Issue No. 05 - October (1995 vol. 15)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/40.464591
We describe a digital architecture to evaluate fuzzy inferences. This architecture is based on a novel systolic structure in the field of fuzzy processing that is called multilevel systolic approach. Its main features are: high throughput, perfomance independent of the fuzzy model size supported, high flexibility to resize the parameters of the design, max-min inference, and capability for a considerable amount of complex rules without loss of efficiency. The circuit development has been carried out by using a VHDL simulator, with ES2 1um standard cells, and its results has given a perfomace over 10 mega fuzzy logic rule base inferences per second.
fuzzy, fuzzy hardware, fuzzy inference, systolics
L. de Salvador and J. Gutiérrez, "A Multilevel Systolic Approach for Fuzzy Inference Hardware," in IEEE Micro, vol. 15, no. , pp. 61-71, 1995.