An unscented Kalman filtering approach for nonlinear singular systems

Abstract

In this study, an unscented Kalman filtering approach is proposed for nonlinear singular systems to obtain not only the estimation for the states but also for the unknown inputs presented in the measurement equations. No prior information is needed for the unknown inputs to be estimated. The formulation of the proposed approach is based on the weighted least squares estimation (LSE) and the unscented transformation (UT) methods. The restriction of the proposed approach is also mentioned. An illustrative example demonstrates that accurate and consistent state and unknown input estimations are obtained with the proposed approach.