Research on Automatic Intrusion Recognition of Multi-Source Targets in Heterogeneous Industrial Control Network
Multi-source target features of heterogeneous industrial control network are in constant change with ups and downs,resulting in low accuracy of intrusion recognition.An automatic intrusion identification method considering real-time changes in data features is proposed and applied to heterogeneous industrial control network.The normalized intrusion feature space is set,all network data are normalized to this space,and the out-of-range data are renormalized according to the maximum and minimum value comparison.Taking the time-domain matrix skewness feature,kurtosis feature,and envelope undulation feature as the categories of intrusion feature extraction,the three kinds of feature data sizes in the industrial control network are calculated respectively.On this basis,a set of intrusion data samples are first calculated in the network of each feature performance;then the performance parameters are converted to the cluster center value,and the Euclidean distance between the target to be identified and the cluster center is solved;finally,the automatic identification of intrusion targets is completed in accordance with the distance size.The test data proves that the method has high recognition accuracy in the industrial control network,and can realize accurate recognition in variety of network attack environments.The method has good application effect.
Heterogeneous industrial control networkMulti-source targetsAutomatic intrusion recognitionEnvelope heaveCluster centerNormalization
NETL
NSTL
万方数据
维普
