Scholarlink

<p>A Wireless Sensor Network (WSN) refers to a decentralized network architecture where sensor nodes, administrative units, and sink nodes work together to observe conditions and gather data efficiently. Due to the rapid progress in wireless communication, WSNs have gained significant popularity across various industries, as they overcome the limitations of existing techniques in collecting and monitoring data in harsh environments. The rapid growth in data transmission has made security a critical concern in WSN. The prevailing methods used to analyze network traffic data exhibit poor detection accuracy during intrusion attacks. This paper proposes a novel hybrid framework named Snow Leopard Beetle Optimization_Convolutional eXtreme Gradient Boosting (SLepBO_ ConvXGB) for intrusion detection in WSN, which introduces a unique integration mechanism, combining the global exploration capability of Snow Leopard Optimization Algorithm (SLOA) with the local exploitation strength of Dung Beetle Optimization (DBO) to achieve faster and more stable convergence in routing decisions. The adaptive synergy between SLOA and DBO ensures optimized routing with reduced energy consumption and minimal packet delay, outperforming conventional routing approaches. Routing in the simulated WSN is achieved using Snow Leopard Beetle Optimization (SLepBO), taking multiple objective fitness parameters into account. Next, the input data undergoes normalization using the Tanh estimator, after which feature fusion is performed through a Deep Neural Network (DNN) with Jeffreys similarity to improve classification performance. Subsequently, intrusion detection is carried out using ConvXGB, with its performance optimized through SLepBO. Further, the SLepBO for routing achieved distance, residual energy, and delay of 0.100&#xa0;J, 81.346&#xa0;m, and 0.718&#xa0;ms, whereas the SLepBO_ConvXGB for intrusion detection obtained 92.777% of True Negative Rate (TNR), 91.988% of accuracy, 91.157% of True Positive Rate (TPR), 90.878% of precision, and 91.818% of F1-score. These results highlight the technical innovation and synergistic mechanisms of the proposed framework, demonstrating superior convergence, energy efficiency, and intrusion detection performance compared to existing methods.</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

Snow leopard beetle optimization and deep learning: a hybrid approach for routing and intrusion detection in WSN

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月1日

查看全文
  • Geerthik S,
  • Raghuraman V,
  • Girija P,
  • Aarthi K. C

摘要

A Wireless Sensor Network (WSN) refers to a decentralized network architecture where sensor nodes, administrative units, and sink nodes work together to observe conditions and gather data efficiently. Due to the rapid progress in wireless communication, WSNs have gained significant popularity across various industries, as they overcome the limitations of existing techniques in collecting and monitoring data in harsh environments. The rapid growth in data transmission has made security a critical concern in WSN. The prevailing methods used to analyze network traffic data exhibit poor detection accuracy during intrusion attacks. This paper proposes a novel hybrid framework named Snow Leopard Beetle Optimization_Convolutional eXtreme Gradient Boosting (SLepBO_ ConvXGB) for intrusion detection in WSN, which introduces a unique integration mechanism, combining the global exploration capability of Snow Leopard Optimization Algorithm (SLOA) with the local exploitation strength of Dung Beetle Optimization (DBO) to achieve faster and more stable convergence in routing decisions. The adaptive synergy between SLOA and DBO ensures optimized routing with reduced energy consumption and minimal packet delay, outperforming conventional routing approaches. Routing in the simulated WSN is achieved using Snow Leopard Beetle Optimization (SLepBO), taking multiple objective fitness parameters into account. Next, the input data undergoes normalization using the Tanh estimator, after which feature fusion is performed through a Deep Neural Network (DNN) with Jeffreys similarity to improve classification performance. Subsequently, intrusion detection is carried out using ConvXGB, with its performance optimized through SLepBO. Further, the SLepBO for routing achieved distance, residual energy, and delay of 0.100 J, 81.346 m, and 0.718 ms, whereas the SLepBO_ConvXGB for intrusion detection obtained 92.777% of True Negative Rate (TNR), 91.988% of accuracy, 91.157% of True Positive Rate (TPR), 90.878% of precision, and 91.818% of F1-score. These results highlight the technical innovation and synergistic mechanisms of the proposed framework, demonstrating superior convergence, energy efficiency, and intrusion detection performance compared to existing methods.