虚假数据注入攻击下的非线性网络物理系统状态估计  

State estimation of nonlinear cyber-physical systems under false data injection attacks

作　　者：魏源初 丁博 王东振 WEI Yuanchu;DING Bo;WANG Dongzhen(School of Information Engineering(School of Artificial Intelligence),Yangzhou University,Yangzhou 225127,China)

机构地区：[1]扬州大学信息工程学院(人工智能学院),江苏扬州225127

出　　处：《扬州大学学报(自然科学版)》2025年第1期73-78,共6页Journal of Yangzhou University:Natural Science Edition

基　　金：江苏省自然科学基金资助项目(BK20230559)。

摘　　要：针对传感器和执行器同时存在未知虚假数据注入攻击的非线性网络物理系统,研究其安全状态估计与攻击重构问题。针对系统中注入虚假数据时固有的不稳定性,提出一种递归三步扩展卡尔曼滤波算法,得到系统状态及攻击信号的无偏最小方差估计,并引入无限增广状态无迹卡尔曼滤波算法进一步提高估计精度。最后,通过仿真实验验证所提方法的有效性。This paper is concerned with the secure state estimation and attack reconstruction prob-lem for nonlinear cyber-physical systems under unknown false data injection attacks,where the at-tacker injects unknown malicious data into the communication networks of both sensors and actua-tors simultaneously.Firstly,in response to the inherent instability when false data is injected into the researched system,a recursive three-step extended Kalman filter is proposed to obtain the unbi-ased minimum-variance estimation of the system state and attack signal.Subsequently,to further enhance estimation accuracy,the infinity augmented state unscented Kalman filter is introduced.Ultimately,the efficacy of these approaches is substantiated via an illustrative example.

关 键 词：非线性网络物理系统 虚假数据注入攻击 安全状态估计 非线性滤波 

分 类 号：TP273.1[自动化与计算机技术—检测技术与自动化装置]