ADAPTIVE MICROSTATISTIC VOLTERRA FILTERS

Dušan Kocur - Imrich Hendel

   The microstatistic Volterra filters (MVF) are based on the idea of generalization of a class of conventional microstatistic filters (CMF) by substituting of Wiener filters (WF) applied in CMF by Volterra filters (VF). In this paper, a new class of adaptive filters called adaptive MVF is introduced. The well-known LMS and RLS algorithms are proposed for the adaptive MVF adaptation. By using computer experiment it is shown that the adaptive MVFs have a superior ability to identify nonlinear systems. In this paper, a short description of the basic properties of the LMS and RLS adaptive MVF is also given.

Keywords: Volterra filter, microstatistic filter, threshold decomposition

[full-paper]