An Artificial Intelligence-based Ensemble Technique for Intrusion Detection and Prevention in IoT Systems

Abstract

The rapid application of Internet of Things (IoT) devices across various sectors, including smart homes, industrial automation, and healthcare, has realized the vision of the Industrial Revolution. However, this growth has also brought a surge in attackers targeting IoT networks. Therefore, real-time security measures are crucial to counter unauthorized access and data breaches. This paper presents an Artificial intelligence (AI)-based intrusion detection and prevention system (IDPS) capable of real-time stoppage on the IoT networks. An ensemble feature selection approach was introduced to address this to eliminate irrelevant and redundant dataset features, thereby improving the proposed model performance. The study utilizes the N-BaIoT dataset, known for its high accuracy in intrusion detection for the training of the proposed model. The trained model processes network data at the IoT gateway, predicting real-time attacks, and employs a responsive background script based on machine learning analysis for intrusion prevention. However, the AI-based utilizes the Light Gradient Boosting Machine (LightGBM) Classifier for decision-making, achieving 99.9% accuracy. As IoT continues to evolve, proactive security measures are paramount. Incorporating ensemble algorithms for feature selection, AI-based techniques, and a responsive script forming a promising framework for detecting and preventing attacks is a paradigm shift from laboratory to field in fostering a safer and more secure IoT landscape.