A Reduced-Order Extended Kalman Filter by Local Minimization of the Error Covariance

Abstract

Abstract. This work introduces a new reduced-order extended Kalman filter (REKF) for discrete nonlinear dynamic systems, which is derived by least-square minimization of the local estimation error covariance. Both the filter and the one-step ahead predictor forms are derived. For demonstration purposes, the REKF is applied to a polynomial nonlinear system, and its estimation error is compared to that of the full-order extended Kalman filter.