An attack detection system in Internet of Things with deep learning based on VGG16 architecture and Siberian tiger optimization algorithm (STO)

Document Type : Original Article

Authors

1 Associate Professor, Department of Computer Engineering, Maybod Branch, Islamic Azad University, Maybod, Iran

2 Department of Computer Engineering, Maybod Branch, Islamic Azad University, Maybod, Iran

Abstract

One of the significant challenges in the Internet of Things is the presence of attacking nodes called botnets. Many nodes are infected with malware in these attacks and perform attacks against network services, such as distributed denial of service. In most cases, botnets target application services in the cloud computing layer. For this reason, it is essential to detect attacks in the Internet of Things as an intermediate layer. Providing a distributed intrusion detection system in the Internet of Things increases the ability to detect intrusion and has a high ability to analyze a large volume of network traffic. Deep learning methods such as convolutional neural networks have a high ability to recognize complex patterns in images. In this article, to use CNN network architecture in network intrusion detection, network traffic is coded in the form of images in a new way. Network traffic images are used to train the VGG16 model, a CNN technique. In the proposed method to focus the proposed penetration detection system, the Siberian tiger optimization algorithm is used to select features and reduce dimensions. The proposed intrusion detection system is trained on the NSL-KDD dataset. The evaluations showed that it has accuracy, sensitivity, and precision equal to 99.62%, 99.38%, and 98.74%, respectively. In the feature selection phase, the proposed method is more accurate than WOA, HHO, and AO algorithms. The proposed method is more accurate in detecting network attacks than CNN, VGG16, Muhti-CNN, and PSO-CNN methods.

Keywords

Main Subjects

Smiley face

 

Files

History

  • Receive Date: 05 April 2025
  • Revise Date: 09 May 2025
  • Accept Date: 16 June 2025
  • Publish Date: 22 June 2025

Share

How to cite

Statistics