Scholarlink

<p>Internet of Things-based Wireless Sensor Network (IoT-WSN) have developed as a core sensing infrastructure due to their potential to provide effective resource consumption and reliable service delivery in real-time applications. However, existing IoT-WSN routing methods exhibit some critical limitations including energy imbalance, inadequate Quality of Service (QoS) awareness, and poor robustness against routing attacks. To overcome these challenges, a novel QoS aware secure routing framework based on Gated Graph Neural Network (QAR-GGNN) is proposed to enhance energy balance across the network. The Gaussian mixture model (GMM) is used for efficient cluster formation, while a Mutated Spider-Tailed Horned Viper Optimization (M-STHVO) algorithm selects optimal cluster heads for maintaining energy consumption and prolong network lifetime. The deep learning-based GGNN is introduced to retrieve the graph-structured topology and temporal QoS variations of IoT-WSNs for accurately detecting the malicious nodes, which are subsequently isolated from routing decisions. Finally, a min-cost routing strategy is used over the secure nodes to minimize energy consumption, delay, and packet loss. The proposed QAR-GGNN model is validated by extensive NS-2 simulation with different performance metrics for secure and scalable IoT-WSN deployments. The performance of the QAR-GGNN model improves the network lifetime by 14.25%, 9.64%, and 5.02% better than the existing ML-HSOR, SDLEER, and BITA models respectively. Furthermore, the proposed model demonstrates higher performance than the existing models based on residual energy, delay, throughput, packet delivery rate, energy consumption and overall detection accuracy rate.</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

QAR-GGNN: QoS Aware Secure Routing in WSN via Mincost Strategy based Gated Graph Neural Network

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月12日

查看全文
  • M. Ishwarya Niranjana,
  • N. Kumareshan

摘要

Internet of Things-based Wireless Sensor Network (IoT-WSN) have developed as a core sensing infrastructure due to their potential to provide effective resource consumption and reliable service delivery in real-time applications. However, existing IoT-WSN routing methods exhibit some critical limitations including energy imbalance, inadequate Quality of Service (QoS) awareness, and poor robustness against routing attacks. To overcome these challenges, a novel QoS aware secure routing framework based on Gated Graph Neural Network (QAR-GGNN) is proposed to enhance energy balance across the network. The Gaussian mixture model (GMM) is used for efficient cluster formation, while a Mutated Spider-Tailed Horned Viper Optimization (M-STHVO) algorithm selects optimal cluster heads for maintaining energy consumption and prolong network lifetime. The deep learning-based GGNN is introduced to retrieve the graph-structured topology and temporal QoS variations of IoT-WSNs for accurately detecting the malicious nodes, which are subsequently isolated from routing decisions. Finally, a min-cost routing strategy is used over the secure nodes to minimize energy consumption, delay, and packet loss. The proposed QAR-GGNN model is validated by extensive NS-2 simulation with different performance metrics for secure and scalable IoT-WSN deployments. The performance of the QAR-GGNN model improves the network lifetime by 14.25%, 9.64%, and 5.02% better than the existing ML-HSOR, SDLEER, and BITA models respectively. Furthermore, the proposed model demonstrates higher performance than the existing models based on residual energy, delay, throughput, packet delivery rate, energy consumption and overall detection accuracy rate.