18-Mar-26

Intelligent
Control Systems

Lecture notes
Ha Le Nhu Ngoc Thanh, Ph.D.

Mechatronics Engineering – Intelligent Control 1 © Ha Le Nhu Ngoc Thanh, Ph.D.

CONTROL SYSTEMS
 Classical Control: PID, Root Locus (Undergraduate)
 Modern Control: (Graduate)
- Optimal control (LQR, LQG)
- Adaptive control
- Robust control
- Nonlinear control
+ Advantage: - Mathematically rigorous foundation
- Stability of the system can be guaranteed and proven by math
+ Disadvantage: - System modelling is necessary
- Expert knowledge of automatic control technique, Dynamic model
 Intelligent Control: (Graduate)
- Fuzzy control
- Artificial Neural Network
- Genetic Algorithm
- Particle Swarm Optimization - PSO
+ Advantage: - System modelling is not necessary
+ Disadvantage: - Mathematically rigorous foundation is not guaranteed
- Can not be guaranteed the stability of the system

Mechatronics Engineering – Intelligent Control 2 © Ha Le Nhu Ngoc Thanh, Ph.D.

1
 18-Mar-26

Lecture 1:
Fuzzy Control System
Main contents
1. Fuzzy Set and Membership Function
2. Mathematic operation in Fuzzy Set
3. Fuzzy Logic
4. Fuzzy Rules

Mechatronics Engineering – Intelligent Control 3 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY SETS & MEMBERSHIP FUNCTION

 Level Control: ON/OFF control

1
0
SV

1
0

How to control liquid level at any setpoint value (SV)?

Automatic control  Impossible

Mechatronics Engineering – Intelligent Control 4 © Ha Le Nhu Ngoc Thanh, Ph.D.

2
 18-Mar-26

FUZZY SETS & MEMBERSHIP FUNCTION

 Level Control: Fuzzy control

How large? How small?

Depend of operator
SV (do not follow any criteria)

CV

Mechatronics Engineering – Intelligent Control 5 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY SETS & MEMBERSHIP FUNCTION

1. Crisp Sets and Fuzzy Sets:
Crisp Sets

Crisp set

Meaning of 𝝁𝑨 (𝒙): The membership function, 𝝁𝑨 (𝒙), quantifies how much an element 𝒙 ∈
𝑿 belongs to the crisp set 𝑨. In the crisp set, the value of MF: 𝝁𝑨 𝒙 : 𝑿 → {𝟎, 𝟏}.
In the crisp set, two options are available: 𝑌𝐸𝑆 (100%) or 𝑁𝑂(0%)
• 𝜇𝐴(𝑎) = 1  Element 𝑎 entirely belongs to a set A is 100%
• 𝜇𝐴(𝑏) = 0  Element 𝑏 entirely belongs to a set A is 0% (not belong to A)

Mechatronics Engineering – Intelligent Control 6 © Ha Le Nhu Ngoc Thanh, Ph.D.

3
 18-Mar-26

FUZZY SETS & MEMBERSHIP FUNCTION

1. Crisp Sets and Fuzzy Sets:
Fuzzy Sets

Fuzzy boundary

Fuzzy set

Meaning of the membership function, 𝝁𝑨 𝒙 : quantifies how much an element 𝒙 ∈ 𝑿

belongs to the Fuzzy set 𝑨. In the fuzzy set, the value of MF: 𝝁𝑨 𝒙 : 𝑿 → [𝟎, 𝟏].
• 𝜇 (𝑎) = 1  Element 𝑎 entirely belongs to a fuzzy set 𝐴 to be 100%
• 𝜇 (𝑏) = 0  Element 𝑏 belongs to a fuzzy set 𝐴 to be 0% (not belong to 𝐴)
• 𝜇 𝑐 = ε ∈ [0,1]  Element 𝑐 belongs to a fuzzy set 𝐴 (ε × 100%), and belongs to a
universe set X (100 − (ε × 100%) )

Mechatronics Engineering – Intelligent Control 7 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY SETS & MEMBERSHIP FUNCTION

1. Crisp Sets and Fuzzy Sets:
Computation MF of Fuzzy set 𝜇 𝑥 =?

Mechatronics Engineering – Intelligent Control 8 © Ha Le Nhu Ngoc Thanh, Ph.D.

