Scholarlink

<p>Machine Learning (ML) with Physical Layer Security (PLS) is a promising technique for future 6G wireless systems. An active eavesdropper is an intelligent malicious attacker who not only listens to the communication channel but also actively intercepts and manipulates the information by pretending to be a legitimate user. Detection of such kind of attacker is particularly challenging in dynamic wireless environments. A novel Ensemble-Bagging-based ML with PLS approach is proposed to address this issue. The Wireless Characteristics Dataset (WCD) was created, capturing Channel State Information (CSI) for various subcarrier values, modulation schemes, and Signal-to-Noise Ratio (SNR) levels. The proposed Ensemble bagging with spectral clustering enhances feature differentiation and the attack detection is based on the average voting of decision trees. The simulated results demonstrate the attack detection accuracy and precision of the proposed technique is improved to 95% in dynamic wireless environment. The high Receiver Operating Characteristic (ROC) value of 0.98 of the proposed method indicates high True Positive Rate(TPR) and low False Positive Rates(FPR). Additionally, it offers reduced computational complexity compared to existing methods, making it a highly efficient solution for dynamic wireless security.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Ensemble bagging-based classifier for eavesdropper detection in 6G wireless networks

  • 期刊论文

  • 发表日期：

    2025年7月8日

查看全文
  • Swethambri Mohan,
  • Nandhini S,
  • Aparna H,
  • Gunaseelan K

摘要

Machine Learning (ML) with Physical Layer Security (PLS) is a promising technique for future 6G wireless systems. An active eavesdropper is an intelligent malicious attacker who not only listens to the communication channel but also actively intercepts and manipulates the information by pretending to be a legitimate user. Detection of such kind of attacker is particularly challenging in dynamic wireless environments. A novel Ensemble-Bagging-based ML with PLS approach is proposed to address this issue. The Wireless Characteristics Dataset (WCD) was created, capturing Channel State Information (CSI) for various subcarrier values, modulation schemes, and Signal-to-Noise Ratio (SNR) levels. The proposed Ensemble bagging with spectral clustering enhances feature differentiation and the attack detection is based on the average voting of decision trees. The simulated results demonstrate the attack detection accuracy and precision of the proposed technique is improved to 95% in dynamic wireless environment. The high Receiver Operating Characteristic (ROC) value of 0.98 of the proposed method indicates high True Positive Rate(TPR) and low False Positive Rates(FPR). Additionally, it offers reduced computational complexity compared to existing methods, making it a highly efficient solution for dynamic wireless security.