FILTER BANK APPROACH TO ADAPTIVE FILTERING METHODS USING INDEPENDENT COMPONENT ANALYSIS

摘要

PURPOSE: A method for approaching a filter bank of a filter adaptive algorithm using independent component analysis is provided to improve the performance of the filter adaptive algorithm. CONSTITUTION: According to the filter adaptive algorithm using independent component analysis, each input signal is separated into a number of frequency bands through an analysis filter bank, and then a signal as many as the number of analysis filters is constituted by taking one sample per randomly fixed samples as to the filter output corresponding to each frequency band. Then, the filter adaptive algorithm using independent component analysis is applied to the whole. And a sample having 0 value less than the number of the fixed samples by 1 is inserted between samples, and then a result signal is recovered through an integrated filter bank.