Information Technology Reference
In-Depth Information
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
−150
−100
−50
0
50
100
150
x
Fig. 2.9 Final antecedents: — EKF(c), - · -EKF ( c + a )
1
0.8
0.6
0.4
0.2
0
−150
−100
−50
0
50
100
150
x
1
0.8
0.6
0.4
0.2
0
−150
−100
−50
0
50
100
150
x
Fig. 2.10 Final antecedents: — EKF(c), -
·
-EKF
(
c
+
a
)
distributed according to the EKF(c) graph of Fig. 2.10 . In order to demonstrate the
operation of
in ( 2.18 ), a restriction on the antecedents is defined, so that each of
them depends on the previous antecedent, and they are never overlap in the regions
of full membership. If the membership are straight lines overlapping by pairs, the
adaptation problem does not have an unique solution which can lead to convergence
problem of the algorithm. This fact and a solution has been discussed in Al-Hadithi
et al. ( 2012 ).
In the EKF
is used both to define the relationship between
the parameters itself of the antecedents and as the consequent. To define the relation-
ship between consequents,
(
c
+
a
)
algorithm,
matrix is always the identity, whereas the relationship
between the parameters of the antecedents is determined by ( 2.37 ). Note that bold
numbers in ( 2.37 ) marks changes from the identity matrix.
Search WWH ::




Custom Search