Issue No.06 - December (1995 vol.15)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/40.476266
The fuzzy controller architecture presented in this paper consists of a synchronous digital system which process fuzzy sets at two abstraction levels: fuzzy set level and domain element level. Its main features are the use of trapezia, encoded with four parameters, to represent the antecedent labels. The defuzzification is based in a novel algorithm which speeds up this process by avoiding the operation of division. The performance of the processor is not a constant but depends on the location of the centroid within the domain of the consequent, being better when the centroid is located at the centre of the domain, and worse when it becomes near to one of the ends. Consequently, it takes advantage from the fact that the centroid in an end is the least probable situation. Any case the processor gets significant improvements in speed and latency.
Fuzzy Hardware, Fuzzy Control, Fuzzy Inference, Defuzzification Algorithms, Fuzzy set abstraction level, Domain element abstraction level, Encoded Trapezia
Antonio Ruiz, Julio Gutiérrez, J.A. Felipe Fernández, "A Fuzzy Controller with an Optimized Defuzzification Algorithm", IEEE Micro, vol.15, no. 6, pp. W40-W49, December 1995, doi:10.1109/40.476266