The development of flight trajectory models that can efficiently predict and simulate air traffic operations is an important step to support novel concepts in Air Traffic Management (ATM), such as Trajectory-Based Operations (TBO). This paper presents a data-driven approach to developing a high-fidelity trajectory model for the airspace surrounding an airport using machine learning. Historical aircraft tracking data for arrival flights at the Sao Paulo/Guarulhos International Airport (GRU) is used to learn actual trajectory patterns and a probabilistic model of aircraft motion in the extended terminal area. We apply and evaluate the model for both trajectory prediction and generation, envisioning potential online and offline applications. Predictive performance is assessed over different time horizons, considering different amounts of past trajectory information available at the time of the prediction. To evaluate whether the trajectory generation process outputs realistic aircraft trajectories, an expert evaluation process is performed. We find that the model is able to deliver accurate trajectory predictions, outperforming a baseline deterministic model, and to generate realistic patterns of aircraft movement in the airspace analyzed.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    A Data-Driven Probabilistic Trajectory Model for Predicting and Simulating Terminal Airspace Operations


    Beteiligte:


    Erscheinungsdatum :

    11.10.2020


    Format / Umfang :

    2572983 byte





    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch




    Learning Probabilistic Trajectory Models of Aircraft in Terminal Airspace from Position Data

    Barratt, Shane / Kochenderfer, Mykel / Boyd, Stephen | ArXiv | 2018

    Freier Zugriff

    Learning Probabilistic Trajectory Models of Aircraft in Terminal Airspace From Position Data

    Barratt, Shane T. / Kochenderfer, Mykel J. / Boyd, Stephen P. | IEEE | 2019