Analysis And Design Of State Observer For Uav Cyber-physical Systems Under Network Attack

Unmanned aerial vehicle information physical system is a highly complex system that integrates communication network,intelligent control,data calculation and physical entities.Most of the sensors used in the UAV information physical system are wireless sensors.Because the communication network of the wireless sensor is open,the communication network in the UAV information physical system is more vulnerable to malicious network attacks.State observation of UAV information physical system is an important part of UAV control technology.When the UAV information physical system is physically attacked,the design of state observer for the UAV information physical system will change accordingly.Common network attacks include data injection attacks,denial of service attacks and replay attacks.Each network attack has a different impact on communication in the UAV Information Physical System.Therefore,in this paper,the design of state observer for UAV information physical system under data injection attack,the state observer for UAV information physical system under denial of service attack and the state observer for UAV information physical system under multiple network attacks are discussed respectively,and the feasibility of the algorithm is illustrated by numerical simulation.The status observation research of UAV Information Physics System under network attack mainly includes the following aspects:(1)The influence of status data transmitted by UAV sensor when several network attacks are applied in UAV information physical system is analyzed.Several network attacks and noise disturbances(system noise and measurement noise)which affect the UAV information physical system are modeled.At the same time,Kalman filter is used to estimate the state of the UAV information physical system in both cases.(2)A state observer for UAV information physical system under data injection attack is designed.Based on the measurement vectors of the current moment and the state estimation vectors of the previous moment,the triggering mode of UAV information physical system under data injection attack is designed.The state observer of the UAV information physical system with gain feedback matrix is designed,and then the optimal solution of the gain feedback matrix is derived.(3)A state observer for UAV information physical system under denial of service attack is designed.First,the state observer of the UAV information physical system under denial of service attack is designed to minimize the covariance of the UAV state estimation during the design process.Secondly,the impact of two covariances on the system during the design of the state observer is analyzed,and the covariance calculation method of the unmanned aerial vehicle status information in the case of denial of service attack is designed.Finally,a state observer algorithm for UAV information physical system without network attack is designed using the method of multiple overlay estimation.(4)A state observer for UAV information physical system under multiple network attacks is designed.First,use the method of adding auxiliary information to determine if the UAV Information Physical System is under replay attack.Then,based on the algorithms in Chapters III and IV,a state observer algorithm for detecting data injection attacks under multiple network attacks and a state observer algorithm for detecting denial of service or replay attacks under multiple network attacks are designed.