关键词: 工业控制系统安全态势, 层次化分析法, 深度学习, 多源信息融合, 非线性时间序列预测

Abstract: Aiming at the problem that the security protection of industrial control system lacks dynamic defense, situation awareness and other means, combined with the characteristics of the fixed business of industrial control system, this paper proposes the framework of security situation assessment and prediction, and constructs a security situation measurement model based on the analytic hierarchy process. Then, this paper studies the construction and extraction of behavior profile features of industrial control system, security attack detection based on deep learning, security event analysis based on big data mining and other technologies, and proposes security situation assessment based on information fusion and security situation prediction based on statistical learning theory. Finally, the validity and accuracy of the security situation assessment method are verified by experiments.

Key words: Industrial control system security situation, Analytic hierarchy process, Deep learning, Multi-source information fusion, Nonlinear time series prediction