Uploaded July 2024 | Updated September 2026, 5 hours ago
🛫 Validating the OpTT 2.0 machine-learning prediction model 🤖 at our Innovation Hub!
🕙 A replay validation exercise of the Optimisation of Turnaround Times - OpTT 2.0 model was recently conducted at the EUROCONTROL Innovation Hub with the participation of Swiss International Air Lines, Prague, Dusseldorf, Brussels and Rome airports as well as Aéroports de Paris.
🖥️ The model provides machine-learning predictions of Target Off-Block Time (TOBT) values and turnaround durations for all European CDM airports as well as airlines operating at these airports. These predictions could help improve airport and airline operational planning.
🔎 EUROCONTROL historical operational data was used during the trial to test the model in a replay mode. As a main outcome of the exercise, it was demonstrated that, throughout the year, 66% of TOBT predictions were very accurate and did not require any TOBT update by an operational user.
Furthermore, this percentage was 68% on a nominal day and 61% on a day affected by heavy strikes in Europe. As a next step, the model will be tested using live data.
The project is part of EATIN, our user-driven innovation portfolio: https://www.eurocontrol.int/project/eatin
🛫 Validating the OpTT 2.0 machine-learning prediction model 🤖 at our Innovation Hub!
🕙 A replay validation exercise of the Optimisation of Turnaround Times - OpTT 2.0 model was recently conducted at the EUROCONTROL Innovation Hub with the participation of Swiss International Air Lines, Prague, Dusseldorf, Brussels and Rome airports as well as Aéroports de Paris.
🖥️ The model provides machine-learning predictions of Target Off-Block Time (TOBT) values and turnaround durations for all European CDM airports as well as airlines operating at these airports. These predictions could help improve airport and airline operational planning.
🔎 EUROCONTROL historical operational data was used during the trial to test the model in a replay mode. As a main outcome of the exercise, it was demonstrated that, throughout the year, 66% of TOBT predictions were very accurate and did not require any TOBT update by an operational user.
Furthermore, this percentage was 68% on a nominal day and 61% on a day affected by heavy strikes in Europe. As a next step, the model will be tested using live data.
The project is part of EATIN, our user-driven innovation portfolio: https://www.eurocontrol.int/project/eatin
