Ellipsolidal Algorithms of Guaranteed Estimation and the Recurrent Least Square Method in Problems for Filtering States of Dynamical System

摘要

The parameterization of the set-theory ellipsoidal algorithm for estimating states of a linear dynamic object by observations distorted by bounded hindrance is realized in such a way that parameters of the looked for ellipsoidal estimation minimized the additive quadratic functional. Thereby the relation between the algorithm of the ellipsoidal estimating and the classical (recurrent) lest square method is I set.