Probabilistic Constraints Prediction with Uncertainty Quantification in Trajectory Based Operations (TBO)

The SESAR research for 4D trajectory data exchanges between Air Traffic Management (ATM) actors enables Trajectory Based Operations (TBO) by introducing more accurate and comprehensive flight data sharing. However, differences in how various ATM actors consider Air Traffic Control (ATC) Letter of Agreement (LoA) constraints, before flight departure, leads to misalignment in trajectory calculations, impacting TBO implementation. To address this problem, as part of the SESAR 3 Network TBO project Solution 1, this paper presents a machine learning-based model predicting the probability for an ATC LoA constraint to be applied during flight. Additionally, an approach to quantifying the predictions’ uncertainty is developed, aiming at helping users identify potential issues with predictions. The model outcome will be subject to further assessment via operational validation scenarios in a prototyping environment since it should allow better trajectory alignment across ATM actors, facilitating smoother Flight and Flow – Information for a Collaborative Environment (FF-ICE) Release 1 deployment and advancing TBO in Europe.