AN INVESTIGATION INTO USING EXPERT SYSTEMS FOR DIGITAL FILTER DESIGN

Abstract

It can be challenging to select the best architecture for DSP filters for a given application. Design constraints often include both objective and subjective information. This paper discusses the initial results of an investigation into using expert system techniques to address this problem. The goal is a system that allows users to specify traditional constraints such as impulse response, frequency response, stability, SNR, etc., but they may also constrain the filter’s cost, complexity, or any parameter which can be clearly identified for the specific application.