View in EDS

A Zero Extraction and Separation Technique for Surface Acoustic Wave and Digital Signal Processing Fir Filter Implementation

Saved in:
Bibliographic Details
Title: A Zero Extraction and Separation Technique for Surface Acoustic Wave and Digital Signal Processing Fir Filter Implementation
Authors: Lindsay, Keith V.
Committee Members: Malocha, Donald C.; Donald C. Malocha (Q59394643)University of Central Florida. College of Engineering [VIAF]University of Central Florida. College of Engineering [LC]
Summary: Presented is a new method of separating the zeros of a Finite Impulse Response (FIR) filter producing optimal digital filter of surface acoustic wave (SAW) design implementation. Overviews of zero extraction algorithms and of FIR filter design using the Remez Exchange algorithm are presented (McClellen et al. 1973). The computer aided design (CAD) procedure presented allows the designer to specify the general filter characteristic which the Remez algorithm translates to FIR time domain coefficients. These coefficients are readily translated to the frequency (z) domain, producing an Nth order polynomial in z. The characteristic polynomial is factored to determine all roots or zeros using a three-stage factoring program presented by M.A. Jenkins (1975). The roots are optimally separated into two groups, each of which is recombined to form mutually exclusive functions. The two functions are then implemented as transducers of a SAW device or as a two processor digital filter The concept may be extended to more than two subgroups for multi-processor digital filter designs.
URL: https://stars.library.ucf.edu/rtd/4959
Database: OpenDissertations
Description
Abstract:Presented is a new method of separating the zeros of a Finite Impulse Response (FIR) filter producing optimal digital filter of surface acoustic wave (SAW) design implementation. Overviews of zero extraction algorithms and of FIR filter design using the Remez Exchange algorithm are presented (McClellen et al. 1973). The computer aided design (CAD) procedure presented allows the designer to specify the general filter characteristic which the Remez algorithm translates to FIR time domain coefficients. These coefficients are readily translated to the frequency (z) domain, producing an Nth order polynomial in z. The characteristic polynomial is factored to determine all roots or zeros using a three-stage factoring program presented by M.A. Jenkins (1975). The roots are optimally separated into two groups, each of which is recombined to form mutually exclusive functions. The two functions are then implemented as transducers of a SAW device or as a two processor digital filter The concept may be extended to more than two subgroups for multi-processor digital filter designs.