A Comparative Study of Finite Word Length Coefficient Optimization of FIR Digital Filters

Gurvinder S. Baicher; Meinwen Taylor; Hefin Rowlands

doi:10.1007/11539902_108

A Comparative Study of Finite Word Length Coefficient Optimization of FIR Digital Filters

Gurvinder S. Baicher, Meinwen Taylor, Hefin Rowlands

Research output: Chapter in Book/Report/Conference proceeding › Other chapter contribution › peer-review

Abstract

The accuracy of a real-time digital filter frequency response is affected by the finite word length (FWL) constraint of coefficients used in its implementation. In this paper, we consider the FWL problem in regard to the finite impulse response (FIR) digital filters. Some theoretical issues and statistical error bound conditions of the maximum deviation between the exact and the approximate magnitude responses are also considered. We use real-valued genetic algorithms (GA) as an optimisation tool and derive results for the maximum error bounds and error deviation due to FWL effects for a number of design examples. Finally, a comparison is drawn between the simply rounded, the GA optimised, integer programming and the simple hill climber methods.

Original language	English
Title of host publication	Advances in Natural Computation
Subtitle of host publication	First International Conference, ICNC 2005, Changsha, China, August 27-29, 2005, Proceedings, Part III
Editors	Lipo Wang, Ke Chen, Yew Soon Ong
Place of Publication	Berlin
Publisher	Springer
Pages	877-882
ISBN (Electronic)	978-3-540-31863-7
ISBN (Print)	978-3-540-28320-1
DOIs	https://doi.org/10.1007/11539902_108
Publication status	Published - 2005
Event	First International Conference on Natural Computation - Changsha, China Duration: 27 Aug 2005 → 29 Aug 2005

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer
Volume	3612
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	First International Conference on Natural Computation
Abbreviated title	ICNC 2015
Country/Territory	China
City	Changsha
Period	27/08/05 → 29/08/05

Access to Document

10.1007/11539902_108Licence: CC BY-NC-ND

Cite this

Baicher, G. S., Taylor, M., & Rowlands, H. (2005). A Comparative Study of Finite Word Length Coefficient Optimization of FIR Digital Filters. In L. Wang, K. Chen, & Y. S. Ong (Eds.), Advances in Natural Computation: First International Conference, ICNC 2005, Changsha, China, August 27-29, 2005, Proceedings, Part III (pp. 877-882). [Chapter 108] (Lecture Notes in Computer Science ; Vol. 3612). Springer. https://doi.org/10.1007/11539902_108

Baicher, Gurvinder S. ; Taylor, Meinwen ; Rowlands, Hefin. / A Comparative Study of Finite Word Length Coefficient Optimization of FIR Digital Filters. Advances in Natural Computation: First International Conference, ICNC 2005, Changsha, China, August 27-29, 2005, Proceedings, Part III. editor / Lipo Wang ; Ke Chen ; Yew Soon Ong. Berlin : Springer, 2005. pp. 877-882 (Lecture Notes in Computer Science ).

@inbook{08e5b7c7fbdf463cb5546e9a4915d5ff,

title = "A Comparative Study of Finite Word Length Coefficient Optimization of FIR Digital Filters",

abstract = "The accuracy of a real-time digital filter frequency response is affected by the finite word length (FWL) constraint of coefficients used in its implementation. In this paper, we consider the FWL problem in regard to the finite impulse response (FIR) digital filters. Some theoretical issues and statistical error bound conditions of the maximum deviation between the exact and the approximate magnitude responses are also considered. We use real-valued genetic algorithms (GA) as an optimisation tool and derive results for the maximum error bounds and error deviation due to FWL effects for a number of design examples. Finally, a comparison is drawn between the simply rounded, the GA optimised, integer programming and the simple hill climber methods.",

author = "Baicher, {Gurvinder S.} and Meinwen Taylor and Hefin Rowlands",

year = "2005",

doi = "10.1007/11539902_108",

language = "English",

isbn = "978-3-540-28320-1",

series = "Lecture Notes in Computer Science ",

publisher = "Springer",

pages = "877--882",

editor = "Lipo Wang and Ke Chen and Ong, {Yew Soon}",

booktitle = "Advances in Natural Computation",

address = "Germany",

note = "First International Conference on Natural Computation, ICNC 2015 ; Conference date: 27-08-2005 Through 29-08-2005",

