Unmanned aerial vehicle detection method based on frequency modulation continuous wave radar

By combining FM continuous wave radar and deep learning technology in UAV detection, micro Doppler and deep features are extracted and convolutional neural network is constructed for detection, the accuracy and threshold setting difficulties of traditional methods and deep learning methods are solved. The unexplainability problem has improved the accuracy and robustness of drone detection.

CN119179055APending Publication Date: 2024-12-24JILIN UNIVERSITY
View PDF 0 Cites 3 Cited by

Patent Information

Application Number
CN202411333424.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-24
Publication Date
2024-12-24

AI Technical Summary

Technical Problem

Traditional drone detection methods rely on the accuracy of statistical models and difficulty in setting detection thresholds. Deep learning methods require a lot of data training and the detection process is unexplainable, resulting in low accuracy, robustness and reliability of drone detection.

Method used

Using a FM continuous wave radar method, the micro Doppler features and deep features of the UAV echo signal are extracted, and the convolutional neural network is used for feature fusion and detection, and the feature fusion unit and no-based feature based on attention mechanism are constructed. Human-computer detection network realizes dynamic weighted feature fusion and detection result output.

Benefits of technology

It significantly improves the detection accuracy of low-slow small drones, combines the interpretability of traditional methods and the automation characteristics of deep learning, and improves the robustness and reliability of the system, and can accurately distinguish drones in complex environments. Especially when other motor targets are disturbed.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119179055A_ABST
    Figure CN119179055A_ABST
Patent Text Reader

Abstract

The invention relates to an unmanned aerial vehicle detection method based on a frequency modulated continuous wave radar, and belongs to the field of radar technology and deep learning. The objective of the invention is to solve the problems of low detection accuracy, low robustness and low reliability of an existing unmanned aerial vehicle. The method comprises the following steps: obtaining a trained new feature extractor, a feature fusion unit and an unmanned aerial vehicle detection network; extracting the echo signal of the unmanned aerial vehicle from the echo signal to be detected, and performing micro-Doppler feature extraction to obtain micro-Doppler features; inputting a to-be-detected echo signal into the trained feature extractor, and outputting deep features by the trained feature extractor; based on a trained feature fusion unit, performing feature fusion on the micro-Doppler features and the deep features to obtain dynamically weighted fusion features; and inputting the dynamically weighted fusion features into the trained unmanned aerial vehicle detection network, and outputting a detection result by the trained unmanned aerial vehicle detection network.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Cited By

  • Small unmanned aerial vehicle identification method and device based on micro-Doppler feature enhancement

    CN119471624A

  • Unmanned aerial vehicle investigation countering method based on deep learning

    CN119814219A

  • Multi-band external radiation source radar unmanned aerial vehicle detection method and system

    CN121596269A