Number-theoretical analysis of the structures of classical IIR digital filters

To implement IIR digital filters with specified requirements for frequency characteristics, a large number of structures can be used. If we do not take into account the finiteness of the word length, then all structures are equivalent in terms of the accuracy of the realization of the characteristic...

Full description

Saved in:
Bibliographic Details
Published in2018 7th Mediterranean Conference on Embedded Computing (MECO) pp. 1 - 4
Main Authors Lesnikov, Vladislav, Naumovich, Tatiana, Chastikov, Alexander
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.06.2018
Subjects
algebraic numbers
Digital filters
Digital signal processing
Embedded computing
Finite element analysis
IIR digital filters
IIR filters
poles
Poles and zeros
zeros
Online AccessGet full text

Cover

More Information
Summary:To implement IIR digital filters with specified requirements for frequency characteristics, a large number of structures can be used. If we do not take into account the finiteness of the word length, then all structures are equivalent in terms of the accuracy of the realization of the characteristics. However, in the practical implementation of digital Alters, the bitness is finite, and the coefficients of the filter structure, and hence the coefficients of the transfer function, are rational numbers. As a consequence, the zeros and poles of the transfer function are elements of a set of algebraic numbers. For different structures, the maximum degree of algebraic numbers that are zeros and fields themselves can be different. Therefore, the maximal degrees of zeros and poles can be used as the basis for classifying structures and dividing them into equivalence classes. The proposed method is demonstrated on classical structures.
DOI:10.1109/MECO.2018.8406099