Scholarlink

<p>The growing integration of the Internet of Things into fifth-generation networks has raised security aspects, particularly concerning potential threats from network attackers. The pre-existing techniques to detect intrusions in such networks include the most accurate methods without concerning the hardware compatibility of the devices. This research paper presents a novel approach aimed at detecting intrusions within Internet of Things (IoT) devices with the goal of attaining a high level of accuracy while maintaining computational efficiency and minimizing processing time. One of the distinguishing aspects of this research is the incorporation of heterogeneous IoT environments, which uses a clustering technique to integrate data across different network domains, increasing its scalability. The proposed model showcases its adaptability and robustness in the face of diverse network configurations without imposing excessive overhead. The research delves into the architecture’s practical implementation. By focusing on the efficient integration of IoT devices within the network, the proposed model enables intrusion detection with minimal latency. This strategic consideration aligns with the ever-increasing demand for timely responses to potential threats in dynamic IoT environments. The research employs the WUSTL-IIOT and NSL-KDD datasets for the experimentations. The proposed model can successfully identify intrusions with an accuracy rate of 91%, accompanied by a substantial reduction in space-time complexity by approximately 90%.</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

A Combined Supervised and Unsupervised Deep Learning Approach for Intrusion Detection in IoT Traffic in an Edge Computing Environment

  • 期刊论文

  • 发表日期：

    2025年6月23日

查看全文
  • Sangeeta Sharma,
  • Priyanka Verma,
  • Nitesh Bharot,
  • Ashish Kumar,
  • Navdeep Singh Rathore,
  • Shivanshu Sharma

摘要

The growing integration of the Internet of Things into fifth-generation networks has raised security aspects, particularly concerning potential threats from network attackers. The pre-existing techniques to detect intrusions in such networks include the most accurate methods without concerning the hardware compatibility of the devices. This research paper presents a novel approach aimed at detecting intrusions within Internet of Things (IoT) devices with the goal of attaining a high level of accuracy while maintaining computational efficiency and minimizing processing time. One of the distinguishing aspects of this research is the incorporation of heterogeneous IoT environments, which uses a clustering technique to integrate data across different network domains, increasing its scalability. The proposed model showcases its adaptability and robustness in the face of diverse network configurations without imposing excessive overhead. The research delves into the architecture’s practical implementation. By focusing on the efficient integration of IoT devices within the network, the proposed model enables intrusion detection with minimal latency. This strategic consideration aligns with the ever-increasing demand for timely responses to potential threats in dynamic IoT environments. The research employs the WUSTL-IIOT and NSL-KDD datasets for the experimentations. The proposed model can successfully identify intrusions with an accuracy rate of 91%, accompanied by a substantial reduction in space-time complexity by approximately 90%.