A hierarchical feedforward adaptive filter for system identification

Author

Boukis, Christos G. ; Mandic, Danilo P. ; Constantinides, Anthony G.

Author_Institution

Dept. of Electr. & Electron. Eng., Imperial Coll. of Sci., Technol. & Med., London, UK

fYear

2002

fDate

2002

Firstpage

269

Lastpage

278

Abstract

An architecture for adaptive filtering based upon the previously introduced hierarchical least mean square algorithm is proposed. This pyramidal architecture incorporates sparse connections between the architectural layers with a certain variable degree of overlapping between the neighboring subfilters of the same level. A learning algorithm for this class of structures is derived, based on the back-propagation algorithm for temporal feedforward networks with linear neurons. Further, a class of normalized algorithms for this class is derived. The analysis and simulations show the proposed algorithms outperform the existing ones.

Keywords

adaptive signal processing; backpropagation; feedforward; filtering theory; identification; least mean squares methods; Taylor series expansion; adaptive filtering; architectural layers; backpropagation algorithm; global gradient descent; hierarchical LMS algorithm; hierarchical feedforward adaptive filter; hierarchical least mean square algorithm; learning algorithm; learning rate; linear neurons; normalized algorithms; output error; pyramidal architecture; real-time adaptive filtering; signal processing; simulation results; sparse connections; subfilters; system identification; temporal feedforward networks; weight updating techniques; Adaptive filters; Adaptive signal processing; Biomedical signal processing; Feedforward neural networks; Finite impulse response filter; Least squares approximation; Neural networks; Neurons; Signal processing algorithms; System identification;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks for Signal Processing, 2002. Proceedings of the 2002 12th IEEE Workshop on

Print_ISBN

0-7803-7616-1

Type

conf

DOI

10.1109/NNSP.2002.1030038

Filename

1030038