<![CDATA[The rapid development of 5G technology has significantly impacted the implementation of the Industrial Internet of Things (IIoT), enhancing industrial efficiency and productivity; however, as the number of connected devices increases, critical challenges related to intrusion detection and data security arise. This research aims to develop a Hybrid Machine Learning (ML) and Federated Learning (FL) model to improve the performance of Intrusion Detection Systems (IDS) in 5G-based IIoT and compare its effectiveness with a centralized Deep Learning (DL) model, with expectations of improving detection accuracy, bandwidth efficiency, and user privacy. Testing was conducted on intrusion detection and anomaly classification using cybersecurity benchmark datasets, including CICIDS, NSL-KDD, and TON_IoT. The experimental results demonstrated that the Hybrid ML-FL model outperforms the centralized model in detecting DDoS and Malware attacks, delivering higher recall, superior precision, and enhanced bandwidth efficiency by transmitting only model updates rather than the entire dataset. Furthermore, Federated Learning provides significant privacy preservation by mitigating data leakage risks while showing superior scalability as the IIoT network expands. Overall, this Hybrid ML-FL framework provides an adaptive, efficient, and secure solution for 5G-based IIoT security and scalability challenges, opening up promising avenues for future robust system developments.]]>
Copyrights © 2023
