We consider a fuzzy controller that processes fuzzy data in which error, change in error, etc. are all fuzzy sets. We construct, and train, a new type of neural net to model the fuzzy controller. Properties of this new type of neural net include: (1) max-min operations; and (2) a modified delta rule for learning. An example is presented showing the applicability of our fuzzy max-min neural controller
Published in:
Neural Networks, 1995. Proceedings., IEEE International Conference on
(Volume:6
)
Date of Conference: Nov/Dec 1995