Design of Pareto-Optimal Linear Quadratic Estimates, Filters and Controllers

Abstract

Consideration was given to the multicriterial approach to the problems of estimation, filtration, and control; among them under uncertainty in regard to the covariance of random factors. The Pareto-optimal estimates, filters, and controllers in the problems where the data can arrive from any of the more than one data sources with various statistical characteristics, as well as in the dual problems where it is required to minimize the rms deviations of more than one objective output, were designed on the basis of the Germeier convolution and the apparatus of linear matrix inequalities.