Minimax linear filtering of random sequences with uncertain covariance function

Abstract

Consideration was given to the development of a numerical method for determination of the minimax filter in the linear stochastic difference system studied over a finite horizon in the presence of an uncertain covariance function in the model of useful signal. Selection of the considered uncertainty sets relied on the form of the corresponding confidence regions. The developed iterative procedure was applied to filtering of the position of a maneuvering target with inexactly given acceleration covariance function.