Nonlinear Kalman Filter by Hermite-Gauss Quadrature

Abstract

Kalman filter became a popular tool for state estimation of linear dynamical systems especially due to its simplicity. However in the case of nonlinear systems its realization is more complicated. Commonly used approach called Extended Kalman filter consists in local approximation based on linearization. Another possibility is to approximate the corresponding integrals using some feasible procedures. In this paper we apply Gauss-Hermite Quadrature for state estimation of nonlinear systems and compare its accuracy with Extended Kalman filter.