journal article·2023-08-09Open access

Unsupervised Learning-Based Spectrum Sensing Algorithm with Defending Adversarial Attacks

Xinyu Li, Shaogang Dai, Zhijin Zhao

Applied Sciences· Volume 13· Issue 16

Abstract

Although the spectrum sensing algorithms based on deep learning have achieved remarkable detection performance, the sensing performance is easily affected by adversarial attacks due to the fragility of neural networks. Even slight adversarial perturbations lead to a sharp deterioration of the model detection performance. To enhance the defense capability of the spectrum sensing model against such attacks, an unsupervised learning-based spectrum sensing algorithm with defending adversarial attacks (USDAA) is proposed, which is divided into two stages: adversarial pre-training and fine-tuning. In the adversarial pre-training stage, encoders are used to extract the features of adversarial samples and clean samples, respectively, and then decoders are used to reconstruct the samples, and comparison loss and reconstruction loss are designed to optimize the network parameters. It can reduce the dependence of model training on labeled samples and improve the robustness of the model to attack perturbations. In the fine-tuning stage, a small number of adversarial samples are used to fine-tune the pre-trained encoder and classification layer to obtain the spectrum sensing defense model. The experimental results show that the USDAA algorithm is better than the denoising autoencoder and distillation defense algorithm (DAED) against FGSM and PGD adversarial attacks. The number of labeled samples used in USDAA is only 11% of the DAED. When the false alarm probability is 0.1 and the SNR is −10 dB, the detection probability of the USDAA algorithm for the fast gradient sign method (FGSM) and the projected gradient descent (PGD) attack samples with random perturbations is above 88%, while the detection probability of the DAED algorithm for both attack samples is lower than 69%. Additionally, the USDAA algorithm has better robustness to attack with unknown perturbations.
Source: crossref · Quality: baseline

Research context

OpenAlex topics: Wireless Signal Modulation Classification · Anomaly Detection Techniques and Applications · EEG and Brain-Computer Interfaces

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

Affiliations

School of Communication Engineering, Hangzhou Dianzi University, Hangzhou 310020, China

State Key Lab of Information Control Technology in Communication System of No. 36, Jiaxing 314000, China

Funding

National Natural Science Foundation of China

References (32)

Show source-provided references
1. Cognitive Radio Networks: The Spectrum Supply Chain Paradigm10.1109/TCCN.2015.2488627 ↗
2. Optimal spectrum sensing in energy harvesting cognitive radio systems10.1109/CC.2014.6880463 ↗
3. Robust deep sensing through transfer learning in cognitive radio10.1109/LWC.2019.2940579 ↗
4. Spectrum sensing based on deep learning classification for cognitive radios10.23919/JCC.2020.02.012 ↗
5. Spectrum sensing method of deep learning based on attention mechanism
6. Limited data spectrum sensing based on semi-supervised deep neural network10.1109/ACCESS.2021.3135568 ↗
7. Li, X., Zhao, Z., Zhang, Y., Zheng, S., and Dai, S. (2023). Spectrum sensing algorithm based on self-supervised contrast learning. Electronics, 12.10.3390/electronics12061317 ↗
8. Employing machine learning based malicious signal detection for cognitive radio networks10.1002/cpe.7457 ↗
9. Channel aware wireless body area network with cognitive radio technology in disaster cases10.1002/dac.4565 ↗
10. Tian, Q., Zhang, S., and Mao, S. (2022). Adversarial attacks and defenses for digital communication signals identification. Dig. Commun. Net., in press.10.1016/j.dcan.2022.10.010 ↗
11. Multiuser adversarial attack on deep learning for OFDM detection10.1109/LWC.2022.3207348 ↗
12. Denoising autoencoder for spectrum sensing adversarial defense model
13. Chen, R., and Park, J. (2006, January 25). Ensuring trustworthy spectrum sensing in cognitive radio networks. Proceedings of the 2006 1st IEEE Workshop on Networking Technologies for Software Defined Radio Networks, Reston, VA, USA.10.1109/SDR.2006.4286333 ↗
14. Cooperative spectrum sensing With m-ary quantized data in cognitive radio networks under SSDF attacks10.1109/TWC.2017.2707407 ↗
15. When attackers meet AI: Learning-empowered attacks in cooperative spectrum sensing10.1109/TMC.2020.3030061 ↗
16. Security in cognitive radio network: Defense against primary user emulation attacks using ge-netic artificial bee colony algorithm10.1016/j.future.2018.08.022 ↗
17. Defence against pue attacks in ad hoc cognitive radio networks: A mean field game approach10.1007/s11235-018-0472-y ↗
18. Luo, Z., and Lou, C. (2011, January 12–14). Specific primary user sensing for wireless security in ieee 802.22 network. Proceedings of the 2011 11th International Symposium on Communications and Information Technologies, Hangzhou, China.10.1109/ISCIT.2011.6089728 ↗
19. Defeating primary user emulation attacks using belief propagation in cognitive radio networks10.1109/JSAC.2012.121102 ↗
20. An SETM algorithm for combating SSDF attack in cognitive radio networks
21. Parhizgar, N., Jamshidi, A., and Setoodeh, P. (2022, January 17–19). Defense against spectrum sensing data falsification attack in cognitive radio networks using machine earning. Proceedings of the ICEE, Tehran, Iran.10.1109/ICEE55646.2022.9827418 ↗
22. Deep reinforcement learning based reliable spectrum sensing under SSDF attacks in cognitive radio networks10.1016/j.jnca.2022.103454 ↗
23. Reinforcement learning for performance improvement in cooperative spectrum sensing10.1016/j.phycom.2023.102112 ↗
24. Feinman, R., and Shintre, S. (2017). Detecting adversarial samples from artifacts. arXiv.
25. Adversarial deep learning for over-the-air spectrum poisoning attacks10.1109/TMC.2019.2950398 ↗
26. Primary user adversarial attacks on deep learning-based spectrum sensing and the defense method10.23919/JCC.2021.12.006 ↗
27. A deep ensemble-based wireless receiver architecture for mitigating adversarial attacks in automatic modulation classification10.1109/TCCN.2021.3114154 ↗
28. Goodfellow, J., Shlens, J., and Szegedy, C. (2014). Explaining and harnessing adversarial examples. arXiv.
29. Madry, A., Makelov, A., and Schmidt, L. (2017). Towards deep learning models resistant to adversarial attacks. arXiv.
30. Kingma, D., and Welling, M. (2013). Auto-encoding variational bayes. arXiv.
31. Rényi divergence and kullback-leibler divergence10.1109/TIT.2014.2320500 ↗
32. Shea, J., and West, N. (2016, January 26–30). Radio machine learning dataset generation with GNU radio. Proceedings of the GNU Radio Conference, Virtually.