4
 18-Mar-26

FUZZY SETS & MEMBERSHIP FUNCTION

1. Crisp Sets and Fuzzy Sets:
 Summarization
Fuzzy Sets
Crisp Sets

Fuzzy set
Crisp set

Mechatronics Engineering – Intelligent Control 9 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY SETS & MEMBERSHIP FUNCTION

1. Crisp Sets and Fuzzy Sets:
 Example: which one is fuzzy set, which one is crisp set

Crisp set is clear determination,

without any additional information

Fuzzy set can not determine unless the

membership function 𝜇 (𝑥) is defined

Mechatronics Engineering – Intelligent Control 10 © Ha Le Nhu Ngoc Thanh, Ph.D.

5
 18-Mar-26

FUZZY SETS & MEMBERSHIP FUNCTION

1. Crisp Sets and Fuzzy Sets: How to describe 𝑨
 Notation of Discrete Fuzzy set (biểu diễn của tập mờ rời rạc)

Not additional math Not additional math

Mean: Set Mean: Set

Example:

Mechatronics Engineering – Intelligent Control 11 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY SETS & MEMBERSHIP FUNCTION

1. Crisp Sets and Fuzzy Sets:
 Notation of Continuous Fuzzy set (biểu diễn của tập mờ liên tục):
Example:

Not integral math

Mean: Set

Mechatronics Engineering – Intelligent Control 12 © Ha Le Nhu Ngoc Thanh, Ph.D.

6
 18-Mar-26

FUZZY SETS & MEMBERSHIP FUNCTION

2. Types of membership function and Trapezoid MF:
Trapezoid MF: Types of membership function

Right shoulder Left shoulder Triangular

Fuzzy Singleton
Mechatronics Engineering – Intelligent Control 13 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY SETS & MEMBERSHIP FUNCTION

Membership function in Gaussian types:

Mechatronics Engineering – Intelligent Control 14 © Ha Le Nhu Ngoc Thanh, Ph.D.

7
 18-Mar-26

FUZZY SETS & MEMBERSHIP FUNCTION

3. Characteristic of Membership Function

 Support of fuzzy set 𝐴 (Miền nền của tập mờ 𝐴): is the set all
elements 𝑥 ∈ 𝑋 such that 𝜇 𝑥 > 0
↔ 𝑺𝒖𝒑𝒑𝒐𝒓𝒕 𝑨 = {(𝒙, 𝝁𝑨 𝒙 ) 𝝁𝑨 𝒙 > 𝟎}

 Boundary of fuzzy set 𝐴 (Miền biên của tập mờ 𝐴): is the set all
elements 𝑥 ∈ 𝑋 such that 0 < 𝜇 𝑥 < 1
↔ 𝑩𝒐𝒖𝒏𝒅𝒂𝒓𝒚 𝑨 = { 𝒙, 𝝁𝑨 𝒙 𝟎 < 𝝁𝑨 𝒙 < 𝟏}

 Core of fuzzy set 𝐴 (Miền lõi của tập mờ 𝐴 ): is the set all
elements 𝑥 ∈ 𝑋 such that 𝜇 𝑥 = 1
↔ 𝑪𝒐𝒓𝒆 𝑨 = { 𝒙, 𝝁𝑨 𝒙 𝝁𝑨 𝒙 = 𝟏}

 Height: maximum of upper bound of MF 𝜇 𝑥

Mechatronics Engineering – Intelligent Control 15 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY SETS & MEMBERSHIP FUNCTION

3. Characteristic of Membership Function
 Alpha-cut Fuzzy sets (𝜶 − 𝒄𝒖𝒕) (tập mờ cắt ngưỡng 𝛼):
Given a fuzzy set 𝐴 , with the MF is 𝜇 𝑥 . A set 𝛼 − 𝑐𝑢𝑡 of a fuzzy set 𝐴 is denoted by a fuzzy set 𝐴
0 < 𝛼 < 1 , where MF is 𝝁𝑨𝜶 𝒙 = 𝐦𝐢𝐧{𝜶, 𝝁𝑨 𝒙 }

Example 1: Find 𝜇 . 𝑥 of 𝐴 . of the Fuzzy set 𝐴

0.6

Mechatronics Engineering – Intelligent Control 16 © Ha Le Nhu Ngoc Thanh, Ph.D.

