Simultaneous state and fault estimation based on improved adaptive Kalman filter

This paper presents an improved adaptive Kalman filter (AKF) to obtain state and fault estimation for linear time-varying (LTV) discrete stochastic systems. In the stochastic framework, the fault is modeled as parameter change, then the improved AKF is introduced to estimate state and fault. The exponential convergency of the improved AKF is proved. Numerical simulation is used to verify the proposed approach.