Objectives

1) Study DC motors
2) Study open-loop and closed-loop control
3) Control methods
I) Proportional feedback control
II) PID (proportional-integral-derivative) con
trol

CEG2400 Ch16: feedback control V3b

1) DC motors
For robots

CEG2400 Ch16: feedback control V3b

Motors
DC motors: Direct cur
rent motor, easy to c
ontrol and use. For m
aking wheeled robots

Servo motors for mak

ing robot legs http://
www.lynxmotion.com
/

CEG2400 Ch16: feedback control V3b

Small Direct Current D.C.

motors
Speed (~ 1200-2000 rpm).
Operates on a 3~5Volt, Ca
n use gear box (e.g. ratio 5
8:1) to increase torque
Use H-bridge circuit to bo
ost up current from the TL
L level to motor driving le
vel.

15351 PRO ,
picture from http://item.taobao.com/item.htm?
id=1606576457&tracelog=newcardfavirate

CEG2400 Ch16: feedback control V3b

Motor control chip

H-bridge Chips
2 (1A)
LEN

1(EN1/2)

LDIR

7(2A)

REN
RDIR

1Y(3)

(2Y)6

Left-motor

10(3A)
(3Y)11
9(EN3/4)
15(4A)
(4Y)14

Right-motor

L293D: H-bridge ci
rcuit, up 2A
LDIR: left motor di
rection; RDIR: righ
t motor direction
LEN : left motor e
nable; REN : right
motor enable

CEG2400 Ch16: feedback control V3b

2) open-loop and closed-loop

control
Feedback control
PID theory and
implementation
CEG2400 Ch16: feedback control V3b

Open-loop motor control and its prob

lems

Change motor supply power change speed

Problem: How much power is right?

Ans: dont know , depends on internal/external frictio

ns of individual motors.

Problem: How to make the robot move straigh

t?
How to control power (Ton) by ISR & an MCU?

Solution: Use feedback control to read actual wheel:

Slower, increase power (+ Ton)
Faster, reduce power (- Ton)
CEG2400 Ch16: feedback control V3b

Exercise
When using the open-loop control metho
d with a constant PWM signal for both wh
eels, explain why the robot would slow do
wn when climbing up hill.
Ans: When climbing up hill, energy is used to
act against gravity, hence the robot is running
slower.

CEG2400 Ch16: feedback control V3b

Feedback control
The real solution to real speed control is f
eedback control
Require speed encoder to read back the r
eal speed of the wheel at real time.

CEG2400 Ch16: feedback control V3b

First you need to have speed encoders

Read wheel speed.

Use photo interrupter
Use reflective disk to save space
Based on interrupts

CEG2400 Ch16: feedback control V3b

10

Wheel encoder

Our motor and

speed encoder
Each wheel rotation=
88 on/off changes
IR receiver

Darkened
part
blocks light

IR light source

CEG2400 Ch16: feedback control V3b

11

CEG2400 Ch16: feedback control V3b

12

New speed encoder

15351 PRO
,
picture from
http://item.taobao.com/item.htm?
id=1606576457&tracelog=newcardfavir
ate

Demo
movie
CEG2400 Ch16: feedback control V3b

13

Student ID: ___________,Date:_____________

Name: _______________CENG2400 , Ch 16 PID

Exercise 1: (Fill in ?__)

The Gear ratio =48, and the DC motor is rotati

66mm
ng at 200 Rotation/Second
So the wheel is rotating at
200/48 ?___ rotations per second
If the disk has 1 hole, the pulse frequency is ?_
_Hz.
If the disk has 4 holes, the pulse frequency is
200x4=?_____Hz.
Wheel radius is r=0.033m, perimeter is 2*pi*0.
033 m
The car is moving at 4 rotations per second=?
__________ meters per second (at full speed, qui
te fast)
The speed encoder gives out ?_______Hz pulses
This is when the disk has 1 hole,
The waveform is 200 Hz
Data. comes from

http://item.taobao.com/item.htm?spm=2013.1.0.116.KbMxbr&id=15172908917
CEG2400 Ch16: feedback control V3b
14

3) Control methods
I) Proportional feedback control
II) PID (proportional-integralderivative) control

