Silicon Implementation for a Novel High-Speed Fuzzy Inference Engine: Mega-Flips Analog Fuzzy Processor

Article type: Research Article

Authors: Miki, Tsutomu | Matsumoto, Hidetoshi | Ohto, Keishi | Yamakawa, Takeshi

Affiliations: Department of Computer Science, and Control Engineering, Kyushu Institute of Technology, Iizuka 820, Japan | Electronic Components, Division H.Q., IC Business Promotion Center, OMRON Corporation, Kyoto 670, Japan

Abstract: This article describes two types of analog fuzzy processors. One is a rule chip (FP-9000) and the other a defuzzifier chip (FP-9001), and both can achieve high-speed fuzzy inference. The inference speed is more than 1 Mega fuzzy logical inferences per second, excluding defuzzification. The rule chip includes four fuzzy inference engines. Each engine achieves one fuzzy inference, which is characterized by a fuzzy if-then rule, in analog mode. These rules can be directly read out and written in through an external digital computer. The rule chip is fabricated in 2 μm BiCMOS technology. The defuzzifier chip converts a fuzzy value to a crisp one, necessary for a fuzzy control system. The defuzzifier chip is implemented by 3 μm bipolar technology. A high-speed fuzzy controller hardware system can be efficiently constructed with these chips. The chips accelerate to develop fuzzy logic systems, especially high-speed applications.

DOI: 10.3233/IFS-1993-1104

Journal: Journal of Intelligent and Fuzzy Systems, vol. 1, no. 1, pp. 27-42, 1993