8
 18-Mar-26

FUZZY SETS & MEMBERSHIP FUNCTION

Example 2: Find 𝜇 . 𝑥 𝐴 . of the Fuzzy set 𝐴

Mechatronics Engineering – Intelligent Control 17 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY SETS & MEMBERSHIP FUNCTION

4. Fuzzy partition (Phân hoạch mờ)
Fuzzy sets 𝐴 , 𝐴 , … , 𝐴 is defined in universe set X is called a fuzzy partition if 𝐴 ≠ ∅, 𝐴 ≠ 𝑋 , and

Which one is
Fuzzy partition?

Mechatronics Engineering – Intelligent Control 18 © Ha Le Nhu Ngoc Thanh, Ph.D.

9
 18-Mar-26

FUZZY SETS & MEMBERSHIP FUNCTION

4. Fuzzy partition (Phân hoạch mờ)
Given a MF as below diagram, Fuzzy set 𝐴 , 𝐴 , … , 𝐴 is fuzzy partition? Why?

Mechatronics Engineering – Intelligent Control 19 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY SETS & MEMBERSHIP FUNCTION

5. Linguistic variables and linguistic value (Biến ngôn ngữ và giá trị ngôn ngữ)
 Linguistic variables are variables whose values are words or sentences.
 Linguistic values are words which contain inaccuracy information, thus linguistic value can be
described by Fuzzy Sets
 Example:
Linguistic Linguistic
variable value

Temperature is cold

Velocity is high
Linguistic Linguistic
variable value

Mechatronics Engineering – Intelligent Control 20 © Ha Le Nhu Ngoc Thanh, Ph.D.

10
 18-Mar-26

FUZZY SETS & MEMBERSHIP FUNCTION

5. Linguistic variables and linguistic value (Biến ngôn ngữ và giá trị ngôn ngữ)

(Linguistic value)

How can we conclude in temp =65oF?

Freezing, Cool or Warm or Hot?

(Linguistic variable)

Mechatronics Engineering – Intelligent Control 21 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY SETS & MEMBERSHIP FUNCTION

Example
 Given linguistic variable is “temperature” of a drier system. Let draw a graph of fuzzy sets
describing linguistic values such as “very low”, “low”, “medium”, “high”, and “very high”. Given
physical temperature of the system within 25 < 𝑥 < 200 Degree C

Mechatronics Engineering – Intelligent Control 22 © Ha Le Nhu Ngoc Thanh, Ph.D.

11
 18-Mar-26

MATHEMATICAL OPERATION OF FUZZY SET

* Intersection of Crisp sets
 Intersection of crisp set A and B can be described by membership functions
Example:

How to prove?

Mechatronics Engineering – Intelligent Control 23 © Ha Le Nhu Ngoc Thanh, Ph.D.

MATHEMATICAL OPERATION OF FUZZY SET

1. Intersection of Fuzzy sets
 Intersection of fuzzy sets 𝑨 and 𝑩 , (𝑥 ∈ 𝑋) can be described by membership functions

Mechatronics Engineering – Intelligent Control 24 © Ha Le Nhu Ngoc Thanh, Ph.D.

12
 18-Mar-26

MATHEMATICAL OPERATION OF FUZZY SET

1. Intersection of Fuzzy sets
 Solution:
The general MF can be determine by bellow Figure

Mechatronics Engineering – Intelligent Control 25 © Ha Le Nhu Ngoc Thanh, Ph.D.

MATHEMATICAL OPERATION OF FUZZY SET

1. Intersection of Fuzzy sets
 Example 2 (Code Matlab)

Mechatronics Engineering – Intelligent Control 26 © Ha Le Nhu Ngoc Thanh, Ph.D.

13
 18-Mar-26

MATHEMATICAL OPERATION OF FUZZY SET

2. Union of Crisp sets
 Union of crisp set A and B can be described by membership functions
Example:

Mechatronics Engineering – Intelligent Control 27 © Ha Le Nhu Ngoc Thanh, Ph.D.

