Multiple model approach for robust state estimation in presence of model uncertainty and bounded disturbances

Abstract

In the present work, an optimization-based algorithm for state estimation under model uncertainty and bounded disturbances is presented. In order to avoid to solve a non-convex optimization problem, model and state estimation problems are divided into two convex formulations which are solved within a fixed-point iteration scheme with standard available solvers. Guaranty of robust global stability is given for the case of bounded disturbances and uncertainty, and convergence to the true system and vector state are given for the case of vanishing disturbances.