APPLICATION OF THE AI TECHNIQUES TO SIGNAL-PROCESSING - A KNOWLEDGE-BASED SYSTEM FOR DIGITAL-FILTERS SYNTHESIS

This paper presents a knowledge-based system for digital filters synthesis, composed of three parts: the synthesis modules, the testing part and the expert system. The first two parts, which are composed of the traditional components of a computer aided design (CAD) system for digital filters synthesis, perform the synthesis, analysis and simulation of finite impulse response (FIR) and infinite impulse response (IIR) filters in both infinite and finite precision. The third part, i.e. the expert system, is developed to aid the user in the choice of the appropriate design methods. To solve this problem of abductive nature a new approach is adopted: the inference engine of our system invokes a particular type of objects called operators, instead of the usual production rules employed in the conventional expert systems.