The Research On Security Control Of Cyber Physical System Under False Data Injection Attack

The Cyber-physical System(CPS)is vulnerable to cyber attacks due to the randomness of sensor node deployment of execution layer and the openness of the communication networks,such as False Data Injection(FDI)attacks.This thesis studies the security control of CPS under FDI attacks.The specific contents are as follows:First of all,considering the Cyber-physical System state estimation scenario,it is assumed that a residue-based detection algorithm is used at the internal system to detect data anomalies.A novel packet-reordering FDI attack strategy where the attacker is able to change the order of the data packets without being detected by the detection mechanism,thereby the accuracy of the state estimation is affected.In addition,the evolution of the estimation error covariance is given,and the degradation of system performance under the proposed attack is analyzed.And using the terminal estimation error covariance as a performance metric,the sufficient conditions for the best attack strategy are obtained.Secondly,the attack model is established,and a detection method combining the chi-square and the similarity detection is proposed aiming at the impact of attacks on CPS.The state variables of the system are estimated by the Kalman filter,and the estimated values are compared with the observations of the system by the chi-square detector and the similarity detector to determine whether an attack has occurred.The chi-square detector can detect most attacks and faults in the control system,but it has been shown that cannot detect FDI attacks due to the high concealment of it.A method of introducing similarity detection by the model of chi-square detector is proposed,which can effectively determine whether the system has FDI attack intrusion behavior and improve the security of CPS.Finally,in order to ensure that CPS can still operate stably when subjected to FDI attacks,a hybrid controller composed of multiple sub-controllers is proposed inspired by the hybrid system model.Because of each sub-controller is designed for a specific type of network attack,switching between the sub-controllers can be achieved by designing the switching logic between the sub-controllers.The performance degradation of the system then can be effectively alleviated by switching to the most optimal sub-controller according to comparing the worst performance of each subcontroller under FDI attacks based on the reconstructed past attack sequence as well as the system's current state.The simulation results show that the proposed detection method can effectively detect FDI attacks and the effectiveness of the designed control strategy is proved.