Scholarlink

<p>To tackle the ever-increasing security threats in IoT environments, this research aims to design a unique combination of MDI-QKD, blockchain, and a Deep Learning Intrusion Detection System that facilitates the secure distribution of cryptographic keys, provides an immutable mechanism for data verification, and detects cyber threats. The suggested architecture uses MDI-QKD at the gateway layer for safe cryptographic key generation, a permissioned blockchain with PBFT consensus for immutable transaction verification, and a CNN-based IDS for real-time traffic categorization using the N-BaIoT dataset. Secure communication, blockchain-verified metadata validation, and intelligent intrusion detection are included in the proposed framework, unlike current security methods. OMNeT++, INET, Simu5G, and TensorFlow/Keras experiments show 99.4% detection accuracy, 98.5% precision, 98.9% recall, a 1.2% false-positive rate, and an average detection latency of 120 ms. In 99.5% of communication sessions, the MDI-QKD module maintained a QBER &lt; 0.05, proving safe key exchange. Results show that the multi-layered architecture improves communication secrecy, data integrity, and threat detection in resource-constrained IoT scenarios.</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

A Hybrid MDI-QKD, Blockchain, and Deep Learning-Based Intrusion Detection Framework for Secure and Privacy-Preserving IoT Networks

  • 期刊论文

  • 发表日期：

    2026年7月20日

查看全文
  • Divya Priya Degala,
  • Senthil Athithan

摘要

To tackle the ever-increasing security threats in IoT environments, this research aims to design a unique combination of MDI-QKD, blockchain, and a Deep Learning Intrusion Detection System that facilitates the secure distribution of cryptographic keys, provides an immutable mechanism for data verification, and detects cyber threats. The suggested architecture uses MDI-QKD at the gateway layer for safe cryptographic key generation, a permissioned blockchain with PBFT consensus for immutable transaction verification, and a CNN-based IDS for real-time traffic categorization using the N-BaIoT dataset. Secure communication, blockchain-verified metadata validation, and intelligent intrusion detection are included in the proposed framework, unlike current security methods. OMNeT++, INET, Simu5G, and TensorFlow/Keras experiments show 99.4% detection accuracy, 98.5% precision, 98.9% recall, a 1.2% false-positive rate, and an average detection latency of 120 ms. In 99.5% of communication sessions, the MDI-QKD module maintained a QBER < 0.05, proving safe key exchange. Results show that the multi-layered architecture improves communication secrecy, data integrity, and threat detection in resource-constrained IoT scenarios.