Scholarlink

<p>Machine learning (ML) methods are commonly used in optical communication networks to enhance bandwidth analysis through efficient resource management and performance monitoring, and to bolster cybersecurity through threat detection and proactive failure management. This research proposes a novel technique for bandwidth analysis of optical communication network signals, with cybersecurity modelling using frequency modulation and machine learning. Here, the optical communication network signal bandwidth is analysed using a dynamic frequency-modulated support vector machine algorithm. Network cybersecurity analysis is carried out using blockchain-reinforced adversarial Bayesian neural networks (NNs). Experimental analysis is based on signal-based cybersecurity modelling and bandwidth analysis. The proposed technique achieves more accurate bandwidth estimation with an error of 3%, a throughput of 97 Gbps, and a spectral efficiency of 4.82 b/s/Hz. By ensuring secure and efficient optical network operation, it achieves 98% attack-detection accuracy with 1.5% false positives in terms of 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

Optical communication network signal analysis and cyber security modelling by frequency modulation with machine learning algorithm

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月18日

查看全文
  • Ramesh Kumar Ayyasamy,
  • Baddepaka Prasad,
  • Chinnasamy Ponnusamy,
  • Srikanth Cherukuvada,
  • N. V. S. M. Reddy,
  • Sk. Khaja Shareef,
  • Michael Melese Woldeyohannis

摘要

Machine learning (ML) methods are commonly used in optical communication networks to enhance bandwidth analysis through efficient resource management and performance monitoring, and to bolster cybersecurity through threat detection and proactive failure management. This research proposes a novel technique for bandwidth analysis of optical communication network signals, with cybersecurity modelling using frequency modulation and machine learning. Here, the optical communication network signal bandwidth is analysed using a dynamic frequency-modulated support vector machine algorithm. Network cybersecurity analysis is carried out using blockchain-reinforced adversarial Bayesian neural networks (NNs). Experimental analysis is based on signal-based cybersecurity modelling and bandwidth analysis. The proposed technique achieves more accurate bandwidth estimation with an error of 3%, a throughput of 97 Gbps, and a spectral efficiency of 4.82 b/s/Hz. By ensuring secure and efficient optical network operation, it achieves 98% attack-detection accuracy with 1.5% false positives in terms of security.