Estimating the degree of time variance in a parametric model

A method capable of directional considerations in tracking parameters which exhibit different time-variant characteristics is suggested. The method, w hich can be considered an ad hoc modification of the recursive Kalman filte r, is suited to estimate the parameters in a parametric model and, unlike t he Kalman filter, does not require any prior knowledge of the variations. T he efficiency of the proposed method is illustrated through simulated examp les as well as by an application to a full-scale industrial problem. (C) 20 00 Elsevier Science Ltd. All rights reserved.