CEG2400 Ch16: feedback control V3b

15

I) Proportional closed-loop feed back

control system
Show the left motor control only
if (leftErr >deadband)
leftPWM increase by (Pgain * leftErr)
Required speed
=leftRPMset

leftErr
+
-

Alter PWM
for driver
L293

leftPWM
Motor

leftRPM

CEG2400 Ch16: feedback control V3b

16

II) PID (proportional-integral-derivative) c

ontrol
A more formal and precise method
Used in most modern machines

CEG2400 Ch16: feedback control V3b

17

control method:
PID (proportional-integral-derivative) control

Required speed=

leftRPMset

leftErr

Motor

integral control

leftPWM

Igain* leftErr dt
Proportional control
Pgain*leftErr

sum

generates
PWM
for driver
L293

Derivative control
Dgain*[d(leftErr)/dt]

leftPWM

LeftRPM

leftRPM

CEG2400 Ch16: feedback control V3b

18

IR wheel
Speed
encoder

Introduction

Control for better performance

Use PID, choose whatever response you
want
Motor speed (w)

Too much overshoot/undershoot, not stable

Good performance
Criteria depends
on users and
applications

required

Response too slow

time
CEG2400 Ch16: feedback control V3b

19

Values to evaluate a control system

Exercise 2: Describe the terms n the follo
wing diagrams

Steady
state
error

overshoot
Target
value

Typically
value=10%
Depends
on application

undershoot

Rise time

time

Settling time

CEG2400 Ch16: feedback control V3b

20

Use of PID
control terms are intertwined
http://en.wikipedia.org/wiki/PID_controller

Kp (Pgain): Proportional Gain - Larger Kp typically

means faster response since the larger the error, t
he larger the Proportional term compensation. An
excessively
large proportional gain will lead to proc

ess instability and oscillation.

Ki (Igain): Integral Gain - Larger Ki implies steady s
tate errors are eliminated quicker. The trade-off is l
arger overshoot: any negative error integrated duri
ng transient response must be integrated away by
positive error before we reach steady state.
Kd (Dgain): Derivative Gain - Larger Kd decreases
overshoot, but slows down transient response and
may lead to instability due to signal noise amplifica
tion in the differentiation of the error.
CEG2400 Ch16: feedback control V3b

21

Effects of increasing parameters

http://en.wikipedia.org/wiki/PID_controller
Parameter

Rise Time

Overshoot

Settling
Time

Steady
state error

Kp (Pgain) Decrease
step1

Increase

Small
Change

Decrease

Ki (Igain)

Increase

Increase

Eliminate
step3

Decrease
step2

Decrease

Small
Change

Decrease

Kd (Dgain) Small
Change

CEG2400 Ch16: feedback control V3b

22

Software
PIDrobotdemo.c
(at course webpage)
http://www.cse.cuhk.edu.hk/khwong/ceg2400/PIDRobotDemo.c

CEG2400 Ch16: feedback control V3b

23

IR receiver
Speed Encoder
sensor
interrupts

1000 interrupts per second

PID control algorithm using interrupt

Main( )
{
Setup( );
:
:
}

_IRQ( )//1000Hz
{
:
read wheel speed
PID
:
}

CEG2400 Ch16: feedback control V3b

24

time

Overview

//////////////main //////////////////////////////////////
Main()
{ setup()
forward (Lstep, Rstep, Lspeed, Rspeed)..
}
////////////subroutine //////////////////////////////////////////
forward (Lstep, Rstep, Lspeed, Rspeed)
{ lcount=0
if (lcount>=Lstep)
Stop left motor
same for right motor
.}
//////////timer triggered , interrupt service routine, 1000Hz///////
__irq exception// Interrupt() running at 1000HZ
{ lcount++
read wheel speeds
PID control to achieve L/Rspeed
.same for right motor.
}
CEG2400 Ch16: feedback control V3b

25

Interrupt
rate 1000Hz

_IRQ
see next slide

Speed encoder interfacing

(show left motor only)
Block diagram
ARM7-microcontroller
LPC2131
1000Hz timer
interrupt
/int0
IR
transmitter

In our robot
88 pattern changes
in one rotation

GPIO
Parallel
interface

CPU

IR receiver
(speed
encoder
sensor)
CEG2400 Ch16: feedback control V3b

26

Exercise 3
_irq interrupt programming method for the
main PID loop, using a counter (intcount)

__irq() exception //timer interrupt 1000HZ,

{ intcount++; //intcount

//increases at 1000 times/sec

//read motor speed:

