Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
6857131 | Information Sciences | 2016 | 27 Pages |
Abstract
In this paper, we study the state estimation problem for a class of nonlinear Markov jump systems with the moving horizon estimation algorithm, which is an optimization-based filtering method. The optimal estimate is obtained by minimizing a quadratic estimation cost function defined on fixed sliding horizon measurements. For this purpose, the quadratic estimation cost function is formulated from the negative logarithm of the joint states distribution, whose particular factorization is represented by a Bayesian network. By analyzing the corresponding full information estimation as a basic problem for nonlinear Markov jump systems, the formulation of the moving horizon estimation is developed. An example is presented to compare the proposed new estimation technique with classical interacting multiple model particle filtering, and show the effectiveness and advantages of the new estimation scheme.
Related Topics
Physical Sciences and Engineering
Computer Science
Artificial Intelligence
Authors
Qing Sun, Cheng-Chew Lim, Peng Shi, Fei Liu,