Scholarlink

<p>The incorporation of edge-based computing with underwater communication networks is becoming a promising solution for supporting increasing activities, like marine environmental monitoring and ocean resource exploration. However, it faces challenges in data transmission due to high energy consumption and the need for secure data transmission. The underwater environment leads to imbalanced energy consumption across sensor nodes, which underscores the urgent need to improve energy efficiency and balance consumption in underwater wireless sensor networks. To address these issues, a novel approach called Multi-access Edge Computing Internet of things -Underwater Sensor Networks (MECIUSN) has been proposed. This approach commences with the Distance, Energy constrained K-Means Clustering (DEKC) model for clustering. This also suggests integrating federated deep reinforcement learning with an automated trust-aware blockchain-assisted security mechanism with a framework that combines trust-aware secure routing and blockchain-based authentication. This also provides detection of malicious nodes, privacy-preserving federated learning, and energy forecasting to improve secure and energy-efficient underwater communication within MEC-enabled underwater sensor networks. Finally, the optimization algorithm, Hybrid Lion Grey Wolf optimization with Electric Electrophorus electricus larvae Fish Optimization Algorithm (HLGW-EEFOA), is applied for dynamic multipath routing. Additionally, the developed method achieves an effective energy consumption of 45&#xa0;J, 1.8-second latency, 96.5% delivery ratio, and 145 Mbps throughput, compared to any other existing techniques. This approach demonstrated superior performance compared to previously used methods and outperformed all existing techniques.</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

Optimized federated deep reinforcement learning with trust-aware blockchain security in underwater IoT networks

  • 期刊论文

  • 发表日期：

    2026年7月1日

查看全文
  • G. A. Senthil,
  • R. Renuka Devi,
  • R. Prabha,
  • L. Sharmila

摘要

The incorporation of edge-based computing with underwater communication networks is becoming a promising solution for supporting increasing activities, like marine environmental monitoring and ocean resource exploration. However, it faces challenges in data transmission due to high energy consumption and the need for secure data transmission. The underwater environment leads to imbalanced energy consumption across sensor nodes, which underscores the urgent need to improve energy efficiency and balance consumption in underwater wireless sensor networks. To address these issues, a novel approach called Multi-access Edge Computing Internet of things -Underwater Sensor Networks (MECIUSN) has been proposed. This approach commences with the Distance, Energy constrained K-Means Clustering (DEKC) model for clustering. This also suggests integrating federated deep reinforcement learning with an automated trust-aware blockchain-assisted security mechanism with a framework that combines trust-aware secure routing and blockchain-based authentication. This also provides detection of malicious nodes, privacy-preserving federated learning, and energy forecasting to improve secure and energy-efficient underwater communication within MEC-enabled underwater sensor networks. Finally, the optimization algorithm, Hybrid Lion Grey Wolf optimization with Electric Electrophorus electricus larvae Fish Optimization Algorithm (HLGW-EEFOA), is applied for dynamic multipath routing. Additionally, the developed method achieves an effective energy consumption of 45 J, 1.8-second latency, 96.5% delivery ratio, and 145 Mbps throughput, compared to any other existing techniques. This approach demonstrated superior performance compared to previously used methods and outperformed all existing techniques.