The management of data is a great challenge for industries in aviation as the amount of available data has been increasing over the last decades. Aviation data are complex and can be derived from heterogeneous data sources. To handle this challenge, ontologies can be applied to facilitate the modelling of the data across multiple data sources. An aviation domain ontology should be able to promote data driven collaboration between the domains that are directly or indirectly linked with the aviation sector, bringing a common understanding for data assets from diverse domains such as Aerospace, Retail and Weather.
The ICARUS ontology has been developed as a key component of the ICARUS activities towards integrating and semantically enriching aviation-related data from all the three core ICARUS tiers in different formats and from different data sources. Specifically, the core uses of the ICARUS ontology are for: (i) data mapping, as each provided data asset will be mapped with concepts from the ICARUS Data Model (that derives from the ICARUS Ontology) in a semi-automatic way, (ii) data linking where any provided dataset will be linked with other relevant sources (or data assets) that exist in ICARUS and (iii) data recommendation which involves the development of algorithms and software for supporting the selection of the most appropriate ICARUS data assets that best match user preferences.
Design and Development
To design a domain ontology, one can adopt various methods. The most known examples are either to extend existing ontologies or to develop the ontology bottom-up. For the ICARUS ontology, we used a combination of these methods by using a multi-layer approach so that the ICARUS ontology can represent both metadata and aviation-specific concepts. More specifically, the design process consists of the following steps:
- Ontology Capture: All the concepts, relationships and data fields from the ICARUS demonstrators’ datasets (aviation-related and aviation-combined) were extracted.
- Ontology Coding: The representation of the ontology capture was transformed into a formal ontology language (e.g. OWL) using Protégé.
a) Top-level Ontology: Create a top-level ontology that describes aviation-related datasets that were obtained from the ICARUS consortium.
c) Expand the Domain-level Ontology: Create a new domain-level ontology that contains new concepts, data fields and relationships based on the concepts that were extracted from the demonstrators’ datasets.
The ICARUS ontology can be easily maintained and extended during the ICARUS platform operation as new individual domain ontologies could be added under the top-level ontology. For example, existing domain ontologies from diverse aviation-related domains such as transportation, tourism or health can be added as independent domain ontologies (C3, C4, C5…) and their entities/properties can be connected to some of the already incorporated domain ontologies (C1-C2,..). Furthermore, it can answer several different competency questions that can be useful to the aviation industry domain. At the moment, the ICARUS ontology consists of 132 classes, 134 object properties and 406 data properties.
Overall, the ICARUS ontology is able to capture the structural and semantic characteristics of the various entities involved in the aviation domain, whereas the underlying conceptual models facilitate the use of lightweight reasoning during the matchmaking process. Furthermore, it can be utilized to search the integrated data from multiple sources (e.g. flights data, data from social networks accounts of airlines and airports, etc.) using SPARQL queries, as well as to provide high quality recommendations of datasets, based on users’ preferences, using a semantic recommendation system. Finally, a detailed description of the ICARUS Ontology is provided in the ICARUS Deliverable D1.3 (Updated ICARUS Methodology and MVP).
Blog post prepared by UCY.
 Data Tier 1: Primary Aviation Data (e.g. aircraft sensor data, scheduled route plans); Data Tier 2: Extra-Aviation Data (e.g. passengers’ profiles); Data Tier 3: Aviation-derived & Aviation-combined Data (e.g. data from sectors like Health, Tourism, Public Sector)
 Keller, Richard M. “Ontologies for aviation data management.” In 2016 IEEE/AIAA 35th Digital Avionics Systems Conference (DASC), pp. 1-9. IEEE, 2016.
 Corsar, David, Milan Markovic, Peter Edwards, and John D. Nelson. “The transport disruption ontology.” In International Semantic Web Conference, pp. 329-336. Springer, Cham, 2015.
 Ferrandez, Oscar, Christian Spurk, Milen Kouylekov, Iustin Dornescu, Sergio Ferrández, Matteo Negri, Rubén Izquierdo et al. “The QALL-ME framework: A specifiable-domain multilingual question answering architecture.” Web semantics: Science, services and agents on the world wide web 9, no. 2 (2011): 137-145.
 Schriml, Lynn M., Elvira Mitraka, James Munro, Becky Tauber, Mike Schor, Lance Nickle, Victor Felix et al. “Human Disease Ontology 2018 update: classification, content and workflow expansion.” Nucleic acids research 47, no. D1 (2018): D955-D962.
 Jannach, Dietmar, Markus Zanker, Alexander Felfernig, and Gerhard Friedrich. Recommender systems: an introduction. Cambridge University Press, 2010.