An Intrusion Detection Method for Industrial Control System Based on Gate Recurrent Unit

Abstract

We proposed an industrial control system intrusion detection method based on gate recurrent unit neural network to handle the problem that the intrusion detection method based on traditional machine learning algorithm such as SVM, decision tree, NN etc., cannot effectively deal with massive, high-dimensional, time related network traffic data in industrial control system. We used the update gate and the reset gate of GRU to save the information of the data in the time dimension. Its deep structure can fully learn the data features, and we used Adam algorithm to optimize the gradient training process of the neural network. Comparison experiments were conducted with the intrusion detection method based on machine learning algorithms such as SVM, decision tree, NN, RNN, LSTM etc. The results show that the proposed method has higher classification accuracy than SVM, decision tree, NN and RNN, and the accuracy is basically the same as LSTM but the training time is greatly reduced.