Robust, exponentially fast state estimator for some non-linear stochastic systems

Abstract

The design of exponentially fast state estimators for a class of non-linear stochastic systems with non-linear noisy measurement models is considered. First, finite-time exponential data-weighting filtering is taken up, then it is shown for systems with mean square stability that by using this technique, the design of lime-invariant filters with guaranteed exponential convergence rate is possible, A robustness property of these estimators is also demonstrated.