Secure and Efficient Intrusion Detection System in Medical Cyber-Physical Systems

As technology advances toward Cyber-Physical Systems (CPS), there is a significant growth in the development of Medical Cyber-Physical Systems (MCPS). In this 21<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">st</sup> century, the cyber-attacks eventually increase as technology increases. Cyberattacks in medical cyber-physical systems cause a great disruption in the patient’s life as well as in the patient's information. A novel Intrusion Detection System (IDS) has been created using Graph Neural Networks (GNNs) and Recurrent Neural Networks (RNNs). GNNs manage relational dependencies by analyzing the behavior of nodes, which represent the data. On the other hand, Recurrent Neural Networks handle temporal dependencies and sequential data like sensor reading, communication logs, and time-stamped series. ICU healthcare security dataset is used in this study to train the model. This dataset is constructed as a graph, then the graph has been fed to the Graph Neural Networks and Recurrent Neural Networks (GNN+RNN) framework. The proposed model has demonstrated a strong performance with 99% detection accuracy, 100% precision, 98% recall, and 99.13% f1-score.