Hybrid IDS for IoT Approach Combining Deep Extraction and Robust Classification

Abstract

The rapid adoption of the Internet of Things (IoT) has significantly increased security risks, exposing networks to advanced cyber threats. In this paper, a structured analysis of the IoT architecture has been provided to identify vulnerabilities and attack vectors for each layer. To address these challenges, we developed a deep learning-based intrusion detection system, evaluated on the ToNIoT dataset. The experimental results confirm its effectiveness in detecting intrusions with high precision, demonstrating its potential to improve IoT security. This work highlights the potential of deep learning to improve IoT network security and resilience, providing a robust framework for future research and practical applications.