A New Diagonally Implicit Implementation of the Continuous-Discrete Extended Kalman Filters for State and Parameter Estimation of Nonlinear Dynamic Systems

An efficient and robust implementation of the Extended Kalman Filter (EKF) dedicated to the real-time estimation of states and parameters of nonlinear and under-sampled dynamic systems is studied in this article. The EKF is here considered in its continuous-discrete (CD) form, which enhances the accuracy of the estimation but requires the numerical integration of a nonlinear continuous-time model. Several efficient integration schemes and implementations of the EKF are proposed in this article, and applied to the parameter estimation of a High Speed Permanent Magnet Synchronous Motor (HSPMSM) as well as several highly nonlinear models, with computation time and accuracy as main comparison criteria. Compared to classical EKF, the proposed implementation offers a good stability even at low sampling rates, and a lower computation time for equal accuracy.