IEICE Trans - Design of CSD Coefficient FIR Filters Using PSO with Penalty Function

Design of CSD Coefficient FIR Filters Using PSO with Penalty Function

Kazuki SAITO  Kenji SUYAMA

Publication
IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences   Vol.E98-A    No.12    pp.2625-2632
Publication Date: 2015/12/01
Online ISSN: 1745-1337
DOI: 10.1587/transfun.E98.A.2625
Type of Manuscript: PAPER
Category: Digital Signal Processing
Keyword:
digital filter,  particle swarm optimization,

Full Text: PDF>>

Summary:
In this paper, we propose a method for designing finite impulse response (FIR) filters with canonic signed digit (CSD) coefficients using particle swarm optimization (PSO). In such a design problem, a large number of local minimums appear in an evaluation function for the optimization. An updating procedure of PSO tends to stagnate around such local minimums and thus indicates a premature convergence property. Therefore, a new framework for avoiding such a situation is proposed, in which the evaluation function is modified around the stagnation point. Several design examples are shown to present the effectiveness of the proposed method.