1. |
- Buselli, Irene, et al.
(författare)
-
Natural language processing for aviation safety : Extracting knowledge from publicly-available loss of separation reports
- 2022
-
Ingår i: Open Research Europe. - : F1000 Research Ltd. - 2732-5121. ; , s. 1-39
-
Tidskriftsartikel (refereegranskat)abstract
- Background: The air traffic management (ATM) system has historicallycoped with a global increase in traffic demand ultimately leading toincreased operational complexity.When dealing with the impact of this increasing complexity on systemsafety it is crucial to automatically analyse the losses of separation(LoSs) using tools able to extract meaningful and actionableinformation from safety reports.Current research in this field mainly exploits natural languageprocessing (NLP) to categorise the reports,with the limitations that theconsidered categories need to be manually annotated by experts andthat general taxonomies are seldom exploited.Methods: To address the current gaps,authors propose to performexploratory data analysis on safety reports combining state-of-the-arttechniques like topic modelling and clustering and then to develop analgorithm able to extract the Toolkit for ATM Occurrence Investigation(TOKAI) taxonomy factors from the free-text safety reports based onsyntactic analysis.TOKAI is a tool for investigation developed by EUROCONTROL and itstaxonomy is intended to become a standard and harmonisedapproach to future investigations.Results: Leveraging on the LoS events reported in the publicdatabases of the Comisi n de Estudio y An lisis de Notificaciones deIncidentes de Tr nsito A reo and the United Kingdom AirproxBoard,authors show how their proposal is able to automaticallyextract meaningful and actionable information from safetyreports,other than to classify their content according to the TOKAItaxonomy.The quality of the approach is also indirectly validated by checking theconnection between the identified factors and the main contributor ofthe incidents.Conclusions: Authors' results are a promising first step toward the fullautomation of a general analysis of LoS reports supported by resultson real-world data coming from two different sources.In the future,authors' proposal could be extended to othertaxonomies or tailored to identify factors to be included in the safetytaxonomies.KeywordsATM, Safety, Resilience, Natural Language Processing, Losses ofSeparation, Safety Reports, TOKAI
|
|