A new sequential estimator for non-linear models
V.P. Leung, Lakshminarasimha Padmanabhan · International Journal of Systems Science · 1972
A new iterative scheme is presented for the sequential estimation of parameters in non-linear measurement models. The iterations lead to a limiting estimate that fully accounts for the non-linearities in the models. Comparative studies reveal that the proposed scheme is more efficient than the iterated extended Kalman filter duo to Denham and Pines. It is shown that this algorithm can be combined with a suitable prediction scheme to yield a filter for non-linear dynamical systems with discrete measurements. The estimation scheme is applied to the problem of finding the rate constants in a kinetics model.