Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A radio signal forgery method for combating deep learning

A technology of radio signal and deep learning, applied in neural learning methods, interference to communication, secure communication, etc., can solve problems such as easy detection, large visual difference of fake signal signal, inability to directly transfer signal field, etc., to reduce accuracy Effect

Active Publication Date: 2022-07-12
XIDIAN UNIV
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, due to the large visual difference between the fake signal and the original signal, it is easy to be detected, so the existing methods cannot be directly transferred to the signal domain.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A radio signal forgery method for combating deep learning
  • A radio signal forgery method for combating deep learning
  • A radio signal forgery method for combating deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0028] In electronic countermeasures, with the continuous improvement of the accuracy of the signal classifier based on deep learning, the classification accuracy of the radio signal modulation method by the enemy is getting higher and higher, and the communication security is threatened, and the existing data forgery methods against deep learning , mainly used to counter the image classifier in the field of computer vision. When it is used for forging signals, there will be a significant visual difference between the forged signal and the original radio signal. The enemy can recognize the existence of the forged signal with the naked eye, which reduces the deception of the forged signal. properties, therefore, cannot be directly transferred for adversarial radio signal classifiers. From the perspective of eliminating the visual difference between the forged signal and the original radio signal, the present invention studies a radio signal forgery method for combating deep lear...

Embodiment 2

[0054]The overall technical scheme of the radio signal forgery method for combating deep learning is the same as that of Embodiment 1. The visual restriction of the countermeasure signal generation described in step (5) is applied to the differential evolution algorithm, and the signal noise scheme set N' is obtained. The visual constraints of adversarial signal generation, and the visual constraints are used in the differential evolution algorithm, by adopting a deep learning-based signal classifier M to evaluate the noise scheme, and finally generate a final generation population T' representing the set of signal-noise schemes, see figure 2 , the specific implementation steps are:

[0055] (5a) Generating offspring populations with visual constraints Applying the proposed visual constraints of adversarial signal generation to a differential evolution algorithm to generate offspring populations Generate the daughter population of the differential evolution algorithm for t...

Embodiment 3

[0071] The overall technical scheme of the radio signal forgery method for combating deep learning is the same as that in Embodiment 1-2. The evaluation noise scheme described in step (5b) is to use a deep learning-based signal classifier M to calculate T and The evaluation score of the medium noise scheme, the implementation steps include:

[0072] (5b1) Convert the interference scheme into candidate adversarial signals: convert the current parent population T and the offspring population The jamming scheme in is transformed into a candidate adversarial signal set X'=[x 1 ',...,x k ',...,x 800 '], where 1≤k≤800, is the kth candidate adversarial signal, The conversion rules are:

[0073]

[0074]

[0075] (5b2) Obtaining the evaluation score: Input each candidate adversarial signal in the candidate fake signal set into the deep learning-based signal classifier M to be confronted, as the evaluation score of the corresponding interference scheme:

[0076] Score ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a radio signal forgery method for countering deep learning, which solves the technical problem that the deep learning-based signal classifier is difficult to counter in the technical field of artificial intelligence and electronic countermeasures. The implementation includes: randomly generating candidate adversarial signals for the modulated radio signals, using the candidate adversarial signals as the initial parent population, and generating a set of signal interference schemes through the differential evolution method based on visual constraints; using deep learning-based radio signal classifier evaluation to obtain Interfere with the signal and counter the signal to complete the counterfeiting of the counter signal. The forged signal of the present invention has extremely high similarity with the original radio signal, effectively counteracts the signal classifier based on deep learning, and reduces the classification accuracy of the modulation type of the radio signal. The forged signal of the present invention basically does not affect the understanding of the content of the radio signal when the interference signal is unknown. It is used in electronic countermeasures in the military field to prevent the type of radio signal modulation from being identified.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence and electronic countermeasures, and further relates to the forgery generation of radio signals, in particular to a radio signal forgery method specially used for combating deep learning. In electronic countermeasures, it can effectively interfere with the recognition of the modulation type of our signal by the enemy's deep neural network classifier, further concealing our own attempts and improving the survivability of the equipment. The invention can be used for application and research in the field of military communication security. Background technique [0002] With the development of deep learning, signal classifiers based on deep learning have been remarkably developed. For example, the accuracy rate of modulation classification of radio modulated signals with low signal-to-noise ratio has reached more than 90%. However, in electronic countermeasures, the improvement of the ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): H04K1/02H04K3/00G06N3/08G06K9/62G06N3/04
CPCH04K1/02H04K3/65G06N3/049G06N3/086G06N3/045G06F18/24
Inventor 杨淑媛马宏斌冯志玺王敏段韵章杨晨刘慧玲王能国孙泽培焦李成王翰林
Owner XIDIAN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com