Data-driven methodology for uncertainty quantification of aircraft trajectory predictions

One of the main objectives of the so-called trajectory-based operations (TBO) concept is to increase the predictability of the aircraft behavior within the air traffic management (ATM) system, thus reducing inefficiencies and increasing the robustness and resiliency of operations. Most systems involved in TBO, such as flight planning systems or on-ground trajectory predictors, focus on proposing deterministic trajectories in the strategic phase and do not take into account the uncertain factors that affect the trajectory prediction process. While TBO is enabled by the automated updating of trajectories in reaction to developing uncertainties, an excessive frequency of trajectory updates in later planning and tactical phases could lead to degraded stability, resulting in an overall decrease of the performance of the ATM network. The use of probabilistic trajectories instead of deterministic ones would allow to reduce the frequency of these updates, as well as increasing to increase the situational awareness of the ATM system. Nonetheless, the analysis of the uncertainty affecting the prediction of a flight is a complex problem that has been tackled in the literature. The main difficulty regarding aircraft trajectory uncertainty is that it cannot be estimated in a post-processing study based on the comparison between the predicted and the actual trajectories. This is because the uncertainty is represented by the estimation of those potential deviations in an a priori phase, based on the identification and quantification of the possible sources impacting that uncertainty and the propagation of the joint effect of those sources to obtain the probability distribution describing the potential trajectory