Anomaly Detection in Cyber-Physical Systems based on BiGRU-VAE

Various problems inevitably arise in cyber-physical systems, such as equipment failure, performance degradation, etc. Untimely detection of an abnormal state caused by a cyber-attack or a failure to operate devices in a cyber-physical system can lead to severe losses for the entire system. This paper proposes a method based on a deep bidirectional gated recurrent unit and variational autoencoder model to detect anomalies in a cyber-physical system. Experiments on a real dataset have shown the effectiveness of the proposed method in detecting anomalies in a cyber-physical system. Comparison with known methods showed the most accurate results according to the precision, recall, and F-measure metrics and amounted to 99.87%, 77.39%, and 87.20%, respectively.