Information Technology Reference
In-Depth Information
Fig. 13.4
xfsg
GUI. When all the fuzzy system components have been configured (identified by
green marks close to them) it is possible to generate the output files by pressing the button 'Generate
Files'. Configurations can be saved (button 'Save Configuration') to be easily reused in a posterior
moment
been generated, which includes different blocks to implement each of the stages of
the architecture for fuzzy inference described in the previous section. Figure
13.3
shows the components included in
XfuzzyLib
grouped by functionalities. MFCs
based on arithmetic techniques are able to generate families of membership functions
corresponding to the types “triangular” and “sh_triangular” defined in
Xfuzzy
(all the
members of a sh_triangular family are functions with triangular shapes, except the
first and the last ones, which can be Z- and S-shaped functions, respectively). The
library provides two options (minimum and product) to implement the antecedent
connectives used in fuzzy rules, as well as different defuzzifiers (Fuzzy Mean,
Weighted Fuzzy Mean, first-order Takagi-Sugeno, and MaxLabel) for interpolators
and decision-making systems. Finally, the library also contains a set of crisp blocks
that implement general purpose arithmetic and logic operations. The dimension and
functionality of
XfuzzyLib
components is defined by means of a set of parameters that
can be assigned using numerical values or bymeans of
Matlab
variables. Numerical
values of these variables can be later defined using the
Matlab
command window
or an “.m” file.
Building a fuzzy inference system with
XfuzzyLib
requires choosing, intercon-
necting, and defining the parameters of the needed blocks. To ease the design
task,
XfuzzyLib
also includes elements describing archetypal fuzzy logic controllers
(FLCs) that differ in the number of inputs, the connective used to calculate rule
activation degrees, and the defuzzification method. These blocks are also parame-
Search WWH ::
Custom Search