A new method for state estimation of dynamic system based on Dempster Shafer theory

The goal of state estimation method is to compute an accurate estimation of the state of the system based on the measurement given by different sensors and a mathematical representation of the system. In this paper a new state estimation method based on Dampster-Shafer theory and interval analysis is presented. This method uses belief structures composed of a finite number of axis-aligned boxes with associated masses. Such belief structures can model partial information on model and measurement uncertainties, more accurately than the bounded error approach alone. Focal sets are propagated in the system equations using tools from interval arithmetics and constraint satisfaction techniques, thus generalizing pure interval analysis. The results of applying the proposed method on a vehicle localization problem show the usefulness of the proposed method.