LSTM‑Based Temporal Modeling for ADS‑B Trajectory Prediction: A Comparative Review of Approaches
Year: 2026
Type: article
Abstract: Recent advances in artificial intelligence and big data, particularly deep learning architectures such as convolutional neural networks, have enabled new capabilities for processing complex spatiotemporal data. Automatic Dependent Surveillance–Broadcast (ADS‑B) trajectory data, characterized by high‑dimensional spatiotemporal features and inherent uncertainty, present significant challenges for accurate prediction. This study focuses on temporal feature modeling for trajectory prediction, systematically examining various Long Short‑Term Memory (LSTM) network‑based approaches. By comparing different LSTM variants and their effectiveness in capturing temporal dependencies, the paper offers insights into model selection and identifies future research directions for enhancing prediction accuracy in air traffic surveillance and management systems.
Author: Chen, Ting
Institution: Civil Aviation Administration of China
Language: English
Subfield: Aerospace Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W7225100410