Design and Analysis of Multilayered Neural Network-Based Intrusion Detection System in the Internet of Things Network

Area of Science:

• Computer Science
• Cybersecurity
• Artificial Intelligence

Background:

• The Internet of Things (IoT) connects numerous devices, generating vast data but facing significant security vulnerabilities.
• Traditional security measures are inadequate for resource-constrained IoT devices, necessitating advanced, scalable solutions.
• Deep learning offers robust capabilities for analyzing heterogeneous data, making it suitable for complex IoT security challenges.

Purpose of the Study:

• To develop and evaluate a multilayered security approach for the transport layer of IoT networks.
• To address the limitations of conventional security solutions in protecting IoT systems.
• To leverage deep learning for enhanced threat detection and mitigation in IoT environments.

Main Methods:

• A multilayered security architecture was designed, integrating deep learning techniques.
• The proposed architecture was implemented and tested on the transport layer of IoT networks.
• Performance was evaluated using established intrusion detection datasets: CIC-IDS-2018, BoT-IoT, and ToN-IoT.

Main Results:

• The deep learning-based multilayered security approach demonstrated superior performance compared to existing methods.
• The system achieved a high accuracy rate of 98% in detecting and mitigating threats.
• The solution proved effective in securing the transport layer of IoT networks against various cyber threats.

Conclusions:

• The proposed deep learning-based multilayered security approach is a scalable and effective solution for IoT security.
• This method significantly enhances the protection of IoT networks, addressing critical vulnerabilities.
• The findings highlight the potential of deep learning in creating robust cybersecurity for the Internet of Things.