MATHEMATICAL OPERATION OF FUZZY SET

2. Union of Fuzzy sets
 Union of fuzzy sets 𝑨 and 𝑩 , (𝑥 ∈ 𝑋) can be described by membership functions

Mechatronics Engineering – Intelligent Control 28 © Ha Le Nhu Ngoc Thanh, Ph.D.

14
 18-Mar-26

MATHEMATICAL OPERATION OF FUZZY SET

2. Union of Fuzzy sets
 Example 2 (code Matlab)

Mechatronics Engineering – Intelligent Control 29 © Ha Le Nhu Ngoc Thanh, Ph.D.

MATHEMATICAL OPERATION OF FUZZY SET

3. Complement (NOT) of Crisp set (Bù, Phủ định)
Complement of a set A includes all elements do not belong to A
Example:

Mechatronics Engineering – Intelligent Control 30 © Ha Le Nhu Ngoc Thanh, Ph.D.

15
 18-Mar-26

MATHEMATICAL OPERATION OF FUZZY SET

3. Complement (NOT) of Fuzzy set
 Complement of fuzzy sets 𝑨 , (𝑥 ∈ 𝑋) can be described by membership functions

Mechatronics Engineering – Intelligent Control 31 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY PROPOSITION LOGIC

1. Clear proposition
 Clear proposition (mệnh đề rõ): is a statement that has only one of two possibilities: TRUE or FALSE
(meaning CLEAR INFOMATION)
 P is clear proposition, then True value of proposition is signed by T(P), (T=TRUE, P=PROPOSITION,
do not mistake T with operator T)
 Example:
 P : room temperature is smaller than 30 degC
If room temperature = 25[oC]  T(P) = 1
If room temperature = 37[oC]  T(P) = 0
 P : Speed of a Car is smaller than 45 kilometers/hour
If Speed = 30km/h  T(P) = 1
If Speed = 50km/h  T(P) = 0

Mechatronics Engineering – Intelligent Control 32 © Ha Le Nhu Ngoc Thanh, Ph.D.

16
 18-Mar-26

FUZZY PROPOSITION LOGIC

2. Fuzzy proposition
 Fuzzy proposition (mệnh đề mờ): is a statement that contain unclear information
 Example:
“Room temperature” is “high”
“Liquid level” of a Tank is “low”
“Speed” of a Car is “fast”
 General Fuzzy proposition is defined by statement as: “Linguistic variable” is “Linguistic value”
𝐱 𝑨
Fuzzy proposition is signed by: 𝑷: 𝐱 ∈ 𝑨
Example:

 Fuzzy set 𝑨 is used to quantify “Linguistic value” in fuzzy proposition

 True value of fuzzy proposition: 𝐓 𝑷 = 𝝁𝑨 𝒙 ∈ [𝟎, 𝟏]
 Note: True value of fuzzy proposition 𝑷: 𝐱 ∈ 𝑨 is dependent possibility of 𝒙 in Fuzzy set 𝑨

Mechatronics Engineering – Intelligent Control 33 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY PROPOSITION LOGIC

3. Mathematic of Fuzzy proposition
 Complement (NOT) of fuzzy proposition

Mechatronics Engineering – Intelligent Control 34 © Ha Le Nhu Ngoc Thanh, Ph.D.

17
 18-Mar-26

FUZZY PROPOSITION LOGIC

3. Mathematic of Fuzzy proposition
 Intersection of fuzzy proposition

Mechatronics Engineering – Intelligent Control 35 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY PROPOSITION LOGIC

3. Mathematic of Fuzzy proposition
 Union of fuzzy proposition

Mechatronics Engineering – Intelligent Control 36 © Ha Le Nhu Ngoc Thanh, Ph.D.

18
 18-Mar-26

FUZZY PROPOSITION LOGIC

3. Mathematic of Fuzzy proposition
 Implication (phép kéo theo) of fuzzy proposition

Mechatronics Engineering – Intelligent Control 37 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY PROPOSITION LOGIC

Exercises

Pressure Temperature

Mechatronics Engineering – Intelligent Control 38 © Ha Le Nhu Ngoc Thanh, Ph.D.

