Industrial Internet of Things Intrusion Detection Model Integrating Graph Attention Network and Gated Temporal Convolutional Network
Chenghe Peng, Ying Zhang · 2024
In recent years, the application of the Industrial Internet of Things (IIoT) in industry has become increasingly important. However, because IIoT devices and systems are connected to the Internet, they may face the risk of malicious intrusions that can cause serious damage to production processes and data security. Intrusion detection systems (IDS) can be used to monitor IIoT for malicious activities and develop response strategies. Therefore, this article proposes an intrusion detection model called GTRGAT to ensure the network security of IIoT. In particular, the GTRGAT model uses GTCN to quickly pull out sequence features from IIoT traffic. GTCN combines convolutional operations and gating mechanisms, aiming to process time series data more efficiently and capture its intrinsic patterns. By using the TON_IoT dataset, the performance of the GTRGAT model is verified, showing that it can accurately detect intrusions. The experiment also showed that the GTRGAT model works better than other advanced methods, so it can be used as a workable way to find intrusions in industrial IoT networks.