Enhancing cybersecurity in Edge IIoT networks: An asynchronous federated learning approach with a deep hybrid detection model

Abstract

In the rapidly evolving field of the Industrial Internet of Things (IIoT), advancements in wireless technology have resulted in significant cybersecurity vulnerabilities. These weaknesses pose serious risks such as damage to manufacturing systems, theft of intellectual property, and substantial financial losses. This study introduces an advanced deep hybrid learning model in an asynchronous federated learning setup, aimed at improving the detection of cyberattacks and ensuring robust data privacy. The combination of Convolutional Neural Networks (CNN), Gated Recurrent Units (GRU), and Long Short-Term Memory (LSTM) networks provides an effective solution for quickly identifying anomalies in IIoT sensor traffic. Our model operates asynchronously, ensuring data remains localised to improve security while avoiding the need for complete node synchronisation. Demonstrating outstanding effectiveness, the model achieves an accuracy of 1.00%, precision of 1.00%, recall of 1.00%, and an F1 score of 1.00% across a variety of IIoT environments. These results highlight the model’s exceptional adaptability and its capability to rapidly respond to emergent threats, marking a significant step forward in the protection of IIoT infrastructures and the rigorous maintenance of data privacy.