19
 18-Mar-26

FUZZY RULES
Fuzzy Rule (“quy tắc mờ” hoặc “luật mờ”) are statements IF…THEN
In the fuzzy rule:

 Condition proposition and conclusion proposition are fuzzy propositions

Ex: “If liquid is low then pump velocity is high”
 In the condition proposition can contain some logical mathematic such as Union, Intersection or
Complement (NOT)

Mechatronics Engineering – Intelligent Control 39 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY RULES
Fuzzy Rules (hệ luật mờ)

Mechatronics Engineering – Intelligent Control 40 © Ha Le Nhu Ngoc Thanh, Ph.D.

20
 18-Mar-26

FUZZY INFERENCE
1. Crisp Inference (suy luận rõ)
 Crisp Inference is process to infer True value of conclusion proposition from True
value of condition proposition.

Mechatronics Engineering – Intelligent Control 41 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY INFERENCE
2. Fuzzy Inference

Principle of fuzzy inference: Compute True value of condition proposition, subsequently infer
true value of conclude proposition, and this value do not larger the True value of condition
proposition

* Method of Fuzzy inference: two methods MAX – MIN, or MAX-PROD

Mechatronics Engineering – Intelligent Control 42 © Ha Le Nhu Ngoc Thanh, Ph.D.

21
 18-Mar-26

FUZZY INFERENCE
2. Method of Fuzzy Inference
 Method MAX – MIN (MIN: And , MAX: Or )

IF AND THEN

Using principle of fuzzy inference, the conclusion

proposition value do not over the true value of
condition proposition. Thus using set of 𝛽 − 𝑐𝑢𝑡

Mechatronics Engineering – Intelligent Control 43 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY INFERENCE
 Method MAX – PROD

IF AND THEN

Using principle of fuzzy inference, the conclusion

proposition value do not over the true value of
condition proposition. Thus using set of 𝛽 − 𝑐𝑢𝑡

Mechatronics Engineering – Intelligent Control 44 © Ha Le Nhu Ngoc Thanh, Ph.D.

22
 18-Mar-26

FUZZY INFERENCE
Example 1: Given a fuzzy rule:
IF (liquid level is Low) AND (pump’s velocity is Medium) THEN (voltage is High)

LOW MEDIUM HIGH

Find the result of fuzzy inference and draw a membership function by method MAX-MIN,
and MAX-PROD. Given liquid level in the tank 𝑥 = 25%, Pump’s velocity is 𝑥 = 1400 RPM

Mechatronics Engineering – Intelligent Control 45 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY INFERENCE
3. Fuzzy Inference from fuzzy rules
 The fuzzy inference result of fuzzy rules is union of all fuzzy inference results of each individual rule.
 Example: consider two fuzzy rules such as

Method MAX-MIN, or MAX-PROD,

the word “MAX” used only for
fuzzy rule: 𝑟 OR 𝑟

Mechatronics Engineering – Intelligent Control 46 © Ha Le Nhu Ngoc Thanh, Ph.D.

23
 18-Mar-26

FUZZY INFERENCE
 Method 1: MAX – MIN Inference

Rule 1: IF AND THEN

OR
(Aggregation)

Rule 2: IF AND THEN

Result:
Aggregation

Mechatronics Engineering – Intelligent Control 47 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY INFERENCE
 Method 2: MAX – PROD Inference

Rule 1: IF AND THEN

Rule 2: IF AND THEN

Result:

Mechatronics Engineering – Intelligent Control 48 © Ha Le Nhu Ngoc Thanh, Ph.D.

24
 18-Mar-26

FUZZY INFERENCE
4. Fuzzy Inference based on impact fuzzy rule
 In some cases, the set of fuzzy with many rules, computing α , α , … , β , β …,is very complicated
and confused. Thus, we only concern impact rules.
 Impact rules are the rules that the True value of the condition proposition is larger than zero
(𝛽 > 0 ⟺ min{𝛼 , 𝛼 , … 𝛼 } > 0 ⟺ 𝛼 ≠ 0).

Example 3: set of Fuzzy includes 9 rules

Mechatronics Engineering – Intelligent Control 49 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY INFERENCE
Example 4: Given 6 fuzzy rules as follows
R1: If x1 is LO and x2 is NE, then y is VH
R2: If x1 is LO and x2 is ZE, then y is HI
R3: If x1 is LO and x2 is PO, then y is ME
R4: If x1 is HI and x2 is NE, then y is ME
R5: If x1 is HI and x2 is ZE, then y is LO
R6: If x1 is HI and x2 is PO, then y is VL

Question:
Given x1 = 6, x2 = -10. Let draw a result of fuzzy inference in two cases using MAX-MIN and MAX-PROD

Mechatronics Engineering – Intelligent Control 50 © Ha Le Nhu Ngoc Thanh, Ph.D.

