Scholarlink

<p>Wireless Sensor Networks (WSNs) are increasingly employed in fields such as precision agriculture, where the need for dependable data transmission and energy conservation is critical. Nonetheless, WSNs encounter significant challenges owing to the restricted battery life of sensor nodes and the detrimental effects of erroneous data on the aggregation and routing efficiency. Existing cluster-based routing and fault detection techniques often fail in dynamic settings and have high false alarm rates, which diminish the overall reliability of the network. To address these challenges, this study introduces a hybrid energy-efficient framework consisting of three primary stages: clustering based on fuzzy logic, selection of a Cluster Head (CH) using the Adaptive Human Evolutionary Optimization Algorithm (HEOA), and energy-aware routing through the Horned Lizard Optimization Algorithm (HLOA). To enhance fault resilience, a Deep Graph-Gated Recurrent Unit (GraphGRU) network is incorporated at the CH level to anticipate and filter out faulty data, thereby enhancing data integrity prior to aggregation. The proposed model was assessed against several existing techniques including BFOABMS, COA, IFCM-CHBCO, ASSO-SSO, and PFCRE. The results demonstrate significant performance improvements: the Measure of Dispersion (0.4314), Packet Delivery Ratio (99%), Average End-to-End Delay (65&#xa0;ms), Network Lifetime (3500 rounds), and Fault Prediction Accuracy (98.9%). These findings underscore the efficacy of the framework in improving the energy efficiency, reliability, and fault tolerance in WSNs.</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

Hybrid Optimization and Deep Graph-GRU Based Energy-Efficient Routing and Fault Prediction in WSNs for Smart Agriculture Applications

  • 期刊论文

  • 发表日期：

    2026年2月2日

查看全文
  • R. Bavithra,
  • J. Jaya

摘要

Wireless Sensor Networks (WSNs) are increasingly employed in fields such as precision agriculture, where the need for dependable data transmission and energy conservation is critical. Nonetheless, WSNs encounter significant challenges owing to the restricted battery life of sensor nodes and the detrimental effects of erroneous data on the aggregation and routing efficiency. Existing cluster-based routing and fault detection techniques often fail in dynamic settings and have high false alarm rates, which diminish the overall reliability of the network. To address these challenges, this study introduces a hybrid energy-efficient framework consisting of three primary stages: clustering based on fuzzy logic, selection of a Cluster Head (CH) using the Adaptive Human Evolutionary Optimization Algorithm (HEOA), and energy-aware routing through the Horned Lizard Optimization Algorithm (HLOA). To enhance fault resilience, a Deep Graph-Gated Recurrent Unit (GraphGRU) network is incorporated at the CH level to anticipate and filter out faulty data, thereby enhancing data integrity prior to aggregation. The proposed model was assessed against several existing techniques including BFOABMS, COA, IFCM-CHBCO, ASSO-SSO, and PFCRE. The results demonstrate significant performance improvements: the Measure of Dispersion (0.4314), Packet Delivery Ratio (99%), Average End-to-End Delay (65 ms), Network Lifetime (3500 rounds), and Fault Prediction Accuracy (98.9%). These findings underscore the efficacy of the framework in improving the energy efficiency, reliability, and fault tolerance in WSNs.