An Augmented Extended Kalman Filter Algorithm for Complex-Valued Recurrent Neural Networks

Author

Goh, Su Lee ; Mandic, Danilo P.

Author_Institution

Imperial Coll. London

Volume

5

fYear

2006

fDate

14-19 May 2006

Abstract

An augmented complex-valued extended Kalman filter (ACEKF) algorithm for the class of nonlinear adaptive filters realised as fully connected recurrent neural networks (FCRNNs) is introduced. The algorithm is derived based on the recent developments in augmented complex statistics, and the Jacobian matrix within the ACEKF algorithm is computed using a general fully complex real time recurrent learning (CRTRL) algorithm. This makes ACEKF suitable for processing general complex-valued nonlinear and nonstationary signals and bivariate signals with strong component correlations. Simulations on benchmark and real-world complex-valued signals support the approach

Keywords

Jacobian matrices; Kalman filters; adaptive filters; nonlinear filters; statistics; Jacobian matrix; augmented complex statistics; augmented extended Kalman filter algorithm; bivariate signals; complex real time recurrent learning; complex-valued recurrent neural networks; nonlinear adaptive filters; nonstationary signals; Adaptive filters; Algorithm design and analysis; Computational modeling; Educational institutions; Jacobian matrices; Neural networks; Recurrent neural networks; Signal processing; Statistics; Wind forecasting;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech and Signal Processing, 2006. ICASSP 2006 Proceedings. 2006 IEEE International Conference on

Conference_Location

Toulouse

ISSN

1520-6149

Print_ISBN

1-4244-0469-X

Type

conf

DOI

10.1109/ICASSP.2006.1661337

Filename

1661337