An Event-triggered Student’s t Extended Kalman Filter Under Multiple Attacks

‍ ‍Wireless network systems are susceptible to various attacks such as DoS （Denial of Service Attacks） and FDI （False Data Injection Attacks）， which disrupt the integrity and effectiveness of sensor measurement data. They reduce the accuracy of filter state estimation， and even cause filter divergence. Moreover， the system modeling error， multipath propagation and other factors in the data transmission process make the process noise and measurement noise will not obey the standard Gaussian distribution， showing heavy-tail characteristics. These factors will reduce the estimation performance of filters. Aiming at the problem of state estimation for a class of networked nonlinear systems with heavy tailed noise distribution， considering the poor performance of filter estimation caused by multiple attacks， such as DoS attacks and FDI attacks， an anti-attack extended Kalman filter based on the Student’s t distribution is proposed. Firstly， the Student’s t distribution is used to approximate the probability density function of the heavy tailed noise. Two binary random variables are introduced to represent DoS attacks and FDI attacks， based on the estimation error covariance upper bound minimization method， the optimal filter gain matrix is derived. Then， the Student’s t extended Kalman filter under multiple attacks is obtained； secondly， considering the limited bandwidth and the issues of channel congestion and information loss faced during data transmission， an event trigger mechanism is introduced to improve the channel utilization； finally， simulations verify the effectiveness of the proposed method.