Scholarlink

<p>Internet of Things (IoT) and blockchain technologies are employed to solve security and privacy concerns. The linearity of conventional blockchains raises storage costs and restricts intrusion detection with the attack severity, which leads to scaling issues. Traditional intrusion detection systems struggle to handle large-scale, resource-constrained environments while ensuring accuracy and privacy. To address these challenges, a novel BLOCKchain Intrusion Detection using deep learning (BLOCK-ID) is proposed in this paper. The BLOCK-ID model leverages the Betweenness Centrality (BeCe) based Spiking Convolutional Neural Network based Bidirectional Gated Recurrent Unit (Spiking CNN-BiGRU) model for detecting the malicious data packets. The chaotic enhanced firefly algorithm has been used for selecting the optimal consensus node for maintaining a fault tolerant consistent blockchain model. The proposed BLOCK-ID model is implemented in NS2 and evaluated against CACVO, GuardianAI, and FLADEN in terms of accuracy, network lifetime, energy consumption (EC), packet delivery ratio (PDR), and security strength. Experimental outcomes show that the proposed model achieves 99.75% accuracy, whereas methods like CACVO, GuardianAI, and FLADEN achieves accuracies of 93.8%, 94.42%, and 96.9%. The proposed BLOCK-ID strategy performs better in terms of security by 5.85%, 4.9%, and 2.89% than the existing methods, including CACVO, GuardianAI, and FLADEN approaches respectively.</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

Enhanced Chaotic Based Firefly Optimization Scalable Blockchain in Attack Detection IoT Devices

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月8日

查看全文
  • Desayar Dhanya,
  • Ahilan A,
  • Ramya Devi M,
  • Mohammed Uveise S A

摘要

Internet of Things (IoT) and blockchain technologies are employed to solve security and privacy concerns. The linearity of conventional blockchains raises storage costs and restricts intrusion detection with the attack severity, which leads to scaling issues. Traditional intrusion detection systems struggle to handle large-scale, resource-constrained environments while ensuring accuracy and privacy. To address these challenges, a novel BLOCKchain Intrusion Detection using deep learning (BLOCK-ID) is proposed in this paper. The BLOCK-ID model leverages the Betweenness Centrality (BeCe) based Spiking Convolutional Neural Network based Bidirectional Gated Recurrent Unit (Spiking CNN-BiGRU) model for detecting the malicious data packets. The chaotic enhanced firefly algorithm has been used for selecting the optimal consensus node for maintaining a fault tolerant consistent blockchain model. The proposed BLOCK-ID model is implemented in NS2 and evaluated against CACVO, GuardianAI, and FLADEN in terms of accuracy, network lifetime, energy consumption (EC), packet delivery ratio (PDR), and security strength. Experimental outcomes show that the proposed model achieves 99.75% accuracy, whereas methods like CACVO, GuardianAI, and FLADEN achieves accuracies of 93.8%, 94.42%, and 96.9%. The proposed BLOCK-ID strategy performs better in terms of security by 5.85%, 4.9%, and 2.89% than the existing methods, including CACVO, GuardianAI, and FLADEN approaches respectively.