Securing Smart Grid IoT System using A Robust RNN based Cyberattack Detection Framework

Cybersecurity in smart grid IoT systems is now a concern of high priority, as the integration of connected devices increases, and these systems become more vulnerable to sophisticated cyberattacks. This paper proposes a new RNN+GRU-based cyberattack detection system to meet such security challenges in smart grid IoT networks. The proposed model is based on Recurrent Neural Networks (RNN) and Gated Recurrent Units (GRU) to efficiently capture temporal and contextual patterns of the attack data for improving detection accuracy of different types of cyber-attacks. Comparing the performance of our proposed model with other models like Support Vector Machines (SVM), Random Forest (RF), Convolutional Neural Network (CNN), Deep Neural Network (DNN), and Long Short-Term Memory (LSTM) network is shown. Experimental results show that the RNN+GRU-based model outperforms these approaches in some key performance metrics: accuracy, precision, recall, F1-score, and ROC-AUC, better than other compared methods by achieving superior attack detection rates. Moreover, correlation analysis is also included for both the existing DNN model and the proposed approach; thus, the effectiveness of the RNN+GRU model in adapting to various scenarios of cyberattacks is noted. In the final analysis, the RNN+GRU-based cyberattack detection system shows great promise in securing smart grid IoT networks and shows better performance compared to some of the existing techniques.