You are on page 1of 4

9/4/2001

Fuzzy Rules and Fuzzy Reasoning Fuzzy Rules and Fuzzy Reasoning

Fuzzy Inference System


a.k.a.:
- fuzzy rule based system
Fuzzy Inference Systems - fuzzy expert system
(Chapter 4)
- fuzzy model
- fuzzy associative memory
Kai Goebel, Bill Cheetham - fuzzy logic controller
GE Corporate Research & Development - fuzzy system
goebel@cs.rpi.edu
cheetham@cs.rpi.edu

1 2

Fuzzy Rules and Fuzzy Reasoning Fuzzy Rules and Fuzzy Reasoning

The Big Picture Outline


Inference systems consist of three Introduction/Motivation
components: Mamdani Model
rule base TSK Model
data base (defines membership functions) Tsukamoto Model
reasoning mechanism (aggregation) et cetera
defuzzification

3 4

Fuzzy Rules and Fuzzy Reasoning Fuzzy Rules and Fuzzy Reasoning

Motivation Fuzzy Inference System


Fuzzy Reasoning left us with something similar
to:
w1
C x is A1 y is B1
Z
z is C w2
Input x x is A2 y is B2 Aggregator Defuzzify y
...
Real world: wr
x is Ar y is Br
Sensor measurements are crisp (albeit imprecise)
Output to controller has to be crisp

5 6

Page 1

1
9/4/2001

Fuzzy Rules and Fuzzy Reasoning Fuzzy Rules and Fuzzy Reasoning

Design of IS Mamdani Model


Consistency: -Coherent modeling of inference system
no rules with the same antecedents but
different consequents -Appropriate defuzzification strategy

Completeness:
for any x X
there is at least one rule j s.t. A j ,i
0

7 8

Fuzzy Rules and Fuzzy Reasoning Fuzzy Rules and Fuzzy Reasoning

Defuzzification Requirements Defuzzification: COA


Center of Area
Intuition: + intuitive
A crisp value should represent the fuzzy C + smooth
Z
set from an intuitive point of view (e.g., - comp. burden
max. membership grade) zCOA

Computational Burden:
simple (real-time constraints)
Continuity: small changes in fuzzy sets A
( z )zdz
zCOA = z
should not result in large changes of z
z
A
( z )dz
9 10

Fuzzy Rules and Fuzzy Reasoning Fuzzy Rules and Fuzzy Reasoning

Defuzzification: BOA Defuzzification: MOM


Bisector of area: Mean of maximum:

C
C
Z
Z
zBOA
zMOM

zdz
z BOA



A
(z )dz = A
(z )dz
z MOM = Z

dz
z BOA

= min {z z Z }
Z

11
= max{z z Z } 12 { }
Z = z A ( z ) = max A ( z )

Page 2

2
9/4/2001

Fuzzy Rules and Fuzzy Reasoning Fuzzy Rules and Fuzzy Reasoning

SISO: max-min composition


Center Average Defuzzifier centroid defuzzification
Approximation of COA Defuzzifier by average If X is small then Y is small
of center of areas of fuzzy sets s.t. If X is medium then Y is medium
n

z w
j =1
*
j j
If X is large then Y is large
z CA = n

w
j =1
j

where
z*j center of area of jth fuzzy set B
wj height of jth fuzzy set B

13 14

Fuzzy Rules and Fuzzy Reasoning Fuzzy Rules and Fuzzy Reasoning

MISO: max-min composition


centroid defuzzification Sugeno Model (TSK)
If X is small and Y is small then Z is negative large Combines fuzzy sets in antecedents with crisp
function in output:
If X is small and Y is large the Z is negative small
IF x is A AND y is B THEN z=f(x,y)
If X is large and Y is small the Z is positive small
(Does not follow compositional rule of inference)
If X is large and Y is large then Z is positive large
IF x is small THEN
Y=4
IF X is medium THEN
Y=-0.5X+4
IF X is large THEN
Y=X-1
15 16
(sug1_1.m)

Fuzzy Rules and Fuzzy Reasoning Fuzzy Rules and Fuzzy Reasoning

Sugeno: MISO Tsukamoto Model


IF X is small AND Y is small THEN z=-x+y+1 - Consequent of rule is represented by monotonical
IF X is small AND Y is large THEN z=-y+3 MF
IF X is large and Y is small THEN z=-x+3 - crisp output is induced by firing strength
IF X is large and Y is large THEN z=x+y+2 - overall output: weighted average of rule output
- BUT: not as transparent

17 18

Page 3

3
9/4/2001

Fuzzy Rules and Fuzzy Reasoning Fuzzy Rules and Fuzzy Reasoning

Tsukamoto Example Summary


IF X is small THEN Y is C1
IF X is medium THEN Y is C2 Fuzzy Inference:
IF X is large THEN Y is C3 - Mamdami
(a) Antecedent MFs

small medium large


(b) Consequent MFs
Defuzzification derived from area
Me m b e rs hip G ra de s

Me m b e rs hip G ra de s

1 1

0.8 0.8 via appropriate method


0.6 0.6

0.4 0.4
C1 C2 C3
- Sugeno
0.2 0.2

0
-10 -5 0 5 10
0
0 5 10
functional relation of output
X Y

12
(c) Each Rules Output
12
(d) Overall Input-Output Curve
- Tsukamoto
10 10

8 8

6 6
Y

4 4

2 2

0 0
19 -10 -5 0 5 10 -10 -5 0 5 10 20
X X

Fuzzy Rules and Fuzzy Reasoning

last slide

21

Page 4

You might also like