New results on stochastic stability of discrete-time Unscented Kalman Filter

Abstract

Performance of the Unscented Kalman Filter, UKF, for nonlinear stochastic discrete-time systems is investigated. It is proved that under certain conditions, the estimation error of the UKF remains bounded. Furthermore, it is shown that the design of noise covariance matrix plays an important role in improving the stability of the UKF algorithm. It is further shown the estimation error remains bounded the nonlinear observability rank condition is satisfied. These results are verified by numerical simulations for a relevant illustrative example.