基于分布式检测与告警切换的多传感器CPS安全状态估计  

Secure state estimation for multi-sensor CPS with distributed detection and alarm switching

作　　者：杨彪 贺光[1] 辛亮 龙志强[1] YANG Biao;HE Guangy;XIN Liang;LONG Zhi-qiang(College of Intelligence Science and Technology,National University of Defense Technology,Changsha Hunan 410073,China)

机构地区：[1]国防科技大学智能科学学院,湖南长沙410073

出　　处：《控制理论与应用》2024年第11期2121-2130,共10页Control Theory & Applications

基　　金：国家自然科学基金项目(62003356);国家自然科学基金“叶企孙”科学基金重点项目(U2241214)资助.

摘　　要：针对多传感器信息物理系统(CPS)在执行器和传感器遭受隐蔽性攻击情况下的安全状态估计问题,本文提出一种基于分布式检测与告警切换的安全状态估计方法.首先,利用卡尔曼滤波器组,构建基于状态差异的残差信号,用于多传感器的分布式检测;其次,定义传感器置信因子,建立基于检测告警切换的安全状态估计框架,设计攻击情况下的攻击隔离与状态估计一体化算法;最后,利用四驱麦克纳姆轮移动机器人,搭建了多传感器CPS实验平台,并分别在零动态攻击、隐蔽攻击和重放攻击场景下进行实验验证,结果表明,本文提出的安全估计方法具有较好的可行性和有效性.A method for secure state estimation of multi-sensor cyber-physical systems(CPS)under stealthy attacks on the actuators and sensors is proposed based on distributed detection and alarm switching.By using a group of Kalman filters,a novel residual signal based on state differences is constructed for distributed detection of multiple sensors.The sensor confidence factor is defined,and a secure state estimation framework based on detection and alarm switching is established.Based on this,an integrated algorithm for attack isolation and state estimation is designed under attack scenarios.Finally,a multi-sensor CPS experimental platform is built using a four-wheel mobile robot,and experiments are performed in zerodynamics attack,covert attack,and replay attack scenarios,respectively.The results show that the secure state estimation method proposed in this paper has good feasibility and effectiveness.

关 键 词：多传感器CPS 隐蔽性攻击 状态差异残差 分布式检测 安全状态估计 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]