Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Signal Adversarial Examples Generation for Signal Detection Network via White-Box Attack

Dongyang Li, Linyuan Wang, Guangwei Xiong, Bin Yan, Dekui Ma, Jinxian Peng

TL;DR

A signal adversarial examples generation model for signal detection network from the perspective of adding perturbations to the signal using the inequality relationship of L2-norm between time domain and time-frequency domain to constrain the energy of signal perturbations is defined.

Abstract

With the development and application of deep learning in signal detection tasks, the vulnerability of neural networks to adversarial attacks has also become a security threat to signal detection networks. This paper defines a signal adversarial examples generation model for signal detection network from the perspective of adding perturbations to the signal. The model uses the inequality relationship of L2-norm between time domain and time-frequency domain to constrain the energy of signal perturbations. Building upon this model, we propose a method for generating signal adversarial examples utilizing gradient-based attacks and Short-Time Fourier Transform. The experimental results show that under the constraint of signal perturbation energy ratio less than 3%, our adversarial attack resulted in a 28.1% reduction in the mean Average Precision (mAP), a 24.7% reduction in recall, and a 30.4% reduction in precision of the signal detection network. Compared to random noise perturbation of equivalent intensity, our adversarial attack demonstrates a significant attack effect.

Signal Adversarial Examples Generation for Signal Detection Network via White-Box Attack

TL;DR

A signal adversarial examples generation model for signal detection network from the perspective of adding perturbations to the signal using the inequality relationship of L2-norm between time domain and time-frequency domain to constrain the energy of signal perturbations is defined.

Abstract

With the development and application of deep learning in signal detection tasks, the vulnerability of neural networks to adversarial attacks has also become a security threat to signal detection networks. This paper defines a signal adversarial examples generation model for signal detection network from the perspective of adding perturbations to the signal. The model uses the inequality relationship of L2-norm between time domain and time-frequency domain to constrain the energy of signal perturbations. Building upon this model, we propose a method for generating signal adversarial examples utilizing gradient-based attacks and Short-Time Fourier Transform. The experimental results show that under the constraint of signal perturbation energy ratio less than 3%, our adversarial attack resulted in a 28.1% reduction in the mean Average Precision (mAP), a 24.7% reduction in recall, and a 30.4% reduction in precision of the signal detection network. Compared to random noise perturbation of equivalent intensity, our adversarial attack demonstrates a significant attack effect.
Paper Structure (13 sections, 1 theorem, 22 equations, 4 figures, 2 tables, 1 algorithm)

This paper contains 13 sections, 1 theorem, 22 equations, 4 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Method
  3. Signal Adversarial Examples Generation Model
  4. Mathematical Framework for Signal Detection Networks
  5. Adversarial Examples for Signal Time-frequency Diagram
  6. Signal Adversarial Examples Generation Model
  7. White-Box Signal Adversarial Attack Method for Signal Detection Network
  8. Experiment
  9. Experimental Setting
  10. Error in Signal Time-Frequency Inversion and Selection of Perturbation L2-norm Ratios
  11. Signal Adversarial Attack Results
  12. Conclusion
  13. Proof for Theorem 1

Key Result

Theorem 1

Suppose ${{\delta }_{r}}$ is the perturbation vector added to the signal in the time domain, ${{\beta }_{r}}$ is the perturbation matrix added to the magnitude of the signal in the time-frequency domain, and $N$ is the length of the time window, the inequality relationship between the two is as foll

Figures (4)

  • Figure 1: Signal detection network structure
  • Figure 2: Signal time-frequency diagram and detection results
  • Figure 3: Schematic diagram of signal detection network adversarial examples' time-frequency conversion
  • Figure 4: Signal time-frequency diagram and detection results

Theorems & Definitions (1)

  • Theorem 1