Couscous-constant approximation in signal filtration task

The paper considers the task of assessing state a nonlinear dynamic system, based on couscous-linear approximation non-linear functions included in and observa-tion equation. Examples method presented filtration tasks are given it is shown that use allows at least half margin sampling error by Kalman-Busey filter compared to first-order approximation. Dynamic systems processing algorithms form vector-matrix equations have been obtained for multidi-mensional systems.