Modification of the Kalman Filter with Residual Separation and Localization

Abstract

A modification of the Kalman filter with residual separation and localization is proposed, which makes it possible to develop an estimate of the measured and restored variables affected by noise under conditions of inaccurately known noise intensities and parametric uncertainty of the plant model. The results of mathematical modeling are presented, demonstrating the advantages of the proposed modification over the classical approach to constructing a Kalman filter.