On the 3rd of September 2021, the ARTIMATION Project attended the 11th EASN International Conference on Innovation in Aviation & Space to the Satisfaction of the European Citizens, hosted online.  


ATCOs are currently supported by several technologies and automations. However, nowadays automation systems do not need to provide additional information on top of the Data Processing result. Therefore, the outcome is a “Black Box”, where solutions are not explained to the human actor. With the adoption of Artificial Intelligence (AI), this “Black Box” system may not be sufficient anymore; and for different reasons, such as, reliability, data feed biases, errors, type of algorithm chosen, liability and reliability for certifications, and so on. ARTIMATION’s goal is to provide a transparent and explainable AI model (XAI), which produces an understandable outcome through an understandable process, to put the human in the loop again; in other words, to provide a “White Box”.  

AI in the ATM field: our presentation at the EASN Conference

After presenting the Consortium and the Advisory Board, our speaker Daniele Ruscio provided an overview of AI challenges in ATM, presenting the results of the State-of-Art review of the past ten year of research in the field. ARTIMATION performed this State-of-Art by reviewing several papers from different conferences and journals; mostly from the ICRAT conference, ATM seminar events, and the Transportation part C journal. After a first selection, we fully reviewed the papers. The results of this review led ARTIMATION to identify four main categories of ATM that could best connect to AI in general. These represent the main trends of AI in ATM:
  • Prediction: AI foresees the future behaviour of an “Object”.
  • Optimisation/Automation: AI enhances the behaviour of an “Object”.
  • Analysis: AI seeks to understand the past/observed behaviour.
  • Modelling/Simulation: AI simulates the air traffic airspace.
However, ARTIMATION found out that research does not equally cover all the 4 categories; indeed, very few research exists on categories like “Analysis” and “Optimization” that represent key enablers for allowing XAI. Therefore, starting from those results, the roadmap to support ATM tasks with XAI will need to involve 4 steps:
  • Prioritize ATM tasks that could benefit from XAI (see the first ARTIMATION workshop).
  • Find the best algorithms and related visualisation tools to produce the best XAI for the prioritized tasks.
  • Find the best ways to build acceptance and trust.
  • Refine the ARTIMATION prototype.

The “State-of-Art” of AI support in ATM

The ARTIMATION Project will soon disseminate its report on the “State-of-Art of AI support in ATM”, while proceeding to determine the best algorithms and visualisation tools that will be tested in the validation activities. Stay tuned!