Machine Learning Supporting Enhanced Optimized Spacing Delivery between Consecutive Departing Aircraft

The Optimised Spacing Delivery (further referred to as OSD) tool has the objective of calculating the necessary time spacing between two consecutive departing aircraft in order to fulfil all required spacing and separation constraints. OSD, developed in SESAR 2020 Wave 1 [1] is based on analytical models [2] to predict aircraft trajectory and speed profiles. The use of this tool by Air Traffic Controller supports the safe, consistent and efficient delivery of the required separation or spacing between consecutive departure pairs by providing the time required between departure aircraft pairs via an automated count-down timer to the tower runway controller. In order to improve OSD, this paper introduces the enhanced Optimised Spacing Delivery (further referred to as eOSD) tool which builds on the OSD tool using Machine Learning techniques to make more accurate predictions of aircraft behaviour (e.g. trajectory/climb profile, speed profile) and wind on the initial departure path, so further optimising spacing delivery between consecutive departures. Zurich airport data were used to develop and asses the performance of the eOSD tool compared to the OSD tool.