//update the motor speed in every 1/4 seconds

if(intcount==250) // happens at every seconds

{
PID CORE

read wheel speed

Control the PWM using PID

intcount=0;

}
}

Question: If if(intcount==250) is changed to if(intcount==500)

What happens?
CEG2400 Ch16: feedback control V3b

27

The interrupt service routine enables the

loop to run 4 times in a second

Loop once in seconds

CEG2400 Ch16: feedback control V3b

28

Speed control interrupt core _irq()

part 1: Find motor speed, leftRPM
lefttimeval

Speed
encoder
sensor

void __irq IRQ_Exception()

{ms++; intcount++;
//get the current wheel sensor values
Part1
lcur=IO0PIN & LWheelSen;rcur=IO0PIN & RWheelSen;
Read

lcount++;lold=lcur;lefttimeval++;}
time
Wheel if(lcur!=lold) {
:
Interrupts 1000 Hz
speeds
//update motor speed by PID feed back control in every Seconds
if(intcount==250) { //calculate the speed of left motor in RPM

leftRPM=lefttimeval;

:
PID core : See following slides

Part 2
PID
control

intcount=0;
lefttimeval=0;
}

CEG2400 Ch16: feedback control V3b

lcount :steps of wheel

lcur : sensor read 1 or 0
lefttimeval: time lapsed
since sensor last change of
29
state

Part 2 :Algorithm for PID core

For every seconds

Set_point

Find error=(desired value - measured value)

{If (error>dead band )
{ find

In our experiment
leftPWM=276000 at the
beginning and
192800 at steady state

Error,
Accumulated error (add up all previous errors)
Derivative error (current error previous error)
PWM+=Kp* Error+
Ka*(Accumulated error)+
Kd* Derivative error ;

}
Else

Error <= dead_band, error too small do nothing

CEG2400 Ch16: feedback control V3b

30

part 2: PID core

1)
2)
3)
4)
5)
6)
7)
8)
9)
10)
11)
12)
13)
14)
15)
16)
17)
18)
19)
20)
21)

PID core
if(intcount==250) {//for every seconds
//caculate the speed of left motor in RPM
leftRPM=lefttimeval;
P=Proportional
leftErr = leftRPMset - leftRPM; //caculate left error
I Integral
if((leftErr<DeadBand*(-1))||(leftErr>DeadBand)) //see next slide
{ //if |left error| > deadband
leftP = Pgain * leftErr;
//calculate P Proportional term
leftI = Igain * leftaccErr; //calculate I Integral term
leftD = Dgain * (leftErr - leftlastErr);//calculate D Derivative term
leftPWM += (leftP + leftI + leftD);//update left motor PWM using PID
if(leftPWM>PWM_FREQ)
leftPWM=PWM_FREQ;//prevent over range (max.=PWM_FREQ)
if(leftPWM<0) leftPWM = 0; // (min. = 0)
leftaccErr += leftErr; // accumulate error
leftlastErr = leftErr; //update left last error
D Derivative
:
: // handle right motor similarly.
current_leftRPM = leftRPM*240/88;
current_leftPWM = leftPWM;//
Pgain, Igain, Dgain are
lefttimeval = 0;
constants found by a trial and
}

because each rotation has 88 counts, the ISR loop is

in seconds, each minutes 60 seconds. see line 18
CEG2400 Ch16: feedback control

error method, here we have

Pgain = 8000;
Igain = 6000;
Dgain = 31
5000;
V3b

control method:
PID (proportional-integral-derivative) control

Required speed=

leftRPMset

leftErr

Motor

integral control

leftPWM

Igain* leftErr dt
Proportional control
Pgain*leftErr

sum

generates
PWM
for driver
L293

Derivative control
Dgain*[d(leftErr)/dt]

leftPWM

LeftRPM

leftRPM

CEG2400 Ch16: feedback control V3b

32

IR wheel
Speed
encoder

Inside the PID core, we will study these li

nes

5) if((leftErr<DeadBand*(-1))||(leftErr>DeadBand))
:
:
7) leftP = Pgain * leftErr;
//calculate P Proportional term
8) leftI = Igain * leftaccErr;
//calculate I Integral term
9) leftD = Dgain * (leftErr - leftlastErr);//calculate D Derivative ter
m
10) leftPWM += (leftP + leftI + leftD);//update motorPWM by PID
:
:
14) leftaccErr += leftErr;
// accumulate error

