Research on Intelligent Intrusion Detection System Model for Train Network Based on TCN

Onboard network of trains has high requirements on the performance of attack detection algorithm, and the network traffic contains characteristics of Time-series data. Therefore, a model of intelligent intrusion detection system based on Temporal Convolutional Network (TCN) is proposed to detect unknown threats. This method not only keeps the advantage of Convolutional neural network in less resource consumption, but also can better process the vehicle network traffic with temporal characteristics. The model was tested using the UNSW-NB15 dataset, compared and analyzed with the LSTM, Lenet-5, MLenet-5, and HAST models. The accuracy of Intelligent Intrusion Detection System Model Based on TCN is 90.1%, it increases by 2.6% compared with the other four models; the false-alarm rate is 6.5%, it decreases by 25.3% compared with the other four models; the false-negative rate is 12.4%, it decreases by 10.8% compared with the other four models. The results show that, the model proposed in this article greatly improves the accuracy of intrusion detection and reduces the false positive rate, which verifies the effectiveness of the method model.