VLSI implementation of an inference processor for approximate reasoning

Abstract

This report outlines the design and implementation of a prototype fuzzy inference processor (FIP) using Xilinx field programmable gate arrays (FPGAs). The fuzzy rules are represented as pre-computed relation matrices and the fuzzy input variables are considered to be in the form of fuzzy singletons. This simplifies the inferencing process significantly and results in hardware architecture that is less complex.