Optimal Partitioned State Kalman Estimator

A new general two-stage algorithm was originally proposed to reduce the computational effort of the augmented state Kalman estimator. The conventional input estimation techniques assume constant input level and there are not covered a generalized input modeling. In this paper an innovative scheme is developed to overcome these drawbacks by using a new partitioned input dynamic modeling. In addition, authors propose a modified two-stage Kalman estimator with a new structure, which is an extension of the conventional input estimation techniques and is optimal for general, linear discrete-time systems.