Abstract
Adaptive filters are systems in digital signal processing designed to work in environments with unknown or time-varying statistical characteristics — according to non-stationary stochastic signals. In order to achieve this capability the adaptive filter's adaptation algorithm has to be able to adapt the filter to a time-dependent error criterion. I.e., given a filter topology, the filter parameters have to be estimated on-line. In this paper, a comparative study of several evolutionary algorithms — evolution strategies (ES), genetic algorithms (GA) and micro-genetic algorithms (ΜGA) — is presented for the on-line adaptation of adaptive filters. The adaptive filter's task is to predict a time-discrete and non-stationary stochastic signal. This important problem arises in a variety of applications, e.g. stochastic signal estimation, system identification and echo cancellation. The evolutionary algorithms optimize the adaptive filter's parameters on-line, with one generation corresponding to one time step of the stochastic signal. Hence, the evolutionary algorithms' ability to adapt to rapidly changing and noisy environments and to track time-varying parameters is of special interest in this study.
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
References
Baker, J.E.: An Analysis of the Effects of Selection in Genetic Algorithms. Ph.D., Vanderbilt University, 1989
Grefenstette, J.J.: Genetic Algorithms for Changing Environments. In: Männer, R.; Manderick, B. (Eds.): Parallel Problem Solving from Nature II. Amsterdam: Elsevier Science Publishers, pp. 137–144, 1992
Haykin, S.: Adaptive Filter Theory. Englewood Cliffs: Prentice Hall, 1986
Holland, J.H.: Adaptation in Natural and Artificial Systems. Cambridge: MIT Press, 1992
Krishnakumar, K.: Micro-Genetic Algorithms for Stationary and Non-Stationary Function Optimization. In: Intelligent Control and Adaptive Systems. Proc. of the SPIE, Vol. 1196, pp. 289–296, 1989
Neubauer, A.: Linear Signal Estimation Using Genetic Algorithms. In: Systems Analysis Modelling Simulation. Vol. 18/19, Switzerland: pp. 349–352, 1995
Neubauer, A.: Real-Coded Genetic Algorithms for Bilinear Signal Estimation. In: Schipanski, D. (Hrsg.): Tagungsband des 40. Internationalen Wissenschaftlichen Kolloquiums. Ilmenau: Band 1, pp. 347–352, 1995
Neubauer, A.: Non-Linear Adaptive Filters Based on Genetic Algorithms with Applications to Digital Signal Processing. In: Proc. of the 1995 IEEE International Conference on Evolutionary Computation. Perth: Vol. 2, pp. 527–532, 1995
Neubauer, A.: Genetic Algorithms for Non-Linear Adaptive Filters in Digital Signal Processing (Invited paper). In: Proc. of the 1996 ACM Symposium on Applied Computing. Philadelphia: pp. 519–522, 1996
Rechenberg, I.: Evolutionsstrategie '94. Werkstatt Bionik und Evolutionstechnik, Band 1, Stuttgart: frommann-holzboog, 1994
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1996 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Neubauer, A. (1996). A comparative study of evolutionary algorithms for on-line parameter tracking. In: Voigt, HM., Ebeling, W., Rechenberg, I., Schwefel, HP. (eds) Parallel Problem Solving from Nature — PPSN IV. PPSN 1996. Lecture Notes in Computer Science, vol 1141. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-61723-X_1026
Download citation
DOI: https://doi.org/10.1007/3-540-61723-X_1026
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-61723-5
Online ISBN: 978-3-540-70668-7
eBook Packages: Springer Book Archive