Supporting air traffic controllers near airports, TADA is an AI tool designed to optimise the efficiency, safety, and sustainability of arrivals and departures. Deep Blue is contributing to the development and evaluation of the tool.
TMAs: many flights, plenty of data to use
Traffic in the areas around airports, known as Terminal Maneuvring Areas (TMAs), is among the most challenging for air traffic controllers to manage. ‘Terminal areas, especially those serving multiple airports, are often congested and pose significant operational challenges,” explains Ana Ferreira, a Human Factors expert and consultant at Deep Blue. “A TMA controller must identify flights and organise their arrival and departure sequences, relying on a set of digital tools that assist in gathering and analysing information”. These tools include AMAN, a system that supports landing sequence management, as well as others for trajectory management, safety, and compliance with operational instructions.
In their daily work, approach controllers analyse large amounts of data in real time, make operational decisions and constantly update the ATM system with the latest information. However, this data is only partially or immediately used. But what if it could be used to train an algorithm capable of predicting future scenarios and suggesting operational solutions to controllers?
TADA: the AI that supports air traffic controllers
This is where TADA (Terminal Airspace Digital Assistant) comes in: a system based on Artificial Intelligence which uses machine learning algorithms to analyse and identify recurring patterns in air traffic. It provides decision-making support to air traffic controllers (ATCOs). These features can help reduce controllers’ workload, which is an increasingly pressing need. According to EUROCONTROL estimates, European air traffic is expected to grow by 2.5% each year until 2030, although there will be different trends across countries.
To develop and test the system, researchers will use operational data from the Milan airspace, which includes three major airports: Linate, Malpensa and Bergamo. The aim is to create a digital assistant that can support controllers in managing TMA traffic in real time, thereby enhancing operational efficiency and smoothness without compromising safety or environmental sustainability. “Existing tools like AMAN already provide valuable support,” Ferreira continues, “but TADA represents a step forward as it is one of the first tools designed to bring artificial intelligence directly into the hands of controllers during the terminal phase of flight”.
A tool tailored to controllers
Within the project, Deep Blue plays a cross-cutting role. Initially, it collects the requirements needed for system design. Then, it defines the human–machine collaboration framework. Finally, it contributes to the design of the user interface and the final validation of the tool. All of this work is carried out in close cooperation with air traffic controllers, who are involved from the earliest stages of the project to ensure that the tool truly meets their operational needs. The chosen test site is the Milan terminal manoeuvring area, which is a complex operational environment due to factors such as traffic intensity, physical and operational constraints (e.g. mountains and military zones), adverse weather conditions and the need to reduce CO₂ emissions in a densely populated area served by multiple airports.
The TADA system is being developed in accordance with human-centred design principles and EASA guidelines on the use of AI in aviation to ensure transparency and explainability in its decision-making processes. “Our expertise in Human Factors will be essential to designing a clear, intuitive and user-friendly interface,” Ferreira explains. ‘We have already started collecting requirements with ENAV controllers, and together with our partner MONAD, we have begun translating them into the graphical and functional elements of the interface”.
Project phases
In November, the first “low-fidelity” simulation will take place, during which controllers will interact exclusively with the system’s prototype interface. This will be followed, in the spring of next year, by the final phase of the project: a real-time simulation where the tool will be tested in realistic operational scenarios. On that occasion, controllers will provide detailed feedback both on the quality of the interface and on the reliability of the system’s recommendations. “The results will be compared with those obtained by colleagues working without TADA,” Ferreira concludes, “to verify whether the tool truly provides added value. We believe it does, and we are ready to prove it”.
