Date of Award

2016

Degree Type

Thesis

Degree Name

M.A.Sc.

Department

Electrical and Computer Engineering

First Advisor

Hon K Kwan

Keywords

custom IC, Differential Evolution, Digital Filter, Linear Phase FIR, multiplierless, optimization

Rights

CC BY-NC-ND 4.0

Abstract

Optimal design of xed coe cient nite word length linear phase FIR digital lters for custom ICs has been the focus of research in the past decade. With the ever increasing demands for high throughput and low power circuits, the need to design lters with reduced hardware complexity has become more crucial. Multiplierless lters provide substantial saving in hardware by using a shift add network to generate the lter coe cients. In this thesis, the multiplierless lter design problem is modeled as combinatorial optimization problem and is solved using a discrete Di erential Evolution algorithm. The Di erential Evolution algorithm's population representation adapted for the nite word length lter design problem is developed and the mutation operator is rede ned for discrete valued parameters. Experiments show that the method is able to design lters up to a length of 300 taps with reduced hardware and shorter design times.

Share

COinS