Synthetic Aircraft Trajectory Generation Using Time-Based VQ-VAE
Abdulmajid Murad, Massimiliano Ruocco
TL;DR
The paper presents an adaptation of TimeVQVAE to synthesize full-flight aircraft trajectories by operating in the time-frequency domain, using dual LF/HF encoders, vector quantization, and transformer priors to capture multi-scale spatiotemporal dynamics. A three-stage training pipeline (VQ-VAE training, prior learning with MaskGIT, and optional fidelity enhancement) enables efficient, coherent generation, with trajectory generation performed via a double-pass iterative decoding. The approach is evaluated with quality, statistical, and domain-specific flyability metrics, showing TimeVQVAE outperforming a temporal convolutional VAE baseline and producing realistic, diverse, and generally feasible trajectories, albeit with some limitations in cruise-phase variability and rare event representation. The work provides an open-source implementation and a robust evaluation framework, highlighting the method’s potential for data augmentation, airspace design, and forecasting, while outlining future extensions like physics-informed constraints and weather conditioning to further enhance realism and applicability.
Abstract
In modern air traffic management, generating synthetic flight trajectories has emerged as a promising solution for addressing data scarcity, protecting sensitive information, and supporting large-scale analyses. In this paper, we propose a novel method for trajectory synthesis by adapting the Time-Based Vector Quantized Variational Autoencoder (TimeVQVAE). Our approach leverages time-frequency domain processing, vector quantization, and transformer-based priors to capture both global and local dynamics in flight data. By discretizing the latent space and integrating transformer priors, the model learns long-range spatiotemporal dependencies and preserves coherence across entire flight paths. We evaluate the adapted TimeVQVAE using an extensive suite of quality, statistical, and distributional metrics, as well as a flyability assessment conducted in an open-source air traffic simulator. Results indicate that TimeVQVAE outperforms a temporal convolutional VAE baseline, generating synthetic trajectories that mirror real flight data in terms of spatial accuracy, temporal consistency, and statistical properties. Furthermore, the simulator-based assessment shows that most generated trajectories maintain operational feasibility, although occasional outliers underscore the potential need for additional domain-specific constraints. Overall, our findings underscore the importance of multi-scale representation learning for capturing complex flight behaviors and demonstrate the promise of TimeVQVAE in producing representative synthetic trajectories for downstream tasks such as model training, airspace design, and air traffic forecasting.