CMOS Neuro-Fuzzy Controller Based on ANFIS

An analog CL,IOS netro-fuzzy controller with fiyo inputs and one output, ttsing integrated circttits, is presented. Each input is divided into three fuzzy membership functions. One advantage of this controller is the ability to use membershipftmctionsa ncl their complementss imultaneouslyw hich provides mttch more freedom and precision in implementing ,,if-then,' rules. On the other hand, both AND and OR operators can be used in combination of rules accordiig ro the application. Llrhole chip design is based on ANFIS (Adaptive Neuro-Fuzzy Inference system) , developed by Jyh-shing Roger Jang in 1993 and has benefit of learning ability due to meural networks theory. CMOS 0.35m m 130 mm . Input voltaged vary from 0 to 3.3v. while output current rangest from 0 up to 1mA (this current can be adjusted by user according to the application

Keywords:

Neuro-Fuzzy , Controller , ANFIS , Analog CMOS , Current mode