CMOS design of analog neuro-fuzzy system with improved circuits

This paper intends outlining several improved analog integrated circuits designed in CMOS, standard VLSI technologies, which could be used in fuzzy inference system as the general modules. Current-mode CMOS implementation of min circuit and a programmable Gaussian-like membership function circuit are proposed, A closed-loop control scheme for the system is employed to avoid the division of a centroid calculation in the normalization block.