BUILDING AN ADAPTIVE ALGORITHM FOR NON-LINEAR NONRECURSIVE FILTERS BY LEAST SQUARES METHOD

Abstract

The algorithm of an evaluation of factors of nonrecursive filters, founded on the leastsquares method, with application of linearization by means of adaptive thresholds of linearity is offered, is demonstrated, the introdution that of adaptive thresholds of linearity in algorithm ofidentification of the non-linear nonrecursive filter allows considerably to reduce time of adaptingas contrasted to by algorithms which are not taking into account linear properties of the filter atsmall input action. In the end, the offered algorithm allows to receive better quality of a filtration non- Gaussian signals on a background additive a noise of Gaussian as constrasted to by optimum linear filters of the Wiener. From results that proved in the article, an algorithm with capability of causing an adaptiveeffect for nonrecursive filters was completely built. It showed a great capability of the nonlinear nonrecursive filter in comparison with the optimum linear filter in nearby - signal noise elimination