With the rise of data analytics, we are developing ontologies for data fusion and reasoning
Point-to-point connections between information systems or third-party data managers glue together the majority of today's systems engineering ecosystems. The focus is on data movement and traceability rather than organizing data in ways to better derive meaning and knowledge from it.
Meanwhile, data complexity increases exponentially as systems become more complex and the number of engineering models increases to cover more aspects of systems engineering processes, including requirements, architecture, independent verification and validation, and sustainment. And, over a system’s life cycle, the models and their data evolve as the system goes into operations and sustainment.
Engineers and logisticians need to track decisions made on a system by the varied domains and disciplines in the ecosystem, with each creating information in its own language and jargon. This presents challenges to understanding system impacts and limiting the complexity of information.
With the rise of data analytics, the systems engineering community is figuring out how to use ontologies to better define, fuse, and reason the data it generates. Ontologies guide the definitions and structuring of different things in an ecosystem based on their relationships with each other.
Ontology development in systems engineering will improve how data generated by different domains and disciplines is organized. The use of ontologies will improve data shareability and aid the discovery of system component relationships from data.
SAIC is making investments in building a set of ontologies for the systems engineering ecosystem, eliminating differences in syntax and semantics and establishing sets of common languages. We are building out a systems engineering process ontology, as well as ontologies for systems engineering artifacts and digital engineering artifacts. A fundamental, top-level ontology and the ISO/IEC/IEEE systems engineering process ontology seed our suite of ontologies, which lays the groundwork for interoperable ontologies and expansion across domains and disciplines.
Leveraging ontologies, semantic tools can perform advanced categorization and processing of data produced across the systems engineering ecosystem. A semantic broker arranges data for processing to a common language and set of rules. It exposes that data for use by any other tool in the ecosystem according to those rules.
A semantic broker not only can assemble data from varied information systems in a more central fashion but also provides domain and discipline traceability and intelligent analysis through different rule sets.
Smart data agents can provide automated advice to users about the data they’re creating or using, such as to look at another part of the system architecture that a component change could impact. For example, architecture analysts will not have to leave their native tools and pull in requirements or build-of-materials data from other sources in the ecosystem to see whether system decisions or changes affect that data.
Ontologies are most effective when they are widely used. We are working with partners in industry and in systems engineering on getting a reference ontology tied to standards so that vendors can develop their tools around it and other organizations can use it as the seed to build their ontologies.
Through a standardized framework that provides a mechanism for interoperable ontologies, organizations from various domains can plug in their application ontologies and a set of rules understood by both computers and humans can manage a variety of information systems.
With ontologies and semantic brokers in place, the systems engineering ecosystem will be able to dictate data curation, thus facilitating better integration of the various pieces of digital engineering data produced and providing a holistic picture of a system to various domains and disciplines.