Abstract

The 21st IEEE International Conference on Intelligent Transportation Systems, Maui - Hawaï, Etats-Unis, 04-/11/2018 - 07/11/2018; The continuous increase in air traffic is pressuring airport operators to augment their capacity while still complying with safety regulations. To achieve this goal, operators need a thorough knowledge of how the airport surface is put to use (i.e., how aircraft are routed) under various conditions. We propose an approach to characterizing traffic in an airport from radar-collected aircraft ground trajectories. The approach does not require prior knowledge of the airport's topology (since it reconstructs it from the available data) and uses network constrained trajectory clustering to produce groups of similar trajectories that represent frequently traveled routes in the airport. Our main findings from experimental results on real data from the Paris-Charles-de-Gaulle airport (France) indicate that the retrieved routes are highly linked to exploitation constraints (maneuver undertaken by the aircraft, runway, and configuration) and that the approach can help uncover the existence of less traveled alternative routes that can be useful for rerouting purposes.