Scholarlink

<p>The Internet of Things, also called the IoT, is a smart device that can wirelessly gather and communicate heterogeneous data because they have processing, sensing, and networking capabilities. The requirement for IoT system security is growing as more and more IoT devices are used in regular tasks. Intruders may easily target these IoT devices to carry out harmful tasks and destroy the underlying network. Therefore, the Integrated Model for Intrusion Detection (IM-ID), a hybridized&#xa0;bio-inspired intrusion detection system (IDS) for the Internet of Things framework, is proposed in this research. The Salp Swarm Algorithm (SSA) and the integrated Sine Cosine Algorithm (SCA) identify the key characteristics of network traffic. Additionally, intrusion detection uses the attention-based bi-directional long short-term memory (ABiLSTM) technology. The Gorilla Troops Optimizer (GTO) is used to do hyperparameter tweaking to&#xa0;improve the ABiLSTM algorithm's intrusion detection performance. The GTO-based hyperparameter tuning procedure is used since manual trial-and-error hyperparameter tuning is time-consuming, illustrating the work's uniqueness. Various tests were conducted to verify the improved IM-ID system solution in terms of intrusion detection. With an average accuracy of 96.88% on the acquired datasets, the simulation figures validate the encouraging outcomes of the IM-ID system when compared to current DL approaches.</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

Designing integrated model for intrusion detection (IM-ID) for internet of things (IoT) using deep learning techniques

  • 期刊论文

  • 发表日期：

    2025年12月17日

查看全文
  • M Santhosh Kumar,
  • R Dhanapal

摘要

The Internet of Things, also called the IoT, is a smart device that can wirelessly gather and communicate heterogeneous data because they have processing, sensing, and networking capabilities. The requirement for IoT system security is growing as more and more IoT devices are used in regular tasks. Intruders may easily target these IoT devices to carry out harmful tasks and destroy the underlying network. Therefore, the Integrated Model for Intrusion Detection (IM-ID), a hybridized bio-inspired intrusion detection system (IDS) for the Internet of Things framework, is proposed in this research. The Salp Swarm Algorithm (SSA) and the integrated Sine Cosine Algorithm (SCA) identify the key characteristics of network traffic. Additionally, intrusion detection uses the attention-based bi-directional long short-term memory (ABiLSTM) technology. The Gorilla Troops Optimizer (GTO) is used to do hyperparameter tweaking to improve the ABiLSTM algorithm's intrusion detection performance. The GTO-based hyperparameter tuning procedure is used since manual trial-and-error hyperparameter tuning is time-consuming, illustrating the work's uniqueness. Various tests were conducted to verify the improved IM-ID system solution in terms of intrusion detection. With an average accuracy of 96.88% on the acquired datasets, the simulation figures validate the encouraging outcomes of the IM-ID system when compared to current DL approaches.