Convergence analysis of block orthogonal projection and affine projection algorithms

Authors:
Kazushi Ikeda
Affiliations:
Department of Systems Science, Graduate School of Informatics, Kyoto University, Sakyo, Kyoto 606-8501, Japan
Venue:
Signal Processing
Year:
2002

Citing 3
Cited 0

Adaptive filter theory (3rd ed.)

Adaptive filter theory (3rd ed.)
Exact convergence analysis of affine projection algorithm: the finite alphabet inputs case

ICASSP '99 Proceedings of the Acoustics, Speech, and Signal Processing, 1999. on 1999 IEEE International Conference - Volume 03
A family of adaptive filter algorithms with decorrelatingproperties

IEEE Transactions on Signal Processing

Quantified Score

Hi-index	0.08

Visualization

Abstract

The affine projection (AP) and the block orthogonal projection (BOP) algorithms are known to have good convergence properties even when the input signal is correlated. However, their convergence rates are not elucidated yet, since they are based on orthogonal projection, and are too complicated to be analyzed. In this paper, we first consider their geometrical view and derive their convergence rates for white signals. Moreover, we provide an approximation method to evaluate the convergence rate for correlated signals and show the results which are confirmed by computer simulations. The results imply that the convergence rate of the BOP algorithm gets saturated as the block size increases.