The model prediction-based two-stage state estimation algorithm

Abstract

For nonlinear coupled systems, the traditional filter algorithm has the disadvantages of large amount of calculation and insufficient estimation accuracy. In order to reduce computational complexity and improve estimation accuracy, this paper proposes a model prediction-based two- stage state estimation algorithm. Firstly, a two-level subsystem is established through model decoupling and reconstruction. Then the linear state estimation is obtained base on first-level linear subsystem by Kalman Filter (KF). Secondly, the predictive model is obtained by substituted the first-level state estimation into the second-level subsystem. The second-level nonlinear state estimation is obtained base on prediction model and KF. Finally, the state estimation of the nonlinear system is transformed into the state estimation of the linear system by the proposed two-stage state estimation filter. And in the end, the effectiveness of the filtering algorithm is verified by simulations.