Structure of Fuzzy Control Module with Neural Network

The paper deals with the fuzzy control module's structure, which has the property that is deficit in "conventional" uncertain systems -the skill to learn. This is attained by offering the control module in the form of a neural-like multilayer network. At the same time, this system is free from the main disadvantage of neural networks -the knowledge distribution. All parameters and weights retain their physical understanding, which makes it likely to examine the information gathered by the system in the process of learning.

Перевод пока недоступен