引用本文:陈峰,孙子文.工业信息物理系统数据注入攻击的事件触发弹性控制[J].控制理论与应用,2020,37(10):2134~2146.[点击复制]
CHEN Feng,SUN Zi-wen.Event-triggered elastic control of data injection attacks in industrial cyber physical systems[J].Control Theory and Technology,2020,37(10):2134~2146.[点击复制]
工业信息物理系统数据注入攻击的事件触发弹性控制
Event-triggered elastic control of data injection attacks in industrial cyber physical systems
摘要点击 2122  全文点击 757  投稿时间:2019-10-17  修订日期:2020-03-13
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2020.90867
  2020,37(10):2134-2146
中文关键词  工业信息物理系统  事件触发  卡尔曼滤波器  输入到状态稳定  数据注入攻击
英文关键词  ICPS  Event-triggered  Kalman filter  Input-to-state stability  Data injection attacks
基金项目  国家自然科学基金项目(61373126), 中央高校基本科研业务费专项资金资助(JUSRP51510), 江苏省自然科学基金资助项目(BK20131107)
作者单位E-mail
陈峰 江南大学 1015887064@qq.com 
孙子文* 江南大学 sunziwen@jiangnan.edu.cn 
中文摘要
      工业信息物理系统(ICPS)恶劣的工业环境使得无线通信网络的通信资源有限, 且通信网络的脆弱性使得系统的稳定性容易受到数据注入攻击的破坏. 本文基于输入到状态理论, 采用H1观测器、预测器和事件触发器建立基于事件触发的ICPS模型来降低通信资源的利用率, 并构造基于改进卡尔曼滤波器的攻击补偿模块进行攻击补偿, 攻击补偿模块的参数随着攻击者攻击策略不同而进行改变, 实现对数据注入攻击的弹性控制. 用matlab软件进行仿真实验, 以倒立摆作为被控对象, 模拟基于事件触发的ICPS模型受到数据注入攻击时系统的稳定性能. 仿真结果表明, 事件触发控制策略和基于改进卡尔曼滤波器的攻击补偿模块的组合不仅能够保证系统期望的稳定性, 还能有效减少数据通过无线网络传输的次数.
英文摘要
      The harsh industrial environment of industrial cyber physical systems makes the communication resources of the wireless communication network limited and the fragility of the communication network makes the stability of the system vulnerable to data injection attacks. Based on input-to-state stability theory, this paper uses H1 observer, predictor and event trigger to establish an event-triggered ICPS model to reduce the cost of communication resources. Besides, an attack compensation module based on an improved Kalman filter which reduces the impact of data injection attacks on the stability of ICPS is constructed for attack compensation. The parameters of the attack compensation module are changed with different attack strategies of adversary to achieve elastic control of data injection attacks. The simulation experiments are conducted by using Matlab tool with the inverted pendulum as the controlled object to simulate the stability of the event-triggered ICPS model under the data injection attack. Simulation results show that the combination of event-triggered control strategy and improved Kalman filter attack compensation module can not only ensure the desired stability of system, but also effectively reduce the number of times data is transmitted through the wireless network.