This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.
Permalink
https://hdl.handle.net/2142/71987
Description
Title
High-Performance Digital Filters
Author(s)
Young, Christopher
Issue Date
1992
Doctoral Committee Chair(s)
Jones, Douglas L.
Department of Study
Electrical Engineering
Discipline
Electrical Engineering
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
Ph.D.
Degree Level
Dissertation
Date of Ingest
2014-12-16T22:23:02Z
Keyword(s)
Engineering, Electronics and Electrical
Abstract
Digital filters, both finite impulse response (FIR) and infinite impulse response (IIR), are widely used in many applications. The motivation for this work lies in the importance of developing methods to reduce the complexity in implementing these types of filters.
The first issue addressed in this thesis is the reduction in computation for a FIR filter. We propose a method called shifted partial products to reduce the amount of computation for a FIR filter in the transpose form. A parallel implementation of this type of FIR filter is designed in very large scale integration (VLSI). Comparisons of this implementation to other conventional methods of implementation are made, primarily with respect to chip area. The VLSI implementation is strictly for an application-specific integrated circuit (ASIC) design, and the results obtained using shifted partial products compare favorably to those using a standard modified Booth's implementation.
The second issue addressed is the design of finite-wordlength linear-phase FIR filters. For a fixed-point implementation, it is known that truncating and/or rounding the infinite precision values does not, in general, result in optimal performance. We propose the log-cascade method for selecting the coefficients for a cascade structure. This method provides performance better than the current best method in many cases. By cascading subfilters to implement an overall filter and using a linearizing approximation to obtain an integer program, it is possible to improve quantized coefficient filter designs. Results are most impressive for filter designs requiring deep stopband suppression, a very common and important application.
We also introduce an extension of Kodek's method, the iterative Kodek method, which yields results slightly superior to the log-cascade method for most designs. However, this iterative method does contain limitations which are explained and discussed.
The last issue addressed in this thesis is the design of both constant and variable wordlength IIR filters. The log-cascade method is modified for the constant wordlength problem, but results are unsatisfactory. Subsequent investigation then focuses on the design of IIR filter coefficients with variable wordlengths. In this problem, the number of bits required to represent each coefficient in an IIR filter is varied, but the total number of bits is constant. Two different methods are presented for selecting these filter coefficients, each resulting in dramatic gains in performance. Both methods are motivated by the desire for a linear error trajectory and exploit the differing sensitivities of the filter coefficients.
Use this login method if you
don't
have an
@illinois.edu
email address.
(Oops, I do have one)
IDEALS migrated to a new platform on June 23, 2022. If you created
your account prior to this date, you will have to reset your password
using the forgot-password link below.
