Artificial intelligence (AI) has the potential to tackle the two major challenges aviation faces today: reducing the environmental impact of air traffic and improving the ability of the network to scale capacity up or down depending on growth or decline of traffic.
At EUROCONTROL we developed a number of AI based applications and are working on more to provide enhanced air traffic management (ATM) performance and new digital services for the EUROCONTROL Network Manager (NM) and our operational stakeholders.
AI enables more accurate predictions and more sophisticated tools to increase productivity improve decision making, enhance use of scarce resources (e.g. airspace, runways) and increase human performance.
EUROCONTROL is fully committed to support the acceleration of AI adoption in European aviation, and more specifically in air traffic management, through the following actions:
- the FLY AI initiative, a coordinated action of European aviation/ATM actors to demystify and accelerate the uptake of AI;
- support to EUROCAE and EASA for the development of AI standards and guidelines for aviation/ATM;
- the development of AI trainings;
- the development of AI-based ATM applications notably through research and innovation actions together with a suitable data and AI infrastructure framework;
- The deployment of AI-based ATM applications at Network Level to be used by the whole aviation.
At EUROCONTROL, we have developed a number of AI based applications enhancing flight planning, traffic predictions and forecast, trajectory optimisation, airport operations, GNSS operations using artificial intelligence (AI) machine learning.
We apply AI to support our stakeholders and make their operations more efficient and predictable. We are developing performance prediction tools to help predict aircraft turnaround times. AI is also powering our forecasting efforts through a set of highly capable new tools in their predictive analysis toolbox. For more details, see our projects below.
More than thirty AI-based applications are under development in different frameworks notably in the Network of Innovation Labs and SESAR.
Data and strong data engineering are essential enablers for AI. Gathering data from producers, storage, providing access to data users all require well-organised infrastructure and very strong data governance.
To support AI across all domains of aviation, we are evolving our PRISME data warehouse into a cloud based aviation data hub that includes an AI-platform. With this, our experts and stakeholders will increasingly benefit from the growth of good practice and expertise, the use of appropriate datasets and ever-improving data quality.
Moreover, we release large data sets to support artificial intelligence research and innovation needs.
AI solutions can come with unintended side effects. That’s why potential failures must be appropriately mitigated in the design phase and new processes, procedures and tools be developed to check algorithms and especially trained/implemented AI ‘solutions’ are integrated and maintained properly to and deliver the intended results.
At EUROCONTROL, to help accelerate the deployment of AI applications with high performance benefit for the network, we collaborate with EASA on the development of their guidance material for trustworthy AI. We contribute with our ATM knowledge, safety expertise and experience and most promising uses cases. .
We also participate in the international standardisation effort on AI for aviation led by EUROCAE (WG114 and SAE G34), notably regarding the provision of ATM-relevant use-cases and their validation against these new standards.
AI-driven projects and applications
At EUROCONTROL, thanks to our innovation labs (EUROCONTROL Air Transport Innovation Network, Network Manager Lab and MUAC Innovation lab), Artificial Intelligence (AI) is developing fast offering significant performance improvements in capacity, safety, security, environment, resilience, and cost efficiency as well as new services to our operational stakeholders.
More than 30 applications are currently in the research or fast track innovation pipeline. ATM domains addressed include flights forecasts, flight plans and trajectory predictions, optimisations of fleet sequences, conflict detection and resolution, airport operations and their integration in the network operations, CNS and cyber monitoring, Speech-To-Text transcription supporting a wide range of applications in ATC, training, simulations or even CNS and many more.
Some of these applications (AI/ML Based augmented 4D trajectory and the Calibration of optimised approach spacing tool with use of machine learning - COAST) supported the development of the EASA Concept paper and the SAE/EUROCAE aviation AI standard.
Our ATM operational systems already rely on several AI applications to support MUAC operations and the Network Manager tasks and functions.