CEG2400 Ch16: feedback control V3b

33

Dead band
line5) if((leftErr<DeadBand*(-1))||(leftErr>DeadBand))

Dead-band : A Dead-band (sometimes called a

neutral zone) is an area of a signal range or ba
nd where no action occurs :
leftErr = leftRPM - leftRPMset; //calculate left error
if(leftErr>DeadBand )
{ activate motor}

only enable motor when leftErr> a small value (deadban

d, ie =1 in our robot )
Otherwise may oscillate when leftErr is small
Dead-band

CEG2400 Ch16: feedback control V3b

34

Exercise 4:

Discuss what will happen if dead-band is changed, say (a) 0.5 or (b) 2.

Example of a dead band ;do nothing if 10-1<leftPRM <10

+1

Dead band
overshoot
+/- 1

In our experiment
leftPWM=276000 at the beginning and
192800 at steady state

Steady
state
error

Target speed=
leftRPMset =10
undershoot

Typically
value=10%
Depends
on application

time
Settling time
0
Rise time
When leftRMPset=10, the real RPM is RPM*240/88=27.3., because each
rotation has 88 counts, the ISR loop is in seconds, each minutes 60 seconds.
CEG2400 Ch16: feedback control V3b
35
see line 18

Parameters for evaluating a control syste

m
near steady state
See next slide

Steady
state
error

overshoot
Target
value

Typically
value=10%
Depends
on application

undershoot

Rise time

time

Settling time

CEG2400 Ch16: feedback control V3b

36

(line 7) Effects of increasing Kp(P)

http://en.wikipedia.org/wiki/PID_controller

Parameter

Rise Time

Overshoot

Settling
Time

Steady
state error

(1) Kp
(Pgain)

Decrease
step1

Increase

Small
Change

Decrease

Ki (Igain)

Decrease

Increase

Increase

Eliminate
step3

Decrease
step2

Decrease

Small
Change

Kd (Dgain) Small
Change

CEG2400 Ch16: feedback control V3b

37

Example: t0t1, The proportional term

when the measurement is below the set point (leftRPMse
t), proportional P term is +ve

(leftErr = leftRPMset-leftRPM )
overshoot

Usage of the P proportional term

Each seconds a new
left RPM (speed of wheel) is measured

Target speed=
leftRPMset =10
8
6
leftErr=
leftRPMset
leftPRM
=10-6=4

undershoot

P is
+ve

P is
+ve

t1 t2 t3 t4
Rise time
leftP = Pgain * leftErr;
leftP= 8000* (10-6)= 8000*4 is positive
This term pushing up leftPWM
(more energy delivered to the wheel).

leftErr=
leftRPMset
leftPRM
=30-28=2

e.g.
Pgain = 8000;
Igain = 6000;
Dgain = 5000;
In our experiment
leftPWM=276000 at the
beginning and
192800 at steady state

time
leftP = Pgain * leftErr;
leftP= 8000* (0-8)= 8000*2 is positive
This term pushing up leftPWM
(more energy delivered to the wheel).

CEG2400 Ch16: feedback control V3b

38

Exercise 5: Fill in ?__: t0t1, The proportional term

when the measurement is below the set point (leftRPMse
t), proportional P term is +ve

overshoot

Usage of the P proportional term

Each seconds a new
left RPM (speed of wheel) is measured

Target speed=
leftRPMset =10
8
6
leftErr=
leftRPMset
leftPRM
=10-6=4

undershoot

P is
+ve

P is
+ve

t1 t2 t3 t4
Rise time
leftP = Pgain * leftErr;
leftP= 8000* (10-6)= 8000*4 is positive
This term pushing up leftPWM
(more energy delivered to the wheel).
0

leftErr=
leftRPMset
leftPRM
=10-8=2

e.g.
Pgain = 8000;
Igain = 6000;
Dgain = 5000;
In our experiment
leftPWM=276000 at the
beginning and
192800 at steady state

time
leftP = Pgain * leftErr;
leftP=?___________ ,is +ve or -ve?
leftPWM is increasedor decreased?__
more/less energy delivered to wheel?__

CEG2400 Ch16: feedback control V3b

39

Example: t1t3,The proportional term

when the measurement is below the set point (leftRPMset)
, the proportional P term is +ve

