A superconducting fuzzy processor

Abstract
The authors propose a novel Josephson fuzzy processor which uses Josephson junctions. The two main circuits in the fuzzy processor, an inference engine and a defuzzifier, are described. An algorithm for performing the function of a defuzzifier by mainly using compare circuits instead of multipliers and dividers is discussed. Simulation results for these circuits are illustrated. This fuzzy processor can be operated in a pipelined fashion so that the performance of the processor can be considerably improved. The simulation results show that the Josephson fuzzy processor for four rules and sixteen input levels can perform fuzzy inferences at the rate of 0.79*10/sup 9/ per second.

This publication has 0 references indexed in Scilit: