Scholarlink

<p>The Internet of Things (IoT) is the intelligent framework that connects smart objects through the Internet. The IoT framework is vulnerable to various attacks because the Internet is unsafe and untrustworthy. The IoT cyber attacks can create significant physical and economic changes that affect the supply chains, manufacturing processes, production lines, impact the vehicle’s physical safety, and affect people’s health. Numerous systems are being connected to the Internet, and the data from the IoT systems is employed in the decision-making process in complex environments such as the military, healthcare, and so on. Hence, it becomes highly significant to safeguard the IoT platform. Traditional authentication and encryption approaches often fail in the IoT-aided framework, and an intrusion may cause damage to the framework. Hence, it is significant to model an Intrusion Detection System (IDS) for a specific network. In this work, a new attack detection, mitigation, and routing system on IoT is suggested for recognizing different kinds of attacks on IoT. At first, the attack detection is carried out by a designed Serial Cascaded and Attention-based Hybrid Convolution Network (SCA-HCN), which is formed by integrating a convolutional autoencoder with Temporal Convolutional Networks (TCN). Especially, in the IoT attack detection model, the proposed SCA-HCN model primarily provides significant advancement in improving the model’s accuracy, efficiency and scalability. Moreover, the integration of cascaded mechanism excels in identifying and understanding the context of attack over time. Further, the attack mitigation and routing tasks are performed by the Enhanced Dung Beetle Optimizer (EDBO). Enabling the EDBO optimization algorithm helps in providing more efficient and stable solutions, which helps in managing the IoT network traffic more effectively. During the routing process, some of the objective functions, including the shortest path, energy, delay, path loss, and Packet Delivery Ratio (PDR), are derived. The simulation outcomes are performed for the implemented framework over existing deep learning approaches and algorithms to analyze the ability of a designed model. From the results, the developed SCA-HCN-EDBO model achieves high performance rates in terms of 97.18% accuracy, 2.91% FPR, 2.5% FNR and 99.1% NPV, respectively. Hence, it is guaranteed that the introduced new attack detection, mitigation, and routing framework achieve superior and promising results.</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

An efficient attack detection with mitigation and routing framework in IoT by deep learning and enhanced dung beetle optimizer

  • 期刊论文

  • 发表日期：

    2026年4月27日

查看全文
  • Ajay Dilip Kumar Marapatla,
  • E. Ilavarasan

摘要

The Internet of Things (IoT) is the intelligent framework that connects smart objects through the Internet. The IoT framework is vulnerable to various attacks because the Internet is unsafe and untrustworthy. The IoT cyber attacks can create significant physical and economic changes that affect the supply chains, manufacturing processes, production lines, impact the vehicle’s physical safety, and affect people’s health. Numerous systems are being connected to the Internet, and the data from the IoT systems is employed in the decision-making process in complex environments such as the military, healthcare, and so on. Hence, it becomes highly significant to safeguard the IoT platform. Traditional authentication and encryption approaches often fail in the IoT-aided framework, and an intrusion may cause damage to the framework. Hence, it is significant to model an Intrusion Detection System (IDS) for a specific network. In this work, a new attack detection, mitigation, and routing system on IoT is suggested for recognizing different kinds of attacks on IoT. At first, the attack detection is carried out by a designed Serial Cascaded and Attention-based Hybrid Convolution Network (SCA-HCN), which is formed by integrating a convolutional autoencoder with Temporal Convolutional Networks (TCN). Especially, in the IoT attack detection model, the proposed SCA-HCN model primarily provides significant advancement in improving the model’s accuracy, efficiency and scalability. Moreover, the integration of cascaded mechanism excels in identifying and understanding the context of attack over time. Further, the attack mitigation and routing tasks are performed by the Enhanced Dung Beetle Optimizer (EDBO). Enabling the EDBO optimization algorithm helps in providing more efficient and stable solutions, which helps in managing the IoT network traffic more effectively. During the routing process, some of the objective functions, including the shortest path, energy, delay, path loss, and Packet Delivery Ratio (PDR), are derived. The simulation outcomes are performed for the implemented framework over existing deep learning approaches and algorithms to analyze the ability of a designed model. From the results, the developed SCA-HCN-EDBO model achieves high performance rates in terms of 97.18% accuracy, 2.91% FPR, 2.5% FNR and 99.1% NPV, respectively. Hence, it is guaranteed that the introduced new attack detection, mitigation, and routing framework achieve superior and promising results.