Digital Signal Processing Reference
In-Depth Information
(a)
[
µ
ik
(
x
i
)]
x
1
[
y
i
]
input1 MFs
Set
of
if-then
rules
Sugeno
output
calcula-
tion
x
2
z
input2 MFs
x
3
input3 MFs
input
MFs
output
MFs
weighted sum
output
(b)
inputs
rules
normalization
factor
Fig. 15.1
FIS studied: a general structure, b adaptive network representation
becomes fuzzified when
• Any T-norm or S-norm is introduced to combine the input data with weighting
factors,
• Aggregation is performed with any T-norm or S-norm,
• u is any continuous function.
The following are just few examples of fuzzy neurons that can be defined in
such a way [
2
,
10
]:
Search WWH ::
Custom Search