}

Baicher, GS, Taylor, M & Rowlands, H 2005, A Comparative Study of Finite Word Length Coefficient Optimization of FIR Digital Filters. in L Wang, K Chen & YS Ong (eds), Advances in Natural Computation: First International Conference, ICNC 2005, Changsha, China, August 27-29, 2005, Proceedings, Part III., Chapter 108, Lecture Notes in Computer Science , vol. 3612, Springer, Berlin, pp. 877-882, First International Conference on Natural Computation, Changsha, China, 27/08/05. https://doi.org/10.1007/11539902_108

A Comparative Study of Finite Word Length Coefficient Optimization of FIR Digital Filters. / Baicher, Gurvinder S.; Taylor, Meinwen; Rowlands, Hefin.

Advances in Natural Computation: First International Conference, ICNC 2005, Changsha, China, August 27-29, 2005, Proceedings, Part III. ed. / Lipo Wang; Ke Chen; Yew Soon Ong. Berlin : Springer, 2005. p. 877-882 Chapter 108 (Lecture Notes in Computer Science ; Vol. 3612).

Research output: Chapter in Book/Report/Conference proceeding › Other chapter contribution › peer-review

TY - CHAP

T1 - A Comparative Study of Finite Word Length Coefficient Optimization of FIR Digital Filters

AU - Baicher, Gurvinder S.

AU - Taylor, Meinwen

AU - Rowlands, Hefin

PY - 2005

Y1 - 2005

N2 - The accuracy of a real-time digital filter frequency response is affected by the finite word length (FWL) constraint of coefficients used in its implementation. In this paper, we consider the FWL problem in regard to the finite impulse response (FIR) digital filters. Some theoretical issues and statistical error bound conditions of the maximum deviation between the exact and the approximate magnitude responses are also considered. We use real-valued genetic algorithms (GA) as an optimisation tool and derive results for the maximum error bounds and error deviation due to FWL effects for a number of design examples. Finally, a comparison is drawn between the simply rounded, the GA optimised, integer programming and the simple hill climber methods.

AB - The accuracy of a real-time digital filter frequency response is affected by the finite word length (FWL) constraint of coefficients used in its implementation. In this paper, we consider the FWL problem in regard to the finite impulse response (FIR) digital filters. Some theoretical issues and statistical error bound conditions of the maximum deviation between the exact and the approximate magnitude responses are also considered. We use real-valued genetic algorithms (GA) as an optimisation tool and derive results for the maximum error bounds and error deviation due to FWL effects for a number of design examples. Finally, a comparison is drawn between the simply rounded, the GA optimised, integer programming and the simple hill climber methods.

U2 - 10.1007/11539902_108

DO - 10.1007/11539902_108

M3 - Other chapter contribution

SN - 978-3-540-28320-1

T3 - Lecture Notes in Computer Science

SP - 877

EP - 882

BT - Advances in Natural Computation

A2 - Wang, Lipo

A2 - Chen, Ke

A2 - Ong, Yew Soon

PB - Springer

CY - Berlin

T2 - First International Conference on Natural Computation

Y2 - 27 August 2005 through 29 August 2005

ER -

Baicher GS, Taylor M, Rowlands H. A Comparative Study of Finite Word Length Coefficient Optimization of FIR Digital Filters. In Wang L, Chen K, Ong YS, editors, Advances in Natural Computation: First International Conference, ICNC 2005, Changsha, China, August 27-29, 2005, Proceedings, Part III. Berlin: Springer. 2005. p. 877-882. Chapter 108. (Lecture Notes in Computer Science ). doi: 10.1007/11539902_108

A Comparative Study of Finite Word Length Coefficient Optimization of FIR Digital Filters

Abstract

Publication series

Conference

Access to Document

Fingerprint

Cite this