An exhaustive search algorithm for checking limit cycle behavior of digital filters

Kamal Premaratne, Ernest C. Kulasekere, Peter H. Bauer

Research output: Contribution to journalArticle

17 Citations (Scopus)

Abstract

The presence of limit cycles that may arise in fixedpoint arithmetic implementation of a digital filter can significantly impair its performance. This paper presents an algorithm to determine the presence/absence of such limit cycles. For generality the filter is taken to be in its state-space formulation. The algorithm is applicable independent of filter order type of quantization nonlincarity and whether the accumulator is single or double length. It may be utilized to construct limit cycle free regions in filter coefficient space. Once a filter is determined to be limit cycle free a technique that provides a robustness region in coefficient space where all filters remain limit cycle free is also presented.

Original languageEnglish
Pages (from-to)2405-2412
Number of pages8
JournalIEEE Transactions on Signal Processing
Volume44
Issue number10
DOIs
StatePublished - Dec 1 1996

Fingerprint

Digital filters

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Signal Processing

Cite this

An exhaustive search algorithm for checking limit cycle behavior of digital filters. / Premaratne, Kamal; Kulasekere, Ernest C.; Bauer, Peter H.

In: IEEE Transactions on Signal Processing, Vol. 44, No. 10, 01.12.1996, p. 2405-2412.

Research output: Contribution to journalArticle

@article{1de21e96e4434328958bfe549e3470af,
title = "An exhaustive search algorithm for checking limit cycle behavior of digital filters",
abstract = "The presence of limit cycles that may arise in fixedpoint arithmetic implementation of a digital filter can significantly impair its performance. This paper presents an algorithm to determine the presence/absence of such limit cycles. For generality the filter is taken to be in its state-space formulation. The algorithm is applicable independent of filter order type of quantization nonlincarity and whether the accumulator is single or double length. It may be utilized to construct limit cycle free regions in filter coefficient space. Once a filter is determined to be limit cycle free a technique that provides a robustness region in coefficient space where all filters remain limit cycle free is also presented.",
author = "Kamal Premaratne and Kulasekere, {Ernest C.} and Bauer, {Peter H.}",
year = "1996",
month = "12",
day = "1",
doi = "10.1109/78.539026",
language = "English",
volume = "44",
pages = "2405--2412",
journal = "IEEE Transactions on Signal Processing",
issn = "1053-587X",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "10",

}

TY - JOUR

T1 - An exhaustive search algorithm for checking limit cycle behavior of digital filters

AU - Premaratne, Kamal

AU - Kulasekere, Ernest C.

AU - Bauer, Peter H.

PY - 1996/12/1

Y1 - 1996/12/1

N2 - The presence of limit cycles that may arise in fixedpoint arithmetic implementation of a digital filter can significantly impair its performance. This paper presents an algorithm to determine the presence/absence of such limit cycles. For generality the filter is taken to be in its state-space formulation. The algorithm is applicable independent of filter order type of quantization nonlincarity and whether the accumulator is single or double length. It may be utilized to construct limit cycle free regions in filter coefficient space. Once a filter is determined to be limit cycle free a technique that provides a robustness region in coefficient space where all filters remain limit cycle free is also presented.

AB - The presence of limit cycles that may arise in fixedpoint arithmetic implementation of a digital filter can significantly impair its performance. This paper presents an algorithm to determine the presence/absence of such limit cycles. For generality the filter is taken to be in its state-space formulation. The algorithm is applicable independent of filter order type of quantization nonlincarity and whether the accumulator is single or double length. It may be utilized to construct limit cycle free regions in filter coefficient space. Once a filter is determined to be limit cycle free a technique that provides a robustness region in coefficient space where all filters remain limit cycle free is also presented.

UR - http://www.scopus.com/inward/record.url?scp=0030270692&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0030270692&partnerID=8YFLogxK

U2 - 10.1109/78.539026

DO - 10.1109/78.539026

M3 - Article

AN - SCOPUS:0030270692

VL - 44

SP - 2405

EP - 2412

JO - IEEE Transactions on Signal Processing

JF - IEEE Transactions on Signal Processing

SN - 1053-587X

IS - 10

ER -