CNN-BiLSTM-Based Classification of RPL Attacks in IoT Smart Grid Networks (Industry Track)
Yue Guan, Morteza Noferesti, Naser Ezzati‐Jivan · 2023
The widespread integration of Internet of Things (IoT) technology in modern electrical power grids has given rise to the emergence of Industrial IoT Smart Grid Networks. However, the utilization of the Routing Protocol for Low-Power Lossy Networks (RPL) in IIoT Smart Grid Networks exposes them to significant risks of routing attacks due to their global connectivity and resource limitations. This paper proposes a hybrid deep neural network approach that utilizes the CNN-BiLSTM network for the detection and classification of attacks in the RPL protocol of IoT Smart Grid Networks. Relevant preprocessing and feature enhancement techniques are applied to engineer and enhance pertinent features. The Synthetic Minority Oversampling Technique (SMOTE) is employed to address the data imbalance problem. The performance of the proposed approach is evaluated and compared with seven different deep learning and traditional classification algorithms on two scenarios. First, we simulate an IIoT network and various types of attacks mentioned in [6]. The results demonstrate the superior performance of the CNN-BiLSTM based approach, achieving an accuracy of 91%, precision of 89%, and recall of 89% in detecting various RPL attacks. Next, we apply the proposed approach to another dataset collected from nine commercial IoT devices infected by two botnets [19]. The approach outperforms other machine learning-based approaches with an accuracy of 90%, precision of 89%, recall of 90%, and F1-score of 89%, which demonstrates the practical applicability of the proposed approach in real-world industrial settings.