VLSI hardware realization of self-learning recursive fuzzy algorithm

This paper presents the hardware realization of a self-learning fuzzy logic algorithm based upon recursive estimation and a probabilistic approach to fuzzy identification which, together with an inherent filtering feature, makes the system highly immune to noise. In the past, the majority of hardware-based fuzzy inference processors were designed as application-specific or custom integrated circuits. However, semi-custom programmable integrated circuits, used in this paper, offer totally reconfigurable nonproprietary hardware designs as well as many other advantages related to flexibility, execution speed, and development cost.