非线性系统下的虚假信息注入攻击研究

Research on Disinformation Injection Attacks in Nonlinear Systems

本文提出了一种基于扩展卡尔曼滤波算法的线性攻击策略。在单传感器网络的非线性系统框架下,通过向数据传输过程中被截获的新息序列注入虚假数据,从而达到攻击该网络系统、使其性能下降的目的。这样不仅有效降低了系统的稳定性,并且能高效地规避被虚假数据检测器发现的风险。最后利用扩展卡尔曼滤波推导出攻击过程中远端估计器的误差协方差矩阵,并且通过误差协方差矩阵的演化合理地展现出系统性能的退化。

In this paper, we present a linear attack strategy based on the Extended Kalman Filter (EKF) algorithm. Within the framework of a single sensor network with nonlinear systems, the objective is to attack the network system and degrade its performance by injecting false data into the intercepted new information sequences during the data transmission process. This not only effectively reduces the stability of the system but also efficiently avoids the risks detected by false data detectors. Finally, utilizing the Extended Kalman Filter, the error covariance matrix of the remote estimator during the attack process is derived, demonstrating the degradation of system performance reasonably through the evolution of the error covariance matrix.