假数据注入攻击下信息物理融合系统的稳定性研究

摘要

Due to the stealthiness behavior, false data injection (FDI) attacks severely threaten the security of cyberphysical systems (CPS) . From the attackers perspective, this paper mainly studies how FDI attacks impact the stability of CPS. First, we give the FDI attack model where the false control and measurement data are injected into the forward and feedback channels, respectively. Then, we propose an FDI effectiveness model to quantify the attack impact on the state estimation and measurement residue of CPS. On this basis, we design a coordination strategy associated with attack vector and further derive the theoretical attack conditions to manipulate the stability of CPS, which are related to the stability of attack matrix H and system matrix A and the selected moment of time parameter ka. Finally, numerical simulations indicate that FDI attacks can effectively manipulate the stability of CPS including two classes of controlled plants: stable and unstable. This study further reveals the coordination behavior of FDI attacks, which provides important reference for securing the CPS and defending cyber attacks.%假数据注入 (False data injection, FDI) 攻击由于其隐蔽性特点, 严重威胁着信息物理融合系统 (Cyber-physical systems, CPS) 的安全.从攻击者角度, 本文主要研究了FDI攻击对CPS稳定性的影响.首先, 给出了FDI攻击模型, 从前向通道和反馈通道分别注入控制假数据和测量假数据.接着, 提出了FDI攻击效力模型来量化FDI攻击对CPS状态估计值和测量残差的影响.在此基础上, 设计了一个攻击向量协同策略, 并从理论上分析出操纵CPS稳定性的攻击条件:攻击矩阵H和系统矩阵A的稳定性及时间参数ka的选取时机.数值仿真结果表明FDI攻击协同策略能够有效操纵两类 (含有稳定和不稳定受控对象) 系统的稳定性.该研究进一步揭示了FDI攻击的协同性, 对保护CPS安全和防御网络攻击提供了重要参考.