Convergence Behaviour of LMS-Type Algorithms for Adaptive Noise Control in Noisy Doppler Environments (862kB)
Convergence Behaviour of LMS-Type Algorithms for Adaptive Noise Control in Noisy Doppler Environments (862kB)
This paper discusses the convergence and tracking behaviour of LMS-type algorithms in a certain type of environment, which is characterised by a Doppler shift in frequency between the two signals available to the algorithm and rapid variations in signal power. We show the linear time-varying characteristics of the underlying system and derive optimum trajectories to which we can compare the adaptation and tracking ability of first order LMS and NLMS adaptive filters. We also present simulations using higher filter orders and real world noise, for which particular emphasis is put on the presence of observation noise. An excursion into the theory of non-stationary convergence and tracking of adaptive algorithms provides justification for the observed behaviour of the algorithms.
Stewart, R W
c819965d-70b4-4689-8206-e6e444dcbf8e
Weiss, S
a1716781-351d-41d2-8d67-3e3d34f16476
Crawford, D H
f17e4330-3b43-4fc5-a120-cd37a30a1efc
September 1996
Stewart, R W
c819965d-70b4-4689-8206-e6e444dcbf8e
Weiss, S
a1716781-351d-41d2-8d67-3e3d34f16476
Crawford, D H
f17e4330-3b43-4fc5-a120-cd37a30a1efc
Stewart, R W, Weiss, S and Crawford, D H
(1996)
Convergence Behaviour of LMS-Type Algorithms for Adaptive Noise Control in Noisy Doppler Environments (862kB).
Third International Symposium on Methods and Models in Automation and Robotics.
Record type:
Conference or Workshop Item
(Other)
Abstract
This paper discusses the convergence and tracking behaviour of LMS-type algorithms in a certain type of environment, which is characterised by a Doppler shift in frequency between the two signals available to the algorithm and rapid variations in signal power. We show the linear time-varying characteristics of the underlying system and derive optimum trajectories to which we can compare the adaptation and tracking ability of first order LMS and NLMS adaptive filters. We also present simulations using higher filter orders and real world noise, for which particular emphasis is put on the presence of observation noise. An excursion into the theory of non-stationary convergence and tracking of adaptive algorithms provides justification for the observed behaviour of the algorithms.
Text
stewart96a.pdf
- Other
Other
stewart96a.ps
- Other
More information
Published date: September 1996
Additional Information:
Invited Paper Address: Szczecin, Poland
Venue - Dates:
Third International Symposium on Methods and Models in Automation and Robotics, 1996-09-01
Organisations:
Electronics & Computer Science
Identifiers
Local EPrints ID: 251906
URI: http://eprints.soton.ac.uk/id/eprint/251906
PURE UUID: f9704751-04c5-4557-905a-760916c5879c
Catalogue record
Date deposited: 08 Nov 2000
Last modified: 14 Mar 2024 05:13
Export record
Contributors
Author:
R W Stewart
Author:
S Weiss
Author:
D H Crawford
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics