A TREE-STRUCTURED PIECEWISE-LINEAR ADAPTIVE FILTER

被引：14

作者：

GELFAND, SB ^{[1
]}

RAVISHANKAR, CS ^{[1
]}

机构：

[1] COMSAT CORP,DEPT VOICEBAND PROC,CLARKSBURG,MD 20871

来源：

IEEE TRANSACTIONS ON INFORMATION THEORY | 1993年 / 39卷 / 06期

基金：

美国国家科学基金会;

关键词：

ADAPTIVE FILTERING; NONLINEAR FILTERING; TREE-STRUCTURED METHODS; STOCHASTIC GRADIENT ALGORITHMS; DEPENDENT DATA; PENALTY METHODS; CONVERGENCE ANALYSIS; ORDER ANALYSIS; ECHO CANCELLATION;

D O I：

10.1109/18.265499

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

We propose and analyze a new architecture for nonlinear adaptive filters. These nonlinear filters are piecewise linear filters obtained by arranging linear filters and thresholds in a tree structure. A training algorithm is used to adaptively update the filter coefficients and thresholds at the nodes of the tree, and to prune the tree. The resulting tree-structured piecewise linear adaptive filter inherits the robust estimation and fast adaptation of linear adaptive filters, along with the approximation and model-fitting properties of tree-structured regression models. A rigorous analysis of the training algorithm for the tree-structured filter is performed. Here, some new techniques are developed for analyzing hierarchically organized stochastic gradient algorithms with fixed gains and nonstationary dependent data. Simulation results show the significant advantages of the tree-structured piecewise linear filter over linear and polynomial filters for adaptive echo cancellation.

引用

页码：1907 / 1922

页数：16

共 23 条

[1] NON-LINEAR ECHO CANCELLATION OF DATA SIGNALS [J].