A Machine Learning Approach to Evaluating Aircraft Deviations from Planned Routes
M.K. Sakyi-Gyinae (TU Delft - Aerospace Engineering)
J. Ellerbroek – Mentor (TU Delft - Control & Simulation)
J.M. Hoekstra – Mentor (TU Delft - Control & Operations)
I. Rudnyk – Mentor (TU Delft - Control & Simulation)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
With the increasing trend in air traffic demand and evidence of large deviations from filed flight plans, airspace capacity is not being optimally utilized. In order to improve air traffic flow and capacity management systems, so that air traffic control operators can handle more aircraft safely, air traffic predictability needs to be improved. Quantitatively, this means reducing the measure of spread of aircraft deviations from filed flight plans. In this thesis, a long short term memory (LSTM) network is proposed to predict trajectories in Maastricht upper airspace in a data-driven approach, using statistical aircraft deviation related features. The results show that the LSTM model has a lower prediction error at predicting trajectories than the current model used by the network manager. The LSTM model finally demonstrates its application within air traffic demand optimization where the LSTM based sector load predictions provide a more accurate estimation than the filed flight plan based predicted occupancy count.