Scholarlink

<p>The growing complexity of cyber threats presents significant challenges to the security of Internet of Medical Things (IoMT) systems, where traditional security and intrusion detection methods often prove inadequate. The Key challenges include inefficient key management, fragmented security protocols, and limited scalability. To address these issues, this paper proposes a Dynamic Adaptive Deep Reinforcement Learning (DA-DRL) framework that enhances Advanced Encryption Standard (AES) encryption by dynamically adjusting key generation in response to real-time threats. Additionally, a multi-layered security architecture integrating AES, SHA-512, Non-Interactive Zero Knowledge Proof (NIZKPs), Practical Byzantine Fault Tolerance (PBFT), and Attribute-Based Access Control (ABAC) is introduced, ensuring robust protection against diverse attack vectors. The InterPlanetary File System (IPFS) is employed for decentralized and immutable data storage, enhancing data security and transparency. The proposed DA-DRL-AES-SHA-512 methodology significantly outperforms conventional encryption techniques, achieving an encryption time of 0.0975&#xa0;s, decryption time of 0.0846&#xa0;s, and a throughput of 75.63 transactions per second (Tx/s) with a network overhead of just 0.1289%. The Energy consumption and computational overhead are reduced to 0.3664&#xa0;J and 0.48%, respectively. The Secure and Dependable Bi-LSTM GRU Intrusion Detection Framework (S-BiLSTMGRU-IDF) achieves 99.94% accuracy in binary classification and 99.89% in multiclass classification, improving detection efficiency by 0.6–3.5% over state-of-the-art models. This blockchain-based framework ensures real-time threat mitigation, enhanced data integrity, and superior system performance, establishing a secure, scalable, and efficient solution for IoMT 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

Multi-layered security architecture for IoMT systems: integrating dynamic key management, decentralized storage, and dependable intrusion detection framework

  • 期刊论文

  • 发表日期：

    2025年5月3日

查看全文
  • Nikhil Sharma,
  • Prashant Giridhar Shambharkar

摘要

The growing complexity of cyber threats presents significant challenges to the security of Internet of Medical Things (IoMT) systems, where traditional security and intrusion detection methods often prove inadequate. The Key challenges include inefficient key management, fragmented security protocols, and limited scalability. To address these issues, this paper proposes a Dynamic Adaptive Deep Reinforcement Learning (DA-DRL) framework that enhances Advanced Encryption Standard (AES) encryption by dynamically adjusting key generation in response to real-time threats. Additionally, a multi-layered security architecture integrating AES, SHA-512, Non-Interactive Zero Knowledge Proof (NIZKPs), Practical Byzantine Fault Tolerance (PBFT), and Attribute-Based Access Control (ABAC) is introduced, ensuring robust protection against diverse attack vectors. The InterPlanetary File System (IPFS) is employed for decentralized and immutable data storage, enhancing data security and transparency. The proposed DA-DRL-AES-SHA-512 methodology significantly outperforms conventional encryption techniques, achieving an encryption time of 0.0975 s, decryption time of 0.0846 s, and a throughput of 75.63 transactions per second (Tx/s) with a network overhead of just 0.1289%. The Energy consumption and computational overhead are reduced to 0.3664 J and 0.48%, respectively. The Secure and Dependable Bi-LSTM GRU Intrusion Detection Framework (S-BiLSTMGRU-IDF) achieves 99.94% accuracy in binary classification and 99.89% in multiclass classification, improving detection efficiency by 0.6–3.5% over state-of-the-art models. This blockchain-based framework ensures real-time threat mitigation, enhanced data integrity, and superior system performance, establishing a secure, scalable, and efficient solution for IoMT security.