Professional Documents
Culture Documents
Fuzzy System
2.1
(2.4-a)
(2.4-b)
(2.5-a)
(2.5-b)
2.6
2.7
2.8
2.9
2.10
2.11
2.12
2.13
2.14
• Inputs:
The error between the desired signal and the plant output
State variables: Position, velocity, temperature
• Outputs:
Voltage, current, power, ….
The input to the plant
• Fuzzify inputs:
Determine ranges for inputs (universal sets)
Choose membership functions (types, number of
membership functions)
• Aggregation method:
MAX, SUM
• Defuzzification method:
Mamdani:
Sugeno:
• Simulate fuzzy logic system
Matlab – Fuzzy
WinFACT – FLOP
http://www.kahlert.com/web/english/e_fuzzyneuro.php
fuzzyTECH - http://www.fuzzytech.com/
• Hardware:
Computer
Microprocessor, Micro controller, DSP
Texas Instruments (DSP) - fuzzyTECH MCU-320
Hard disk control, noise cancellation, sonar signal
classification, servo motor control
Microchip Technologies (PIC) - fuzzyTECH MCU-MP
DC motor control, furnace, air conditioner control, car door
control, fake card detection
Intel
Anti-lock braking system, CNC position control, Metal
heating furnace, household device, hand-writing recognization.
Allen-Bradley / Rockwell (PLC) - fuzzyTECH IA-PLC5
Siemens (Fuzzy166)
Motorola
National Instruments
Mitsubishi
Wonderware (Process control)
• System test
• Adjust and complete system
r e u y
Plant
Controller
(or object)
r: setpoint Controller:
y: plant output - PID
u: control signal - Fuzzy
e = r-y: the error - Neural networks
- u = f(e)
Fuzzy system
(ANFIS)
• Example:
• x = (0:0.1:10)';
• y = sin(2*x).*exp(x/5);
• epoch_n = 20;
• in_fis = genfis1([x y],5,'gbellmf');
• out_fis = anfis([x y],in_fis,epoch_n);
• plot(x,y,x,evalfis(x,out_fis));
• legend('Training Data','anfis Output');
1 1
e u
10/20/2020 Dr. Nam H. Nguyen 30
2.3.3 Basic functions
-b e
b
-a
1 1
e u
r e u y
P fuzzy
G(s)
controller
de
de/dt
r e u y
Fuzzy logic
Object
system
ie
PD fuzzy controller
de: derivative of e(t)
PI fuzzy controller
ie: integral of e(t)
PID fuzzy controller
10/20/2020 Dr. Nam H. Nguyen 39
2.4.2 PD fuzzy controller
de
de/dt
r e u y
Fuzzy logic
Object
system
e
0
Rule Table
Negative Zero Positive
Negative Negative Negative Small
e negative
Zero Zero Zero Zero
Positive Small Positive Positive
positive