Skip to main content

When airlines and ANSPs come together


The project team came together for the last Consortium Meeting on November 6th and 7th in Majorca. Big thanks to Air Europa who supported and hosted the meeting.

For two days, five airlines (namely Air Europa, Iberia, Norwegian, Pegasus and Vueling) met with the 3 ANSPs participating in the project (Austrocontrol, ENAIRE and LFV), along with Eurocontrol, AESA and EASA (Spanish and European Safety Authority respectively). The last meeting was to collaboratively discuss their broad experience in safety. The group combination of airspace users, including pilots, ATCOs, FDM safety analysts, and safety authorities representatives provided a very inspiring and clear overview of present-day aviation safety analysis, its challenges and opportunities on the transition from event-driven to data-driven safety intelligence. These meetings provide critical insight for data scientists and are key to support the users-driven approach adopted for the project since its conception. The users have defined relevant safety scenarios where data science and ML techniques can provide an added value over the incident-analysis tools they currently have. The scenarios, Runway performance, unstable approaches, group proximity, and airprox drive the descriptive and predictive analytics for The consortium meetings are an important to present results from the data analysis and discuss and capture their requirements (both individually and in groups) for future work. As the final users of the data analysis work performed within, it is key to ensure this alignment so their visualization dashboards provides relevant and usable ML tools.

SafeClouds is currently immersed in running the analytics based on three years of FDM data, which is merged with traffic data from Eurocontrol, weather data and surface radar data, among other data sources as required by the use case. This comes after investing the first months of the project to develop the legal and technical framework for securely managing and protecting the data. Considering this, the DataBeacon development, a data infrastructure that through several security layers and applying innovative cryptographic techniques, enables the data protection and merging while preserving its confidentiality. This Aviation ML platform, and the different implemented features and applications, enables data analysts to perform their analysis over various aviation data sources without actually having access to the databases. In all, this provides the necessary level of trust to the users and data owners.

With these developments, is going one step further by providing breakthrough analytics on safety precursors based on ML techniques. This analysis will combine airline FDM data with traffic, ADS-B and METEO data, providing improved information on the scenario that individual airspace users cannot otherwise access. This provides airlines, ANSPs and airports an enhanced understanding on the main causes that influence a safety incident which can support decision making for developing customized mitigation actions. Interested in more details on the techniques and results? A follow-up post will be published soon.

Aviation, data analytics, data infrastructures, Data Science, Safety