• Demo to show the differences between open
loop and closed loop control including P, PI
and PID
Master Project – Month Year 1
International University Biomedical Engineering
VNU-HCMC Department
Implementing a digital PID
Controller on Arduino board
Prepared: Nguyễn Vũ Duy Hậu
Author Scientific Advisor(s)
[email protected] Contents
• Overview
• Implementation
• Conclusion
• Anatomy of Arduino,
• How does it work?
• Functions of Arduino
• System using Arduino (w or w/o laptop)
• How to use Arduino in the feedback control?
Master Project – Month Year 4
• Anatomy of Arduino,
Master Project – Month Year 5
Overview
Comparision
VS
About 75,000 VND Consume more time and money
Which one do you prefer for prototyping
and testing?
September 2016 6
• How does it work?
• Functions of Arduino
Master Project – Month Year 7
• System using Arduino (w or w/o laptop)
Master Project – Month Year 8
• How to use Arduino in the feedback control?
Master Project – Month Year 9
Master Project – Month Year 10
Overview
Our PID demo system in detail
Counterclockwise
Motor - Actuator
Clockwise
Arduino - Controller
Sensor - Angle
September 2016 11
Implementation
Control Actuator by PWM (Pulse Width Modulation)
Duty Cycle RPM Connect control signal to
100% or Always On ~ 2000
digital pin (2) of Arduino
50% ~ 1000
• digitalWrite(2, LOW);
10% ~200
• digitalWrite(2, HIGH);
September 2016
Implementation
Control Sensor
Connect sensor signal to analog pin (A7) of
Arduino
• analogRead(A7);
Arduino automatically convert analog signal
from sensor to digital signal range from 0 -
1023
Sensor voltage Digital value (VCC=5V)
5V 1023
2.5 512
0V 0
September 2016
Implementation
PID controller
September 2016
Implementation
One cycle of PID controller
• Arduino get the desired angle
1
• Arduino get the recent angle from sensor
2
• Arduino compute the PID control signal
3 and send to motor
• Arduino wait for the next cycle
4
The first
1st 2nd 3rd 4th …
cycle
Time domain Period
September 2016
Implementation
The important of period
Natural PID system cycle:
1. Get the desired color in brain
2. Check the recent color
3. Adjust the position
• Decide how quick the system
response
• Must be selected approriately
to ensure Tasks 1,2,3 must be
done
5 hours/cycle
5 seconds/cycle1st 5 miliseconds/cycle
2nd 3rd 4th …
Time domain Period
September 2016
Implementation
Proportional - P term
Consider the current
Decide how strong
error at the time of the
the system response
controller calculation
Max control signal: 360 degree
Min control signal: 0 degree
September 2016
Implementation
Proportional - P term
How to get 100km/h with P controller?
Cycle Setpoint Process Value Kp Control Signal
1 100 km/h 0 km/h 3 300 degree
2 100 km/h 200 km/h 3 - 300 degree
or 0 degree
3 100 km/h 50 km/h 3 150 degree
4 100 km/h 150 km/h 3 -150 degree or
0 degree
Cycle Setpoint Process Value Kp Control Signal
1 100 km/h 0 km/h 0.5 50 degree
↑ Kp: ↑ Oscillation
2 100 km/h 50 km/h 0.5 25 degree
↓ Kp: ↓ Oscilation
3 100 km/h 110 km/h 0.5 -5 degree or 0
but can not reach
degree the setpoint → need
4 100 km/h 90 km/h 0.5 5 degree Integral term
September 2016
Implementation
Proportional - P term
In each cycle process of Arduino:
error = Setpoint - ProcessVariable;
Pterm = Kp*error;
September 2016
Implementation
Integral - I term
Decide how strong
the system response
Consider the history of the error over time
P controller cannot reach setpoint
With I term sum up all small value
overtime → increase signal
control to reach the setpoint
September 2016
Implementation
Integral - I term
Cycle Setpoint Process Value Kp Control Signal
1 100 km/h 0 km/h 0.5 50 degree I term can add
2 100 km/h 50 km/h 0.5 25 degree more to control
3 100 km/h 110 km/h 0.5 -5 degree or 0
signal → > 5
degree degree
4 100 km/h 90 km/h 0.5 5 degree
This P controller cannot reach
setpoint
With I term sum up all small value
overtime → increase signal
control to reach the setpoint
September 2016
Implementation
Integral - I term
Approximate by Discrete Integral
At the begin: Iterm = 0; ↓ Ki: ↓ Oscilation but
too low or cannot
In each cycle process of Arduino: reach the setpoint
Iterm = Iterm + ↑ Ki: reach setpoint
but ↑ Oscillation →
(Ki*error*period); need Derivative term
September 2016
Implementation
Differential - D term
Decide how strong D
term against P, I terms
Try to predict process value change
then against/slow this change or
the control efforts of P, I terms
At the begin: LastError = 0;
In each cycle process of Arduino:
Dterm = Kd*(error-LastError)/period;
September 2016
Implementation
Differential - D term
At the begin: Iterm =
LastError = 0;
In each cycle process of Arduino:
• Pterm = Kp*error;
• Iterm = Iterm +
(Ki*error*period);
• Dterm = Kd*(error-
LastError)/period;
Time Error Last Error Pterm Iterm Dterm
20 8 0 Kp*8 90 Kd*8/period
30 1 8 Kp*1 130 Kd*(-7)/period
40 -2 1 Kp*(-2) 120 Kd*(-3)/period
September 2016
Implementation
Simple PID controller
September 2016
Conclusion
Summary characteristics of P,I,D
P term = 0; I term largest D term largest
P term I term D term
Look at the error Look at the history of Try to predict the
currently error overtime value error in the
future
Try to drive system to Try to drive system to Find the way to
reach the setpoint reach the setpoint against P and I efforts
P term The more process value Get very strong Largest effect when
closer to the setpoin, overtime, can help P the process is
largest; the more P term get term drive system to changing rapidly in
I term smaller then cannot reach setpoint and one direction and
=0 drive system to reach event make system disapper when system
the setpoint oscilation if Ki too stable at setpoint
large
September 2016
Conclusion
Summary characteristics of P,I,D
Closed-Loop Rise Time Overshoot Settling Time Steady-State error Stability
Response
Increasing Kp Decrease Increase Small Increase Decrease Degrade
Increasing Ki Small Decrease Increase Increase Large Decrease Degrade
Increasing Kd Small Decrease Decrease Decrease Minor Change Improve
September 2016
Master Project – Month Year 28
