Towards a Semantic Knowledge Base for Competency-Based Training of Airline Pilots

Abstract

The acquisition and maintenance of non-technical skills by the pilots are fundamental factors for the prevention of aviation accidents. The aviation authorities are promoting that air crew training be carried out through simulator sessions using scenarios specifically designed to develop and assess the global performance of pilots in such skills. When designing custom flight training scenarios, choosing the correct events and conditions from the myriad of possible combinations with respect to their potential utility in training specific competencies is a costly task that depends entirely on highly specialized expert knowledge. In this paper, we present EBTOnto, an OWL DL ontology that allows to formalize this knowledge and other useful data from real cases, laying the foundations for a semantic knowledge base of scenarios for airline pilots training. Previous advances in this matter and possible applications of this system are reviewed. EBTOnto is built on top of a source validated by experts, the Evidence-Based Training Implementation Guide by the International Air Transport Association, and then checked using an automatic reasoner and a database of 37,568 aviation safety incidents, extracted from the widely regarded Aviation Safety Reporting System by the U.S. National Aeronautics and Space Administration. The results suggest that it is possible to classify real aviation scenarios in terms of non-technical competencies and filter useful incident reports for design and enrichment of these training scenarios. EBTOnto opens up new possibilities for interoperability between incident databases and training organizations, and smoothes the path to represent, share and generate custom simulation training scenarios for pilots based on real data.