Year: 2025
Type: article
Abstract: . While predicting human and vehicle trajectories is a deeply investigated field of research, predicting aircraft trajectories remains a less explored frontier. Still, the long-term prediction of aircraft movements is a fundamental challenge in aviation, influencing Air Traffic Management (ATM), operational efficiency, and flight safety. Traditional trajectory prediction models are often primarily focused on a 2D prediction.With this work, we evaluate different data representation methods in the field of long-term aircraft trajectory prediction using a state-of-the-art mobility prediction method, namely a CVAE-LSTM. We show that the H3 grid presents advantages for this task. With that, we explore a fascinating field of future mobility research, as the used data allows for various technical analyses without implying threats to personal privacy-relevant information.
Source: AGILE GIScience Series
Authors: Natália Soares, Paul Walther, Martin Werner
Institution: Technical University of Munich
Language: English
Subfield: Aerospace Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
Funder: Deutsche Forschungsgemeinschaft
OpenAlex ID: W4411134908