Information Technology Reference
In-Depth Information
set of responses as shown in Figure 1. Once the TLR receives the signal (ie.
the temperature reading), it evaluates it according to predefined condition and
response pairs shown in Table 1. The TLR also returns the inflammation level
associated with the particular response:
y
=
f
(
TLR
x
)
(2)
where
x
denotes the TLR in question and
y
denotes the inflammation level asso-
ciated with the action that should be performed when the robot is in that state.
Such functions can be implemented in terms of simple mathematical functions,
lookup tables, fuzzy logic operators or any other appropriate technique. An ex-
ample input could be the value 50, which represents the temperature of one of
the motors and a response is generated according to the following lookup table
(for example):
Table 1.
Function table
Condition
Response
T
x
<
40
o
C
∅
40
o
C<T
x
<
80
o
C
Fan On
T
x
>
80
o
C
Fan On, Motor Off
This means, that the outcome of the function
f
will be the action
Fan On
.This
is a local immediate response to the trigger of a single TLR. If the temperature
is within the acceptable range, no action will be taken.
Response
1
Response
2
... Response
n
TLR
1
1.0
1.0
...
f
(
TLR
1
)
TLR
2
0.0
0.0
...
f
(
TLR
2
)
...
...
...
...
...
TLR
x
...
...
...
f
(
TLR
x
)
x
m
=1
f
(
TLR
m
)
...
...
...
...
Fig. 2.
Input Feature Vector
In this model implementation, the stable state inflammation level is repre-
sented with the real value 0
.
0, while the TLR triggered state is 1
.
0. This is the
contribution to the inflammation level described above. Once the model collects
the outputs of the TLR functions of each individual TLR, a vector is created
from the responses as shown in Figure 2. This vector is used as input to the
SOM and the sum of its components is used to update the inflammation level
according to equation 1.
