Design And Analysis Of Stealthy Deception Attacks Against Networked Systems

Networked control systems,which embed communication technology into control system,have been applied into all sorts of fields,such as smart vehicles,smart factories,smart grids,space exploration and intelligent building systems.However,due to the openness of the Internet,NCSs are easily intruded by various malicious attackers,which usually causes huge losses for the society.And stealthy FDI attacks can cause losses for NCSs meanwhile avoiding the detection.Therefore,a class of linear time invariant networked control systems with multiple inputs and multiple output are considered in this paper.From the perspective of security,the design of FDI attacks and the resilient control method are studied in networked control systems.The specific contents are as follows:(1)For the problem that NCSs cannot run stably under FDI attacks against control instruction,a resilient control method is designed to ensure that the system can run stably under FDI attack.That is,an AKF is designed to estimate the system state and attack signal simultaneously,and a resilient controller is designed to eliminate the impact of the FDI attack based on estimation values.The effectiveness of the method is verified by theoretical analysis and numerical simulation.(2)For NCSs with the resilient control method mentioned above,a new stealthy FDI attack strategy is designed from the perspective of attacker.That is,the attacker,which with a model similar to the system,estimates the system state and designs the stealthy FDI attack based on estimation values.The goal is to break the system without being detected by the detector.The stealthy condition of the FDI attack is obtained through theoretical analysis.Finally,the corresponding stability analysis and numerical simulation verify the effectiveness of the proposed FDI attack method.(3)A novel stealthy FDI attack strategy is designed,for the problem that the attacker cannot always hold the same such knowledge as the defender due to the information privacy.And the attack effect is analyzed when the attacker and the defender hold different system knowledge.The main content is summarized as follows: 1)The attacker and defender hold the same system knowledge.2)Based on the case 1,the defender actively modifies Kalman filter gain,that is,the attacker and defender hold different filter gain.3)the attacker designs the attack signal based on the inaccurate system model.The stealthy condition of the proposed attack scheme is derived and proven.Theoretical derivation and numerical simulation verify the stealthiness of the attack strategy.