EUVIMEDEUVIMED entdecken
Crossref

PFRTF: A Robust Training Framework to Counter Adversarial Attacks in Signal Classification for Next-G Consumer Electronics

Zhida Bao, Sicheng Zhang, Songlin Yang, Qiao Tian, Yun Lin

IEEE Transactions on Consumer Electronics · 2025-02 · 71 (1) · 1235-1248

DOI: 10.1109/tce.2024.3411668 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Consumer Electronics, 2025-02. Verfasst von Zhida Bao und 4 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 50 Literaturangaben. Gefördert von: National Natural Science Foundation of China, National Key Research and Development Program of China, Key Laboratory of Advanced Marine Communication and Information Technology, Ministry of Industry and Information Technology, Harbin Engineering University, Har….

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Adversarial attacks in wireless communications pose significant challenges to the reliability of next-generation (Next-G) consumer electronics. Attackers can cleverly modify signals to mislead deep learning-based signal classification models, compromising the integrity and confidentiality of the system. To address this, we propose a phase flipping-based robust training framework (PFRTF) to bolster adversarial robustness in signal classification. PFRTF employs phase flipping data augmentation to create diverse training samples and utilizes soft label training to support nuanced decision-making. A balanced adversarial training strategy is integrated to consider both original and adversarial losses, effectively mitigating overfitting to adversarial examples. The effectiveness of the PFRTF is validated through rigorous experiments on both open-source datasets and a high-quality, locally-collected dataset. The results demonstrate that the PFRTF-trained model significantly improves defense capabilities, outperforming the widely adopted PGD adversarial training method in the wireless domain with an 18.12% higher robust accuracy against the most potent optimization-based CW attacks and a 55.24% improvement over the baseline model. This research advances adversarial robustness in signal classification for Next-G consumer electronics and provides valuable insights into developing reliable model training architectures, contributing to safer wireless communication technologies.

Förderung

  • National Natural Science Foundation of China
  • National Key Research and Development Program of China
  • Key Laboratory of Advanced Marine Communication and Information Technology, Ministry of Industry and Information Technology, Harbin Engineering University, Har…

Zitieren

APA

Bao, Z., Zhang, S., Yang, S., Tian, Q., & Lin, Y. (2025). PFRTF: A Robust Training Framework to Counter Adversarial Attacks in Signal Classification for Next-G Consumer Electronics. IEEE Transactions on Consumer Electronics, 71(1), 1235-1248. https://doi.org/10.1109/tce.2024.3411668

Vancouver

Bao Z, Zhang S, Yang S, Tian Q, Lin Y. PFRTF: A Robust Training Framework to Counter Adversarial Attacks in Signal Classification for Next-G Consumer Electronics. IEEE Transactions on Consumer Electronics. 2025;71(1):1235-1248. doi:10.1109/tce.2024.3411668

BibTeX
@article{Bao2025,
  author = {Bao, Zhida and Zhang, Sicheng and Yang, Songlin and Tian, Qiao and Lin, Yun},
  title = {PFRTF: A Robust Training Framework to Counter Adversarial Attacks in Signal Classification for Next-G Consumer Electronics},
  journal = {IEEE Transactions on Consumer Electronics},
  year = {2025},
  volume = {71},
  number = {1},
  pages = {1235--1248},
  doi = {10.1109/tce.2024.3411668}
}
Kostenlos starten →