A data converter for analog/fuzzy‐logic interface

Abstract

Input signals are often fed as analog quantities into fuzzy logic processing systems. This paper proposes a method for configuring interface necessitated by fuzzy logic systems featuring analog inputs and digital processing. This interface is a data converter which processes input analog voltage signals and outputs digital signals suitable for fuzzy logic processing. The conversion principle is based on the concept of level sets within the fuzzyset theory, i.e., depending on whether a given analog signal belongs to level sets, the signal is quantized and converted to membership grade, simultaneously. Having programmable conversion characteristics, it can realize typical membership functions such as Z‐, ‐, and S‐functions. The operation of the converter, which is configured of counters, a memory, a digital‐to‐analog converter, and a comparator, is verified experimentally as regards a non‐monotonic and asymmetric membership function. the conversion accuracy is then discussed and the causes for error are described.

Fuzzy inference based on families of alpha -level sets

“…5 [13], [14]. In this figure, {P,} denotes the family of a-level sets which defines the fuzzy set P. The processing time with this algorithm is proportional to logam, where m is the number of quantized membership grades.…”