An aircraft flight trajectory prediction method based on frequency domain calibration and space-time fusion

By constructing a deep collaborative prediction system integrating space, time, and frequency domains, and utilizing frequency domain calibration and spatiotemporal fusion methods, the prediction error and trajectory deviation problems of existing flight trajectory prediction models are solved, achieving high-precision flight trajectory prediction.

CN122432629APending Publication Date: 2026-07-21SHENYANG AEROSPACE UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
SHENYANG AEROSPACE UNIVERSITY
Filing Date
2026-04-23
Publication Date
2026-07-21

AI Technical Summary

Technical Problem

Existing deep learning flight trajectory prediction methods lack a unified prediction method that deeply couples long-term temporal dynamics, spatial maneuvering features, and global frequency domain statistical information. This results in the model being unable to fully utilize the complementary advantages between heterogeneous features, making it difficult to effectively model the low-frequency overall flight trend in the trajectory, leading to prediction errors and trajectory deviations.

Method used

A deep collaborative prediction system integrating space, time, and frequency domains is constructed. Through frequency domain calibration and spatiotemporal fusion, using the Informer encoder and decoder, combined with spatial-frequency domain dual-branch feature extraction, dynamic weighted fusion, frequency domain calibration, and graph attention calculation, high-precision flight trajectory prediction is generated.

Benefits of technology

It significantly enhances the model's ability to characterize complex flight patterns, suppresses noise interference and trajectory deviation in long-term prediction, and improves the accuracy and robustness of flight trajectory prediction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122432629A_ABST
    Figure CN122432629A_ABST
Patent Text Reader

Abstract

The application provides an aircraft flight trajectory prediction method based on frequency domain calibration and space-time fusion, and relates to the technical field of aircraft flight trajectory prediction. The method first extracts local maneuvering features and global frequency domain features of trajectory data in parallel, and generates an initial embedding representation; then captures the time sequence dynamic evolution law of the long sequence trajectory through the Informer encoder, learns the time sequence dynamic change of the trajectory; at the same time, the time-space probability sparse cross attention is used to further fuse the space features with the time sequence evolution features; finally, the improved graph attention is used to calibrate the space-time fusion features in the frequency domain, calculate the cosine similarity of the frequency domain features, and distribute the weights of the nodes, so that the predicted trajectory conforms to the global flight trend. The method can deeply fuse the complementary advantages of time domain, space and frequency domain, ensure the response to local maneuvering at the same time, correct the trajectory deviation by using global frequency domain information, and make the predicted trajectory more accurate.
Need to check novelty before this filing date? Find Prior Art