Window length selection in linear receding horizon filtering

Abstract

A receding horizon filtering problem for linear systems with norm-bounded time-varying uncertainties is considered. The main goal of this paper is to choose the reasonable window length (WL) which enables us to adjust the modeling uncertainty considering not only computational cost but also the accuracy. The decision of the WL is an important step for receding horizon filter designing. This paper presents a novel algorithm for decision of optimal WL. Two methods are proposed. The first algorithm decides the optimal WL by considering lower bound and upper bound of the uncertainty. Secondly hybrid approach which is the combination of the Kalman filter and the optimal receding horizon filter for suitable situations respectively. The performance of the receding horizon filter with proposed WL is illustrated and compared to other finite memory filters.