Event-triggered PMI Extended Kalman Filter and Its Application to Sensor Fault Estimation

This paper is concerned with the event-triggered fault estimation (FE) problem in combination with extended Kalman filter (EKF) and proportional gain and multi-integral (PMI) techniques. A nonlinear system with sensor fault is transformed into an augmented discrete time model, where the fault signal is described in the form of polynomials. Based on this, a new event-triggered fault estimator is designed for the augmented system so that the generated estimation error is completely decoupled from the event-triggered transmission error. In addition, the FE issue for nonlinear system is converted into the design of the estimator gain matrix, which is addressed by minimizing the estimation error covariance. Finally, a simulation study based on a three-tank system is presented to verify the applicability of the developed scheme, by comparing with the existing event-triggered EKF schemes for FE.