A moving horizon estimation approach to constrained linear system with uncertain model

• Published in: 2003 Conference American Control Vol. 3; pp. 2726 - 2731 vol.3
• Main Authors: Wang, Zhao, Liu, Zhiyuan, Pei, Run, Ban, Xiguang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 2003
• Subjects: Covariance matrix; Filtering algorithms; Kalman filters; Linear systems; Nonlinear filters; Robust control; Robustness; Signal processing algorithms; Stability; State estimation
• ISBN: 9780780378964; 0780378962
• ISSN: 0743-1619
• DOI: 10.1109/ACC.2003.1243491

In the framework of moving horizon strategy, a robust estimation problem is formulated as a min-max problem subject to system dynamics and constraints on state and disturbance. In this paper two algorithms of the state estimation for the constrained linear system with an uncertain model are presented. First, we present an approximate recursive covariance matrix for the min-max problem with moving horizon N=1. Then a new recursive covariance matrix algorithm for the worst-case of the uncertain system is discussed and the covariance matrix is proved bounded for the unconstrained linear system. Simulation results show that the robust moving horizon estimation proposed in this paper is effective for constrained linear systems with uncertain model.