Finite-Time Mean-Square Boundedness, H_2, and H_∞-Properties of the Discrete-Time Extended Kalman Filter for Systems with Independent Sensor Failures

摘要

The discrete-time extended Kalman filter designed for systems with independent sensor failures is analyzed for its finite-time mean-square boundedness as a function of initial conditions and its ability to attenuate effects of finite-energy disturbances as a function of the disturbance magnitudes. In addition to this extended Kalman filter being designed as a locally optimal minimum variance estimator, this work proves that it also has local finite-time mean-square boundedness of the estimation error, stochastic H_2, and H_∞ properties. Simulations are provided that analyze the effect of the initial estimation error and the disturbance magnitude.