Intelligent Tuning of Fuzzy Controllers by Learning and Optimization - Fuzzy Modeling and Control: Theory and Applications - page 132

Fig. 7.3 Main design parameters of a fuzzy controller
Indeed, the incorporation of human know-how is not straightforward, and fuzzy
rules can be obtained from numerical data gathered during manual process opera-
tion performed by expert operators. This method provides the objectivity of actual
measurements, after processing large data files containing the information gathered
during manual operation. Several computational procedures can be applied such as
fuzzy clustering and neurofuzzy strategies as will be futher analyzed in this chapter.
The use of fuzzy or neural network models to generate a set of fuzzy control rules
and the use of learning to create the set of fuzzy rules have been widely investigated.
The research field on automatic learning, which entails the ability to create fuzzy
control rules and to modify them based on experience or self-organizing algorithms
have yielded outstanding results (Shaw 1998 ).
In next sections, a brief description of fuzzy techniques simple design methods
called “learning by example” is presented. Two main approaches are considered:
fuzzy clustering and neuro-fuzzy strategies. The first one will be focused on design-
ing fuzzy controllers without significant information about the type of membership
functions, rule base, etc. The second technique will be used when there is an initial
fuzzy controller and the objective is to achieve a better adjustment of its parameters.
Two case studies will be presented to demonstrate the suitability of both strategies.
7.3.1 Fuzzy Clustering
If information about the template linguistic values is not available and only input-
output data is available, the structure of the fuzzy controller (relationship between
variables, rough estimates of the membership functions of the antecedent and con-
sequent fuzzy sets, and the number of rues) can be obtained by clustering the