舰船网络入侵行为的检测与识别研究

Research on detection and recognition of naval ship network intrusion behavior

入侵行为检测是保证舰船网络安全的核心技术,当前入侵行为检测与识别存在检测误差大,识别准确性差等严重不足,为此设计基于核主成分分析和聚类分析算法的舰船网络入侵行为的检测与识别方法。首先对舰船网络入侵行为的检测的原理进行分析,并收集大量的舰船网络入侵行为检测特征。然后采用核主成分分析对舰船网络入侵行为检测特征进行选择,并通过聚类分析算法建立训练样本。最后建立舰船网络入侵行为检测与识别模型。利用标准舰船网络入侵数据集的仿真测试结果表明,本文方法不仅可以大幅度减少舰船网络入侵行为特征数量,降低舰船网络入侵行为检测的复杂度,舰船网络入侵行为检测的实时性增强,而且能够获得更高正确率的舰船网络入侵行为检测结果。

Intrusion detection is the core technology to ensure the security of warship network.At present,there are serious shortcomings in intrusion detection and recognition,such as large detection error and poor recognition accuracy.A new method of detection and recognition of warship network intrusion is designed.Firstly,the principle of detection of network intrusion of warships is analyzed,and a large number of detection features of network intrusion of warships are collected.Then,the detection features of network intrusion of warships are selected by using kernel principal component analysis,and training samples are established by clustering analysis algorithm.Finally,the detection and identification model of network intrusion of warships is established,and the intrusion data of standard network of warships are obtained.The simulation results show that this method can greatly reduce the number of characteristics of network intrusion,reduce the complexity of network intrusion detection,real-time detection of network intrusion,and obtain higher accuracy of detection results of network intrusion.