An Aircraft Trajectory Prediction Method Based on Trajectory Clustering and a Spatiotemporal Feature Network

The maneuvering characteristics and range of motion real aircraft are highly uncertain, which significantly increases the difficulty trajectory prediction. To solve problem that high-speed maneuvers excessive trajectories in airspace cause a decrease prediction accuracy to find out laws hidden large number trajectories, this paper proposes deep learning algorithm based on clustering spatiotemporal feature extraction, aims better describe regularity movement for higher accuracy. First, abnormal public dataset automatic dependent surveillance–broadcast (ADS-B) were analyzed, ensure uniform sampling data, cleaning interpolation data performed. Then, Hausdorff distance was used measure similarity between K-Medoids clustering, corresponding model established according results. Finally, extraction network constructed convolutional neural (CNN) bidirectional long short-term memory (BiLSTM) network, joint attention mechanism obtain important features points. A actual experiments showed proposed method is more accurate than existing algorithms BP, LSTM, CNN–LSTM models.