25
 18-Mar-26

FUZZY CONTROL SYSTEM

General diagram of Fuzzy control system

Crisp value Crisp value

Fuzzy inference

Preprocessor Postprocessor

Mechatronics Engineering – Intelligent Control 51 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY CONTROL SYSTEM

1. Preprocessor (bộ tiền xử lý)
 Input signal of preprocessor is crisp value from sensor/measurement system
 Preprocessor is used to:
 ADC
 Signal mapping: from sensor/measurement value to criteria signal: [-1, 1] or [0, 1]
 Noise filtering
 …

Mechatronics Engineering – Intelligent Control 52 © Ha Le Nhu Ngoc Thanh, Ph.D.

26
 18-Mar-26

FUZZY CONTROL SYSTEM

2. Fuzzifier (bộ mờ hóa)
 Convert from crisp value to fuzzy value
 Example: given a specific value 𝑥 = 𝑥

Mechatronics Engineering – Intelligent Control 53 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY CONTROL SYSTEM

3. Fuzzy Rule-based (hệ quy tắc mờ cơ bản)
 Fuzzy rule-based can be considered as mathematical models representing human knowledge
and experience in solving problems in the form of linguistic statements.
 Two fuzzy rule-based systems:
1. Mamdani
2. Sugeno
 Mamdani rule is a fuzzy rule where the conclusion statement is a fuzzy proposition.

 Ex:
i. If “error” is “large” and “velocity” is “small”, then “voltage” is “medium”.
ii. If “error angle” is “small negative” and “angular rate” is “small positive”, then “control voltage” is
zero

Mechatronics Engineering – Intelligent Control 54 © Ha Le Nhu Ngoc Thanh, Ph.D.

27
 18-Mar-26

FUZZY CONTROL SYSTEM

 Sugeno rule is a fuzzy rule where the conclusion proposition is a function of input variables.

 Ex:
If “error (e)” is “large” and “velocity (v)” is “small”, then 𝑢 = 5𝑒 + 𝑣 .

Mechatronics Engineering – Intelligent Control 55 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY CONTROL SYSTEM

4. Defuzzifier (Bộ giải mờ)
 "Defuzzification" is the process of converting fuzzy output values of a fuzzy system into crisp values. It
implies that find 𝑦.

 Defuzzification for Mamdani model

a. Maxima method (phương pháp cực đại)
 Height method
 MOM
 LOM
 SOM Height method Smallest of Maximum (SOM)

Largest of Maximum (LOM) Mean of Maximum (MOM)

Mechatronics Engineering – Intelligent Control 56 © Ha Le Nhu Ngoc Thanh, Ph.D.

28
 18-Mar-26

FUZZY CONTROL SYSTEM

 Defuzzification for Mamdani model
b. Centroid method (phương pháp trọng tâm): This method is used for Control problem
 Centre of gravity (COG or centroid)

CoG

2 3 5
1 4 6

Centre of Gravity

Mechatronics Engineering – Intelligent Control 57 © Ha Le Nhu Ngoc Thanh, Ph.D.

FUZZY CONTROL SYSTEM

 Defuzzification for Sugeno model

In fuzzy inference:
• In the Mamdani model, fuzzy rules are combined using the MAX operator before defuzzification.
• In the Sugeno model, fuzzy rules are not combined using the MAX operator before defuzzification;
instead, the output is computed as the weighted average of each rule’s consequent.

Mechatronics Engineering – Intelligent Control 58 © Ha Le Nhu Ngoc Thanh, Ph.D.

29
 18-Mar-26

HCMUTE Univ. - FME

[Link]

Lecture notes info

Intelligent Control
Fall Semester, 2023
------------------------------------------------

Copy right © Ha Le Nhu Ngoc Thanh, Ph.D.

Email: [Link]@[Link]
Sites: [Link]

Mechatronics Engineering – Intelligent Control 59 © Ha Le Nhu Ngoc Thanh, Ph.D.

30