Abstract:
The widespread use of various devices in domains like housing environments, transportation systems, healthcare facilities, and industrial sectors has demonstrated the significant development of Internet of Things (IoT). The introduction and integration of the IoT concept in industrial environments have resulted in substantial modifications to the architecture of Industrial Automation and Control Systems (IACS) and the widespread interconnectivity of various industrial systems. The outcome of this development is often known as the Industrial IoT (IIoT), which eliminates the obstacle of linking IACS with separate traditional information and communication technology (ICT) platforms. In contemporary times, the IoT has begun to affect our individual lives and extend its influence beyond our immediate surroundings, establishing a foundation for imminent cyberattacks targeting the IoT. The extensive utilization of the IoT has produced a productive ground for potential assaults against IoT systems. Machine learning (ML) algorithms have enhanced wireless communication security in IIoT-based systems and addressed various cyber security issues. This study shows a better ensemble model for finding intrusions. It uses Particle Swarm Optimization (PSO) to group network data into groups of harmful behaviours in the IIoT setting. We tested the proposed model's performance on the X-IIoTID dataset, a cyber security dataset founded on the Industrial Internet of Things (IIoT). We looked at its recall, precision, F1 score, F2 score, ROC-AUC, and accuracy metrics. The obtained findings were compared to those of other recent state-of-the-art ML approaches, and it was observed that our model exhibited superior performance.