Target speed=
leftRPMset =10
leftErr=
leftRPMset
leftPRM
=10-13= -3

P is
-ve
overshoot

13

Usage of the P proportional term

Each seconds a new
left RPM (speed of wheel) is measured

e.g.
Pgain = 8000;
Igain = 6000;
Dgain = 5000;

undershoot

P is
+ve
0

t1 t2 t3
Rise time

t4

In our experiment
leftPWM=276000 at the
beginning and
192800 at steady state

time

leftP = Pgain * leftErr;

leftP= 8000* (10-13)= 8000*(-3) is negative
This term lowering down leftPWM (less energy delivered to the wheel).
CEG2400 Ch16: feedback control V3b

40

Understanding PID a little summary for

the P proportional term
When the measurement is below the set point (leftRPM
set)
The motor is slow
The P proportional term is +ve, pushing the speed higher

When the measurement is above the set point (leftRPM

set)
The motor is running too fast
The P proportional term is -ve, lowering down the speed.

Increase in Pgain (Kp) will decrease rise time (meaning

faster to reach set point) , decrease steady state error
(study it later)
But increase overshoot

CEG2400 Ch16: feedback control V3b

41

(line 9 ) Effects of increasing Kd (D) dx/dt

http://en.wikipedia.org/wiki/PID_controller
Parameter

dx/dt

Rise Time

Overshoot

Settling
Time

Steady
state error

Kp (Pgain) Decrease
Igain
step1

Increase

Small
Change

Decrease

Ki (Igain)
gainI

Increase

Increase

Eliminate
step3

Decrease

Kd (Dgain) Small
Change

Decrease Decrease
step2

CEG2400 Ch16: feedback control V3b

42

Small
Change

Derivative term
Derivative control
Dgain*[d(leftErr)/dt]

d(leftErr)/dt =
=Derivative term
=current_Err - last_Err
=leftErr leftlastErr ; in our program

CEG2400 Ch16: feedback control V3b

43

Example: time 0 t1, the Derivative term (=current-previous) When the m

easurement is rising, the Derivative term is -ve

Target
Value=

seconds
overshoot

leftRPMset
=10
leftRPMset =10
leftErr=
leftRPMset
7
leftPRM
=10-7=3
leftlastErr= 3
leftRPMset
eftlastPRM
=10-3=7
0

Usage of the D Derivative term

Each seconds a new
left RPM (speed of wheel) is
measured

e.g.
Pgain = 8000;
Igain = 6000;
Dgain = 5000;

undershoot

D is
-ve

In our experiment
leftPWM=276000 at the
beginning and
192800 at steady state

t1 t2 t3 t4
time
Rise time
leftD = Dgain * (leftErr leftlastErr);
leftD= 5000* (3-7)= 5000*( -4) is negative
This term lowering down leftPWM (less energy delivered to the wheel).
CEG2400 Ch16: feedback control V3b

44

Example: time t1t2, the Derivative term

When the measurement is rising, the Derivative term is -ve

leftErr=
leftRPMset -leftPRM
=10-15= -5
leftPRM
15
12
leftRPMset =10

seconds
overshoot

leftlastErr=
leftRPMset -leftlastPRM
=10-12= -2

D is D is
-ve -ve
t1 t2
Rise time

Usage of the D Derivative term

Each seconds a new
left RPM (speed of wheel) is
measured

e.g.
Pgain = 8000;
Igain = 6000;
Dgain = 5000;

undershoot

t3

t4

In our experiment
leftPWM=276000 at the
beginning and
192800 at steady state

time

leftD = Dgain * (leftErr leftlastErr);

leftD= 5000* (-5- (-2))= 5000*(-3) is negative
This term lowering down
CEG2400 Ch16: feedback control
leftPWM (less energy delivered to the wheel).

