Risk sensitive estimation using cubature quadrature points for system with unmodeled bias

Abstract

In this paper risk sensitive estimation based upon cubature quadrature rule for system with model uncertainty due to unmodeled bias is discussed. The system used here is three dimensional discrete time Lorenz attractor, one of the classic icons in non-linear dynamics. A brief discussion over risk sensitive estimation has been given. Risk sensitive estimation using cubature quadrature rule has been discussed. The improvement in performance of this risk sensitive estimator in comparison to its risk neutral counterpart has been shown for system with unmodeled bias.