Date of Award

1991

Degree Type

Thesis

Degree Name

M.A.Sc.

Department

Electrical and Computer Engineering

First Advisor

Ahmadi, M.

Keywords

Engineering, Electronics and Electrical.

Rights

CC BY-NC-ND 4.0

Abstract

In the process of designing a digital filter, the coefficients of the transfer function are normally evaluated with a high degree of accuracy. To implement the designed filter in hardware, finite word length registers have to be utilized. This may bring about truncation of the filter coefficients to the limited word length registers employed, which in turn may affect the total response of the digital filter. To remedy this, filters are designed with the finite word length coefficients. In this thesis a review of existing techniques for the design of integer coefficients 1-D FIR filters is presented. Where possible extension of these techniques to 2-D are also presented. In this thesis two design methods for 1-D and 2-D FIR filter with integer coefficients based on the suboptimal approach of discretization and reoptimization techniques are presented. These two approaches are somewhat modified versions of the well known branch and bound optimization technique. This thesis also presents a technique for the design of 1-D and 2-D FIR filters with integer coefficients using Mixed integer linear programming. Delta modulation is also employed for a new filter structure for 1-D FIR filter with its coefficients being $-$1, 0, and +1. McClallan transformation is used for designing a class of 2-D FIR filters with integer coefficients and a suitable realization structure. In this thesis, we also present a comparison of various techniques we have proposed in terms of complexity of the design issue, as well as the computational burden for each iteration. Source: Masters Abstracts International, Volume: 31-03, page: 1332. Co-Supervisors: M. Ahmadi; M. Shridhar. Thesis (M.A.Sc.)--University of Windsor (Canada), 1991.

Share

COinS