Scholarlink

<p>Nodes in Wireless Sensor Networks (WSNs) are characterized by limited resources, including energy, memory, and processing power, which necessitates optimized energy usage to extend the network’s lifespan. Efficient CH selection significantly contributes to conserving energy and improving the operational duration of the network. However, anomalies can arise when certain nodes behave abnormally, either due to faulty measurements or limitations on resources, such energy depletion, limited memory, bandwidth, or processing capability. This study proposes an energy-efficient routing framework for WSNs that integrates three distinct anomaly detection. Support Vector Machines (SVM), Isolation Forest and Density-Based Spatial Clustering of Applications with Noise (DBSCAN). Among these, DBSCAN demonstrated the highest anomaly detection accuracy at 98%, followed by Isolation Forest at 85%, and SVM at 79%. For data transmission, a two-hop routing technique is employed, wherein cluster members (CMs) transmit data to their respective Cluster Heads (CHs), which subsequently relay the data to a Global Cluster Head (GCH) for forwarding to the Base Station (BS). To optimize GCH selection among CHs, we propose an objective function that incorporates crucial parameters including residual energy levels, proximity of CHs to the base station, and link quality. Link quality is identified as a critical factor influencing the transmission rates of sensor nodes (SNs). It is evident from the results that the introduced routing algorithm, when paired with SVM, outperforms in terms of total residual energy with higher data packet size, while DBSCAN provides the highest stability period for network operations for all data packet size. The integration of DBSCAN, SVM, and Isolation Forest provides robust mechanisms for identifying faulty nodes or nodes experiencing resource constraints. Fitness-driven selection of CH and GCH focuses on optimizing energy efficiency, which in turn supports a longer operational duration for the network. The clustering-based approach and hierarchical routing ensure scalability, making the protocol effective even in large WSN deployments. Extensive simulations using the Intel Berkeley Research Lab sensor dataset validate the efficacy and robustness of the proposed method. The results demonstrate that the proposed method significantly improves energy utilization and extends network lifetime, supporting the development of more efficient routing strategies 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

Energy-aware Clustering and Lifetime Extension in WSNs Using Evolutionary Learning Algorithms

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月16日

查看全文
  • Tripti Sharma,
  • Gitesh Goyal,
  • Babita Tiwari,
  • G. S. Tomar,
  • Neeru Rathee

摘要

Nodes in Wireless Sensor Networks (WSNs) are characterized by limited resources, including energy, memory, and processing power, which necessitates optimized energy usage to extend the network’s lifespan. Efficient CH selection significantly contributes to conserving energy and improving the operational duration of the network. However, anomalies can arise when certain nodes behave abnormally, either due to faulty measurements or limitations on resources, such energy depletion, limited memory, bandwidth, or processing capability. This study proposes an energy-efficient routing framework for WSNs that integrates three distinct anomaly detection. Support Vector Machines (SVM), Isolation Forest and Density-Based Spatial Clustering of Applications with Noise (DBSCAN). Among these, DBSCAN demonstrated the highest anomaly detection accuracy at 98%, followed by Isolation Forest at 85%, and SVM at 79%. For data transmission, a two-hop routing technique is employed, wherein cluster members (CMs) transmit data to their respective Cluster Heads (CHs), which subsequently relay the data to a Global Cluster Head (GCH) for forwarding to the Base Station (BS). To optimize GCH selection among CHs, we propose an objective function that incorporates crucial parameters including residual energy levels, proximity of CHs to the base station, and link quality. Link quality is identified as a critical factor influencing the transmission rates of sensor nodes (SNs). It is evident from the results that the introduced routing algorithm, when paired with SVM, outperforms in terms of total residual energy with higher data packet size, while DBSCAN provides the highest stability period for network operations for all data packet size. The integration of DBSCAN, SVM, and Isolation Forest provides robust mechanisms for identifying faulty nodes or nodes experiencing resource constraints. Fitness-driven selection of CH and GCH focuses on optimizing energy efficiency, which in turn supports a longer operational duration for the network. The clustering-based approach and hierarchical routing ensure scalability, making the protocol effective even in large WSN deployments. Extensive simulations using the Intel Berkeley Research Lab sensor dataset validate the efficacy and robustness of the proposed method. The results demonstrate that the proposed method significantly improves energy utilization and extends network lifetime, supporting the development of more efficient routing strategies in WSNs.