Year: 2025
Type: article
Abstract: To improve the safety of routes, reduce route delays, and shorten the latency time, 4D trajectory prediction is carried out based on deep learning, and a 4D trajectory prediction method for aircraft in point merge approach mode is proposed. This paper takes a certain airport as an example for research. Firstly, the acquired ADS-B data is preprocessed based on the principle of point merge. Secondly, the point merge data are screened out by dividing the route data. Finally, the two models of CNN and LSTM are used to carry out the simulation and results analysis. In the comparison of the prediction ability of the two models in terms of time, latitude, longitude, and altitude, the results of the research show that the deep learning-based trajectory prediction is feasible in point merge approach mode. Furthermore, compared with the CNN model, the LSTM model can better capture the dynamic changes of the aircraft and show higher accuracy in prediction.
Authors: Yuandi Zhao, Mingxuan Ren, Kefei Wu, Yixin Hu
Institution: Civil Aviation University of China
Language: English
Subfield: Aerospace Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4411680398