A First Course in FUZZY and NEURAL CONTROL

8.4. IDENTIFICATION OF TRASH IN COTTON 277

Initial MF's on Edif

0

0.2

0.4

0.6

0.8

1

1.2

0 0.1 0.2 0.3 0.4 0.5

Edif

Final MF's on Edif

0

0.2

0.4

0.6

0.8

1

1.2

0 0.10.20.30.40.5

Edif

Figure 8.20. Initial andfinal membership functions onEdif: 222-partition

The linguistic variablesoliditywas partitioned assmallandlargeinstead of
small,medium,andlarge(membership partition-232). There are a total of 8
fuzzy rules for identifying the trash types; they are summarized in Table 8.12.
Tables 8.13 and 8.14 are the updated values of the consequent and premise
parameters after 200 epochs of training with the 222-membership partition.

Table 8.12. TSK fuzzy rules for trash

identification with 222-partition^2

If and and then

Rule areais solidityis Edifis trash typeis

R^1 Small Small Small Pepper

R^2 Small Small Large Pepper

R^3 Small Large Small Pepper

R^4 Small Large Large Pepper

R^5 Large Small Small Bark

R^6 Large Small Large Bark

R^7 Large Large Small Leaf

R^8 Large Large Large Stick

Table 8.13. Final values of consequent

parameters: 222-partition

Consequent Parametersaji

ji 0123

1 2. 2184 − 0. 0016 − 0 .6010 3. 5199

2 − 0 .0841 0.0069 1.1953 10. 1662

3 − 0 .0897 0.0010 0. 7641 − 0. 5085

4 − 1 .2554 0.0004 2. 1392 − 1. 8541

5 0 .1818 0. 0000 − 0. 1468 − 0. 1665

6 3 .0904 0. 0000 − 7 .5783 4. 1122

7 0. 8847 − 0. 0001 − 0. 1512 − 0. 1263

8 1 .9480 0. 0000 − 0. 7704 − 1. 0734

(^2) The membership partition of the input variablearea,solidity,andEdifare {(Small,
Large), (Small, Large), (Small, Large)} and is represented as 222-partition in the text.