Enhancing IoT Network Security Through Deep Learning-Powered Intrusion Detection System

• Author(s): James O. Ogunseye ; Afolabi Emmanuel
• Paper ID: 1709373
• Page: 1453-1458
• Published Date: 27-06-2025
• Published In: Iconic Research And Engineering Journals
• Publisher: IRE Journals
• e-ISSN: 2456-8880
• Volume/Issue: Volume 8 Issue 12 June-2025

Abstract

The explosive growth of the Internet of Things (IoT) has introduced increased connectivity and automation with increased cybersecurity threats, especially on resource-constrained devices. Traditional intrusion detection systems (IDS) cannot respond to these dynamic threats due to their reliance on known attack patterns and lack of adaptability. This work investigates the application of deep learning (DL) techniques, in the form of a hybrid Convolutional Neural Network–Long Short-Term Memory (CNN-LSTM) model, to improve IDS performance for IoT networks. The proposed model was evaluated and tested on three publicly available benchmark datasets—CICIoT2023, Bot-IoT, and UNSW-NB15—and has been shown to have improved accuracy, precision, and recall compared to traditional machine learning models. Despite DL's typically high computational requirements, the model showed efficient performance with minimum latency, and it was a suitable candidate for the edge. The results confirm the potential of DL-based IDS to provide robust, adaptive, and elastic solutions for IoT security against new and traditional cyberattacks

Keywords

Internet of Things (IoT), Intrusion Detection System (IDS), Deep Learning, CNN-LSTM, Cybersecurity, Network Security, Edge Computing, Anomaly Detection

Citations

IRE Journals:
James O. Ogunseye , Afolabi Emmanuel "Enhancing IoT Network Security Through Deep Learning-Powered Intrusion Detection System" Iconic Research And Engineering Journals Volume 8 Issue 12 2025 Page 1453-1458

IEEE:
James O. Ogunseye , Afolabi Emmanuel "Enhancing IoT Network Security Through Deep Learning-Powered Intrusion Detection System" Iconic Research And Engineering Journals, 8(12)