Scholarlink

<p>Underwater Wireless Sensor Networks (UWSNs) plays an important role in a range of applications including flood rescue management, environmental monitoring, oceanographic research, and military surveillance. Despite their importance, UWSNs are highly susceptible to security threats due to their constrained communication infrastructure, limited bandwidth, high propagation delays, and dynamic topology. Among the critical threats, Black Hole, Wormhole, and Sinkhole attacks stand out as particularly damaging. These attacks exploit routing vulnerabilities to disrupt data transmission, mislead node communication, and degrade the overall network performance. Detecting such attacks is challenging due to their deceptive and coordinated nature. To counter these threats, trust-based security mechanisms offer a viable solution by continuously evaluating the behaviour of nodes to identify malicious activity. However, traditional trust models struggle to account for the temporal dynamics and adaptiveness required in hostile underwater environments. This paper proposes an intelligent trust model that utilizes Long Short-Term Memory (LSTM) networks to capture the temporal dependencies in node behaviour for accurate and dynamic trust evaluation. By learning from historical interaction patterns, the LSTM effectively identifies anomalies linked to Black Hole, Wormhole, and Sinkhole attacks. To further optimize the model’s performance, the Whale Optimization Algorithm (WOA) is integrated for fine-tuning LSTM hyperparameters. This synergy enhances the model’s accuracy, adaptability, and robustness in detecting sophisticated threats in UWSNs, thereby improving network reliability and security.</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

Trust model for hybrid security attack mitigation in underwater acoustic wireless sensor networks using LSTM and whale optimization algorithm

  • 期刊论文

  • 发表日期：

    2025年12月13日

查看全文
  • Vinay Agrawal,
  • Rakesh Kumar

摘要

Underwater Wireless Sensor Networks (UWSNs) plays an important role in a range of applications including flood rescue management, environmental monitoring, oceanographic research, and military surveillance. Despite their importance, UWSNs are highly susceptible to security threats due to their constrained communication infrastructure, limited bandwidth, high propagation delays, and dynamic topology. Among the critical threats, Black Hole, Wormhole, and Sinkhole attacks stand out as particularly damaging. These attacks exploit routing vulnerabilities to disrupt data transmission, mislead node communication, and degrade the overall network performance. Detecting such attacks is challenging due to their deceptive and coordinated nature. To counter these threats, trust-based security mechanisms offer a viable solution by continuously evaluating the behaviour of nodes to identify malicious activity. However, traditional trust models struggle to account for the temporal dynamics and adaptiveness required in hostile underwater environments. This paper proposes an intelligent trust model that utilizes Long Short-Term Memory (LSTM) networks to capture the temporal dependencies in node behaviour for accurate and dynamic trust evaluation. By learning from historical interaction patterns, the LSTM effectively identifies anomalies linked to Black Hole, Wormhole, and Sinkhole attacks. To further optimize the model’s performance, the Whale Optimization Algorithm (WOA) is integrated for fine-tuning LSTM hyperparameters. This synergy enhances the model’s accuracy, adaptability, and robustness in detecting sophisticated threats in UWSNs, thereby improving network reliability and security.