Research on Aircraft Trajectory Prediction Algorithm Based on Hybrid Neural Network Model

This paper presents a novel approach for aircraft trajectory prediction, focusing on the analysis of flight big data to extract spatiotemporal features from a large number of historical flight trajectories. A neural network prediction model is established to forecast aircraft flight trajectories. In the experiments, a Temporal Convolutional Network (TCN) is first employed to extract spatial features, followed by Long Short-Term Memory (LSTM) networks to capture temporal features of the trajectories. This forms the TCN-LSTM neural network model. Furthermore, an Attention Mechanism is incorporated to capture multi-level periodicity in the historical flight trajectories, aiming to achieve high-precision trajectory prediction. Experimental results on real historical trajectory data demonstrate that the TCN-LSTM-Attention hybrid model outperforms the individual LSTM and TCN models in terms of prediction accuracy.