V3b

45

Little Summary: Negative D Derivative term

When the measurement (leftRPM) is rising, the change

(change= current-previous) of error (error = setpoint-lef
tRPM ) is -ve = d(Err/dt)=-ve
D Derivative term= Kd*d(Err/dt) is VE
Decrease energy to motor decrease overshoot
overshoot
Setpoint

undershoot

D is D is
-ve -ve
t1 t2 t3 t4
time
Rise time
CEG2400 Ch16: feedback control V3b

46

Example: time t2t3, the Derivative term

When the measurement is falling, the Derivative term is +ve

leftRPMset =10

seconds
leftPRM
14
11

leftlastErr=
leftRPMset -leftlastPRM
=10-14= -4

overshoot

Usage of the D Derivative term

Each seconds a new
left RPM (speed of wheel) is
measured

undershoot

D is D is
D is -ve +ve
-ve
t1 t2 t3
0
Rise time

t4

leftD = Dgain * (leftErr leftlastErr);

leftD= 5000* (-1- (-4))= 5000*3 is positive
This term pushing up
CEG2400 Ch16: feedback control
leftPWM (more energy delivered to the wheel).

leftErr=
leftRPMset -leftPRM
=10-11= -1
e.g.
Pgain = 8000;
Igain = 6000;
Dgain = 5000;
time
In our experiment
leftPWM=276000 at the
beginning and
192800 at steady state

V3b

47

Exercise 6:Fill in ?_ time t2t3, the Derivative t

erm
When the measurement is falling,
theofDerivative
termterm
is +ve
Usage
the D Derivative

seconds

leftPRM

overshoot

Each seconds a new

left RPM (speed of wheel) is
measured

leftRPMset =10 9
7
undershoot
D is
leftlastErr=
D is D is +ve
+ve
leftRPMset -leftlastPRM
D is -ve
=10-9= 1
-ve
t1 t2 t3 t4
0
Rise time
leftD = Dgain * (leftErr leftlastErr);
leftD= ?___________, which is +ve or ve?____
leftPWM increased or decreased?____
More/less energy delivered to the wheel?___

leftErr=
leftRPMset -leftPRM
=10-7= 3
e.g.
Pgain = 8000;
Igain = 6000;
time
Dgain = 5000;
In our experiment
leftPWM=276000 at the
beginning and
192800 at steady state
48
CEG2400 Ch16: feedback control V3b

Little Summary: Positive D Derivative term

When the measurement (leftRPM) is falling, the change
(change= current-previous) of error (error = setpoint-lef
tRPM ) is +ve
D Derivative term= Kd*d(Err/dt) is +VE
Increase energy to motor decrease undershoot
overshoot
Setpoint

undershoot

D is D is
+ve +ve
0

t1 t2 t3 t4
time
Rise time
CEG2400 Ch16: feedback control V3b

49

Understanding PID a little summary for

the D derivative term

When the measurement (leftRPM) is rising,

The motor is gaining speed
The D derivative term is ve, so lowering the motor speed decrease
overshoot

When the measurement (leftRPM) is falling,

The motor is reducing speed
The Derivative term is +ve, so pushing the motor speed higher decre
ase undershoot

In conclusion, the gradient of the error (Err) determines t

he adjustment. Depends on whether d(Err)/dt is +ve or v
e.
Increase in Dgain (Kd) will decrease overshoot/undershoot
and settling time (system more stable)
CEG2400 Ch16: feedback control V3b

50

(line 8)Effects of increasing Ki (I) http://en.wiki

pedia.org/wiki/PID_controller
Parameter

Rise Time

Overshoot

Settling
Time

Steady
state error

Kp (Pgain) Decrease
step1

Increase

Small
Change

Decrease

Ki (Igain)

Increase

Increase

Eliminate
step3

Decrease
step2

Decrease

Small
Change

Decrease

Kd (Dgain) Small
Change

CEG2400 Ch16: feedback control V3b

51

control method:
PID (proportional-integral-derivative) control

Required speed=

leftRPMset

leftErr

Motor

integral control

leftPWM

Igain* leftErr dt
Proportional control
Pgain*leftErr

sum

generates
PWM
for driver
L293

Derivative control
Dgain*[d(leftErr)/dt]
At steady state, leftErr0, so
So, Pgain*leftErr=0
leftRPM

leftPWM

LeftRPM

also Dgain*[d(leftErr)/dt] =0
And if no Integral { Igain* leftErr dt} term, left PWM 0
It is a problem.

