Scholarlink

<p>Internet of Things (IoT) is intermittently targeted by Distributed Denial of Service (DDoS) attacks, which occupy computational resources and bandwidth for preventing potential users from accessing services. The attack strategy involves massively flooding of packets. As the use of IoT foundations spreads around the globe, so are the number of attacks and threats that these systems face. The IoT are susceptibility to a range of network attacks due to their low processing capacity and wireless connection. One among these is the HELLO Flood attack, where an attacker who is not a legitimate node in the network floods HELLO requests to every genuine node in the network, compromising Wireless Sensor Network security. In this work we have designed a novel HELLO Flood attack detection and prevention model using the artificial intelligence. Here, the development of Modified Deep Neural Network (MDNN) is introduced for HELLO Flood attack detection, in which the hybrid Coyote Optimization Algorithm (COA) and Elephant Herding Optimization (EHO) called CoYote-Elephant Herding Algorithm (CY-EHA) is used for weight optimization as a training algorithm. The modified DNN works with the input parameters like “RSS, route discovery time, inter-route discovery, distance, data rate, and packet arrival rate”. Once the attack node is identified, it is removed utilizing CY-EHA. Experimental analysis showed that the proposed CY-EHA-MDNN reduced the cost function by 15.22%, 33.90%, 1.68%, and 30% compared to EHO-MDNN, COA-MDNN, GWO-MDNN, and PSO-MDNN, respectively, demonstrating superior performance in HELLO flood attack detection and prevention for IoT.</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

Enhanced detection and mitigation of HELLO flood attacks in IoT networks using modified deep neural networks and optimization algorithms

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月23日

查看全文
  • Pawan Kumar Verma,
  • Mohit Kumar,
  • Abhishek Gupta,
  • Ashwini Kumar Saini,
  • Ajeet Kumar Sharma,
  • Nitin Rakesh

摘要

Internet of Things (IoT) is intermittently targeted by Distributed Denial of Service (DDoS) attacks, which occupy computational resources and bandwidth for preventing potential users from accessing services. The attack strategy involves massively flooding of packets. As the use of IoT foundations spreads around the globe, so are the number of attacks and threats that these systems face. The IoT are susceptibility to a range of network attacks due to their low processing capacity and wireless connection. One among these is the HELLO Flood attack, where an attacker who is not a legitimate node in the network floods HELLO requests to every genuine node in the network, compromising Wireless Sensor Network security. In this work we have designed a novel HELLO Flood attack detection and prevention model using the artificial intelligence. Here, the development of Modified Deep Neural Network (MDNN) is introduced for HELLO Flood attack detection, in which the hybrid Coyote Optimization Algorithm (COA) and Elephant Herding Optimization (EHO) called CoYote-Elephant Herding Algorithm (CY-EHA) is used for weight optimization as a training algorithm. The modified DNN works with the input parameters like “RSS, route discovery time, inter-route discovery, distance, data rate, and packet arrival rate”. Once the attack node is identified, it is removed utilizing CY-EHA. Experimental analysis showed that the proposed CY-EHA-MDNN reduced the cost function by 15.22%, 33.90%, 1.68%, and 30% compared to EHO-MDNN, COA-MDNN, GWO-MDNN, and PSO-MDNN, respectively, demonstrating superior performance in HELLO flood attack detection and prevention for IoT.