Filtering problem for the non-stationary random processes with unknown correlation functions

Abstract

We study the problems of the linear mean-square filtration of non-stationary random processes with unknown correlation functions. We consider the filtering problem for the general basic model and the case of building the linear meansquare estimation with the special type of the estimation. We investigate case when unknown correlation functions belong to certain special sets too. We offered algorithms for building optimal linear mean-square estimation for these cases.