A short wave frequency hopping signal rapid interception method based on deep neural network

By using a deep neural network-based method, the features of shortwave frequency hopping signals are directly extracted from the time-frequency diagram, solving the problem of long detection time in existing technologies. This enables rapid interception in complex electromagnetic environments and improves the reaction speed and accuracy of communication countermeasures systems.

CN120934557BActive Publication Date: 2026-06-23KUSN JIUHUA ELECTRONICS EQUIP FACTORY
View PDF 4 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
KUSN JIUHUA ELECTRONICS EQUIP FACTORY
Filing Date
2025-08-08
Publication Date
2026-06-23

AI Technical Summary

Technical Problem

Existing shortwave frequency hopping signal detection methods have long detection times in complex electromagnetic environments, making it difficult to achieve rapid interception. Especially in the case of low hopping speed and single-sideband voice signals, traditional methods require more than 10 seconds, which affects the reaction speed of communication countermeasures systems.

Method used

By employing a deep neural network-based approach, and through the construction of a training dataset and optimization of the Yolov5s network, the confidence level, hopping rate, and frequency range of the frequency hopping signal are directly extracted from the time-frequency plot. This avoids hopping point detection, achieves global detection, narrows down the target frequency band, and reduces the computational load of subsequent parameter estimation.

Benefits of technology

It significantly shortens the interference response time from 10 seconds to less than 1 second, improving the detection speed and accuracy in complex electromagnetic environments, especially in low-speed and single-sideband voice signal conditions, with a detection accuracy of over 97%.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120934557B_ABST
    Figure CN120934557B_ABST
Patent Text Reader

Abstract

The application discloses a kind of short wave frequency hopping signal fast interception method based on deep neural network, comprising the following steps: S1, model training stage: the training data set comprising short wave frequency hopping signal sample and corresponding category-frequency position label is constructed;Design deep neural network architecture, utilize training data set to optimize network weight, generate frequency hopping signal interception model;Including the following steps: S1.1, wideband IQ signal acquisition;S1.2, STFT conversion;S1.3, pseudo color conversion;S1.4, image compression;S1.5, data normalization;S1.6, network training;S2, real-time inference stage: the short wave IQ signal received is input into the signal interception model of training completion, obtains the result whether there is frequency hopping signal, hop speed and frequency range.Through the above mode, the present application realizes the real-time detection and parameter extraction of short wave narrowband frequency hopping signal by the combination of time-frequency conversion and deep learning, effectively solves the difficult problem of fast interception of short wave narrowband frequency hopping signal under complex electromagnetic background environment.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • A method for frequency hopping signal detection and parameter estimation based on deep neural networks

    CN111832462B

  • Short-wave frequency hopping signal sorting method based on improved YOLOv5

    CN116318249A

  • Industrial internet-oriented frequency hopping signal intelligent detection and identification method and system

    CN118760928A

  • Predicting visible / infrared band images using radar reflectance / backscatter images of a terrestrial region

    US20220335715A1