An Ensemble Approach to Identify Intrusions in Edge IoT and IIoT Networks

This work presents an IDS for Internet of Things (IoT) that makes use of the ML-EdgeIIoT-dataset to safeguard critical IoT infrastructures. The proposed model integrates CNN and GRU with an XGBoost layer to enhance predictive accuracy. The CNN-GRU architecture is specifically engineered to effectively capture both spatial and temporal features. The CNN component in this design is tasked with capturing local patterns, whereas the GRU component is responsible for managing sequential dependencies. The results are combined with the initial characteristics and improved using XGBoost to generate the ultimate forecasts. The ensemble approach ensures robust real-time data processing, enabling fast and accurate risk identification. The effectiveness of the Intrusion Detection System (IDS) is evaluated using the four accuracy metrics, demonstrating its outstanding efficiency and resilience across various workloads. This comprehensive architecture prioritizes the security and dependability of the IoT system and is adaptable to evolving cyber threats. Our Intrusion Detection System (IDS) establishes a higher benchmark for taking initiative and adapting to changing circumstances in order to defend IoT networks. It offers robust safeguards to ensure the stability and security of these networks. The study emphasizes the significance of employing sophisticated machine learning methods to enhance the detection capabilities and overall robustness of IDS in Internet of Things (IoT) environments.