Many Electromagnetic (EM) problems require a multidisciplnary approach as they need the integration of different capabilities, competences and methodologies. Moreover, because of such problems complexity, EM researchers may take great advantages from already existing solutions: very often, research groups that work autonomously have the know-how needed to solve portions of a given EM problem and therefore they could have developed specific software solutions. Cooperation, data gathering and diffusion among diverse EM research groups (that differ both as for the geographical point of view and as for the themes of their researches) are needed to overcome this fragmentation and this possible co-existence of different solutions for the same problem. As a direct consequence of that, EM researchers need a common “tool” in order to codify their knowledge domain. Ontologies are a very promising technology able to provide such a tool.
Nowadays ontologies are commonly used in many knowledge domains: medicine and healthcare, biology, geography, astronomy, physics… On the contrary, the EM community has only seen few ontological applications: a series of studies in Computer Aided Design (CAD) area. Moreover, no efforts have been made in order to codify the entire EM knowledge domain. Such a purpose, certainly ambitious, requires a strong knowledge sharing that means a wide participation among the EM scientific community.
The research unit for the IT techniques applied to the EM area, operating in the EM fields laboratory at the University of Salento, is in charge of such a task: the ontological codification of the EM knowledge domain. Our research unit has developed an ontological infrastructure named “OntoCEM” (Ontological Codification of EM) in order to promote large-scale research and cooperation in the EM area where so many problems need a multidisciplinary approach.
OntoCEM (as depicted in Figure) has a hierarchical modular structure that makes concept reusability very easy. Other advantages given by our modular approach are represented by the sharing and updating of knowledge. In order to increase the OntoCEM interoperability, already available ontologies have been used where possible.
The upper ontology in our framework enlists general concepts that cover different knowledge domains. By using this kind of ontology a general conceptual infrastructure is available and every domain ontology can be built on top of the upper one. We have chosen the Suggested Upper Merged Ontology (SUMO) as a consequence of its better usability.
In our ontological framework, the mid-level has been split into two sub-levels: the first one comprises the “Scientific Domain Ontologies” and the second one the “EM Domain Ontologies”. This characterization acts as a bridge between top-level abstract concepts and EM domain terms. The ontologies published by the Astronomy Dept. of the University of Maryland (UMD) has been used as scientific domain ontologies. They provide a wide range of scientific concepts useful in the EM area. On the contrary, the EM domain ontologies have been built from scratch.
As shown in figure, the EM domain is made up by many sub-domains and each of them is described by a specific ontology. In order to build those ontologies, many research techniques have been exploited: scientific interviews, research of useful topics among EM conferences, IEEE keyword list, research of useful categorizations among EM Societies, Councils, Committees and so on…
The bottom level of the framework is composed by ontolgies populated with concepts describing specific EM applications such as: antenna CAD, microwave circuits… The application ontologies further specialize their EM domains.
ONTOLOGICAL EXPERT SYSTEM
An ontological knowledge base is made up by instances that “populate” concepts; such concepts represent the ontological description of the knowledge domain for a given problem.
In many EM areas, ontological expert systems are very useful research tools: they can provide the correct methodology to tackle a given problem (being this a hard-to-find solution as it involves many aspects – problem definition and typology, constraints, performances, hardware resources…)
OntoCEM describes both the “EM problems” area and the “EM analysis methods” area so it represents a good way to speed up the decisional process for EM problems and a sharable knowledge repository easily accessible, re-usable and modifiable. Moreover, the ontological framework can be processed by a reasoner, a software entity that can infer new knowledge starting from the existing one.
The EML2 IT technologies research unit is involved in the development of an ontological decision system that has to:
1) guide the user towards the better analysis method choice for a given EM problem
2) provide an ordered and logically structured knowledge repository for EM researchers