CEG2400 Ch16: feedback control V3b

52

IR wheel
Speed
encoder

Time near steady state

leftRPMset=leftRPM
hence leftErr =0, leftlastErr=0 leftP=0, leftD=0

In our experiment
leftPWM=276000 at the
beginning and
192800 at steady state

e.g.
leftErr = leftRPMset leftRPM=0
Pgain = 8000;
So as leftlastErr = 0
Igain = 6000;
Therefore
Dgain = 5000;
Steps
7) leftP = Pgain * leftErr//=0
//calculate P Proportional term
8) leftI = Igain * leftaccErr; //calculate I Integral term
9) leftD = Dgain * (leftErr - leftlastErr)//=0
//calculate D Derivative term
10) leftPWM += (leftP + leftI + leftD);//update left motor PWM using PID
near steady state : leftP=0

leftD=0

The only valid term is the integral term leftI

) leftPWM += leftI
) The main idea is to create a small term (left
I) to maintain the leftPWM value
CEG2400 Ch16: feedback control V3b

53

The integral term is found by adding all p

revious errors
same as leftaccErr=sum{leftErr(t=0)+leftErr(t=1)+..

+leftErr(t=now)}

14) leftaccErr += leftErr;//LeftaccErr is the summation of all previous

8) leftI=Igain*leftaccErr// integral term,

errors

:
10) leftPWM += leftI // near steady state, only leftI is valid

Dont worry it will not become infinitive, one measure to safeguar

d this is:

e.g.
Pgain = 8000;
Igain = 6000;
Dgain = 5000;

11) if(leftPWM>PWM_FREQ);// PWM_FREQ=maximum PWM allowed

12) leftPWM=PWM_FREQ;//prevent over range
(max.=PWM_FREQ)

Also, because near steady state , leftaccErr will adjust itself auto
matically, see next slide
In our experiment
leftPWM=276000 at the
beginning and
192800 at steady state
CEG2400 Ch16: feedback control V3b

54

How the integral term adjusts itself auto

matically near steady state
Pgain = 8000;

Igain = 6000;
Dgain = 5000;

Near steady state

8) leftI=Igain*leftacceErr
10) leftPWM += leftI
If measured speed (leftRPM) > set point
leftErr is -ve
leftaccErr (Acculumation) decreases,
hence reducing leftaccErr at a suitab
le value to maintain leftPWM
If measured speed (leftRPM) < set point
leftErr is +ve
leftaccErr (Acculumation) increases,
hence increasing leftaccErr at a suit
able value to maintain leftPWM

CEG2400 Ch16: feedback control V3b

In our experiment
leftPWM=276000 at the
beginning and
192800 at steady state

55

Understanding PID a little summary for

the I Integral term

When the measurement is below the set point (leftRPMset)

The motor is slow
The I Intergral term is +ve, pushing the speed higher

When the measurement is above the set point (leftRPMset)

The motor is running too fast
The I intergral term is -ve, lowering down the speed.

Increase in Igain (Ki) will result in

It is similar to the P term (see above two points)

So increase overshoot, settling time
and decrease rise time
Note: the main function of Integral control is to reduce steady state err
or

CEG2400 Ch16: feedback control V3b

56

PID Tuning
(usually done by
trail and error)
Tune (adjust manually)

Motor speed

Accepted
performance

V1
(V1)/2

step1) Pgain proportional_gain (Kp),

step2) Dgain derivative_gain (Kd),
step3) Igain integral_gain (Ki)

T1
unstable

Set constant speed V1 for a while (5 seconds) and reduc

ed to (V1)/2 at T1
Record the speed by the computer after T1 and see if the pe
rformance is ok or not

Yes (accept Kp,Ki,Kd)

No (tune Kp,Ki,Kd again)

done
CEG2400 Ch16: feedback control V3b

57

time

Application of
PID

CEG2400 Ch16: feedback control V3b

58

Robot Turning

(for smooth circular turns around 90-degree corn

ers)
http://micromouse.cannock.ac.uk/dynamics/smoothcorners.htm

Required speed V (meters per minute)

Turning radius R (meters), from a point to robot center
Wheel diameter = a (meters)
Use similar triangle
V1 V V2
V1/(R+D/2)=V/R
V1=V+(VD)/2R
V2/(R-D/2)=V/R
a
V2=V-(VD)/2R
D/2 D/2
R
wi=rotations per
minute of wheel i
D
Result
w1=V1/(2a/2)=(2RV+VD)/(2Ra)
w2=V2/(2a/2)=(2RV-VD)/(2Ra)

smooth circular turn

CEG2400 Ch16: feedback control V3b

59

Summary
Studies PID control theory and implement
ation

CEG2400 Ch16: feedback control V3b

60

Appendix:
New robot drive circuit ver13.3

CEG2400 Ch16: feedback control

V3b

61