Implementing of Neuro-Fuzzy System with High-Speed, Low-Power CMOS Circuits in Current-Mode

摘要

This paper presents a new general purpose neuro-fuzzy controller to realize adaptive-network-based fuzzy inference system (ANFIS) architecture. ANFIS which tunes the fuzzy inference system with a back propagation algorithm based on collection of input-output data makes fuzzy system to learn. To implementing this idea we propose several improved CMOS analog circuits, including Gaussian-like membership function circuit, minimization circuit, and a defuzziflcation circuit. A two-input/one-output neuro-fuzzy system composed of these circuits is implemented. The control surfaces of controller are obtained by using ANFIS training and simulation results of integrated circuits in less than 0.045mm~2 area in 0.35μm CMOS standard technology. Simulation results show that all the proposed circuits provide characteristics of high operation capacity, high speed, and simple structures. They are very suitable for rapid implementation of high-speed complex neuro-fuzzy system.