understand what is being designed. Our work addresses these limitations through the semantic annotation of UML models in Papyrus (an MBSE Tool), to attach domain-specific semantics to the models. This integration enables not only reasoning capabilities over the annotated models, but the models can be shared with semantic-compatible tools and stakeholders. Moreover, the models can reuse and integrate knowledge generated outside the tooling environment. The approach's feasibility is demonstrated through an implementation that defines a technology stack, with emphasis on the mapping of UML elements and its counterparts in the ontology. We address the coherence and preservation of the semantics throughout the transformation process, which enable the formalization of constraints coming from the UML's system design. Finally, we illustrate the reasoning capabilities by evaluating expert knowledge via SPARQL queries and SWRL rules.