Advance semantic integration of aeronautical information, weather and other data

[bermud]

Semantic integration is an approach to bring heterogeneous data together to make complex queries. Example of a queries are:

• Which flights are affected by a volcano eruption
• Which flights will be affected by a hurricane path one day ahead

This task will use various sources of data, that will be ingested and converted to an on ontology store per for semantic integration and enable clientes to perform semantic querying. An ontology like the NASA ATM ontology might be used to integrate the data.

Related work occurred in Testbed 12 work, where SBVR for aviation was tested. Work on this deliverable can advanced the understanding of how SBVR could be implemented based on ontologies. The Aviation Semantics report explored the FAA Web Service Description Ontological Model to improve service discovery within Spatial Data Infrastructures. The future work sections of the report (OGC 16-039) and OGC 16-061 provide further information.

Examples of these data include:

• aeronautical information with AIXM
• weather WXXM (via WCS)
• other data via WFS (e.g. points of interest, buildings, coastlines etc.)

Components:

• server for AIXM
• server for WXXM
• 2 server for other vector data (WFS)
• knowledge (ontology) server
• Translator (converter) from XMLs from various sources to ontologies (using GeoSPARQL)
• Linked data client to display results
• Engineering Report

[pebau]

Interesting topic! Instead of translating XML to ontologies, queries involving raster data and metadata might use WCPS extended with XPath/XQuery, ie: work directly on XML. Feasibility has been shown by Athena Research, and concepts are likely to go into WCPS 2.0. Enabling search expressions of any complexity opens up new vistas for semantic Big Data services.

[ghobona]

OWS-9 did some work on Aviation ontologies using the FAA Air Transportation Information Ontology (ATIO). The work is documented in the OWS-9 Semantic Mediation ER.

There is also the SESAR ATM Information Reference Model (AIRM) which could be used as a high-level conceptual model. We would need to find out from Eurocontrol what the reuse license is for AIRM. http://www.airm.aero/

Testbed-14 could explore how far natural language processing (NLP) tools such as OpenNLP could go to enable users (e.g. pilots, air traffic controllers) to query data integrated through ontologies. This requires some form of integration of an NLP framework with an ontology framework.

Some example NLP tools are listed below:

https://opennlp.apache.org/

https://www.tensorflow.org/versions/r0.12/tutorials/syntaxnet/

Note the relevance to Machine Learning and Artificial Intelligence

[ghobona]

Another idea relating to aviation ontologies is the potential integration of WXXM, AIXM, TDS and OpenStreetMap through ontologies to provide pilots with detailed maps and charts of areas where there are limited authoritative maps and charts.

[pebau]

BTW, Coverage Implementation Schema 1.1 supports RDF as well - that opens options for ontology integration as well. First research on this has been done by Uppsala University.