A Non-autoregressive Multi-Horizon Flight Trajectory Prediction Framework with Gray Code Representation

TL;DR

This work targets multi-horizon flight trajectory prediction by moving beyond autoregressive, binary-encoded approaches. It introduces FlightBERT++, a non-autoregressive framework that uses Gray Code representation and differential prediction, guided by a horizon-aware context generator and a differential-prompted decoder. The architecture combines a Trajectory Encoder, HACG, and DPD to produce direct multi-horizon outputs with improved accuracy and efficiency on real ATC data, outperforming seven baselines. Empirical results show significant reductions in mean deviation error (MDE) and faster inference, highlighting its suitability for real-time ATC applications such as conflict detection and traffic-flow management.

Abstract

Flight Trajectory Prediction (FTP) is an essential task in Air Traffic Control (ATC), which can assist air traffic controllers in managing airspace more safely and efficiently. Existing approaches generally perform multi-horizon FTP tasks in an autoregressive manner, thereby suffering from error accumulation and low-efficiency problems. In this paper, a novel framework, called FlightBERT++, is proposed to i) forecast multi-horizon flight trajectories directly in a non-autoregressive way, and ii) improve the limitation of the binary encoding (BE) representation in the FlightBERT framework. Specifically, the proposed framework is implemented by a generalized encoder-decoder architecture, in which the encoder learns the temporal-spatial patterns from historical observations and the decoder predicts the flight status for the future horizons. Compared to conventional architecture, an innovative horizon-aware contexts generator is dedicatedly designed to consider the prior horizon information, which further enables non-autoregressive multi-horizon prediction. Additionally, the Gray code representation and the differential prediction paradigm are designed to cope with the high-bit misclassifications of the BE representation, which significantly reduces the outliers in the predictions. Moreover, a differential prompted decoder is proposed to enhance the capability of the differential predictions by leveraging the stationarity of the differential sequence. Extensive experiments are conducted to validate the proposed framework on a real-world flight trajectory dataset. The experimental results demonstrated that the proposed framework outperformed the competitive baselines in both FTP performance and computational efficiency.

Figures (10)

• Figure 1: Comparison of the BE, GC, and Differential GC representations. The examples of bit state changes are marked by the red rectangles.
• Figure 2: Overview of the proposed FlightBERT++ framework.
• Figure 3: The detailed implementation of the channel-mix trajectory point embedding and trajectory encoder.
• Figure 4: The detailed implementation of the Differential-prompted Decoder.
• Figure 5: The comparison of the MDE curve over the 1 to 15 prediction horizons for different models.