Indonesian
Journal
of
Electrical
Engineering
and
Computer
Science
V
ol.
7,
No
.
2,
A
ugust
2017,
pp
.
434
441
DOI:
10.11591/ijeecs
.v7.i2.pp434-441
434
Composite
Nonlinear
Feedbac
k
With
Disturbance
Obser
ver
f
or
Active
Fr
ont
Steering
Sarah
’Atifah
Saruc
hi
1
,
Hairi
Zamzuri
2*
,
Noraishikin
Zulkarnain
3
,
Mohd
Hatta
Mohammed
Ariff
4
,
and
Norbaiti
W
ahid
5
1,2,4,5
Mala
ysia-J
apan
Inter
national
Institute
of
T
echnology
,
Univ
ersiti
T
eknologi
Mala
ysia,J
alan
Sult
an
Y
ah
y
a
P
etr
a,
54100
K
uala
Lumpur
,
Mala
ysia
3
F
aculty
of
Engineer
ing
and
Built
En
viroment,
Univ
ersiti
K
ebangsaan
Mala
ysia,
43600
Bangi
S
elangor
,
Mala
ysia
*
corresponding
author
,
e-mail:
hair
i.kl@utm.m
y
Abstract
One
of
the
dominant
vir
tue
of
St
eer-By-Wire
(SBW)
v
ehicle
is
its
capability
to
enhance
handling
perf
or
mance
b
y
installing
Activ
e
F
ront
Steer
ing
(AFS)
system
without
the
dr
iv
er’
s
interf
erences
.
Hence
,
this
paper
introduced
an
AFS
control
str
ategy
using
the
combination
of
Composite
Nonlinear
F
eedbac
k
(CNF)
controller
and
Disturbance
Obser
v
er
(DOB)
to
achie
v
e
f
ast
y
a
w
r
ate
tr
ac
king
response
which
is
also
rob
ust
to
the
e
xistence
of
disturbance
.
The
proposed
control
str
ategy
is
sim
ulated
in
J-cur
v
e
and
Lane
change
manoe
vres
with
the
presence
of
side
wind
disturbance
via
Matlab/S
im
ulink
sotw
are
.
Futher
more
,
compar
ison
with
Propor
tional
Integ
r
al
Der
iv
ativ
e
(PID)
and
Linear
Quadr
atic
Regulator
(LQR)
controllers
are
also
conducted
to
e
v
aluate
the
eff
ectiv
eness
of
the
proposed
controller
.
The
results
sho
w
ed
that
the
combined
CNF
and
DOB
str
ategy
achie
v
ed
the
f
astest
y
a
w
r
ate
tr
ac
king
capabilit
y
with
the
least
impact
of
disturbance
in
the
AFS
system
installed
in
SBW
v
ehicle
.
K
e
yw
or
ds:
composite
nonlinear
f
eedbac
k,
disturbance
obser
v
er
,
activ
e
front
steer
ing,
st
eer-b
y-wire
Cop
yright
c
2017
Institute
of
Ad
v
anced
Engineering
and
Science
.
All
rights
reser
v
ed.
1.
Intr
oduction
Steer-By-Wire
(SBW)
is
a
steer
ing
system
which
remo
v
es
mechanical
linkages
betw
een
steer
ing
wheel
and
front
wheel
systems
with
electronic
components
and
wires
.
Due
to
the
re-
mo
v
al,
an
A
ctiv
e
F
ront
Steer
ing
(AFS)
system
can
be
applied
independently
without
the
dr
iv
er’
s
interf
erence
because
there
is
no
fix
ed
relationship
betw
een
the
steer
ing
wheel
angle
and
the
front
wheels
angle
[1,
2].
The
aim
of
AFS
implementation
is
to
enhance
v
ehicle
manouv
er
ability
and
stability
[3,
4].
Here
,
AFS
is
installed
in
SBW
v
ehicle
f
or
y
a
w
r
ate
tr
ac
king
control
to
assist
the
dr
iv
ers
to
k
eep
the
v
ehicle
on
the
desired
path.
CNF
controller
is
a
combination
of
linear
and
nonlinear
f
eedbac
k
la
ws
without
an
y
s
witch-
ing
element.
According
to
the
pre
vious
research
findings
,
CNF
has
been
pro
v
en
to
ha
v
e
strong
capabilites
in
achie
ving
f
ast
y
a
w
r
ate
tr
ac
king
perf
or
mance
with
minimal
o
v
ershoot
in
AFS
system
[5,
6].
Ho
w
e
v
er
,
Hasan
et.
al
addressed
the
CNF
perf
or
mance
issue
when
a
side
wind
distur-
bance
is
applied
to
the
AFS
system
[7].
According
to
the
results
,
the
y
a
w
r
ate
tr
ac
king
response
is
major
ly
aff
ected
b
y
the
disturbance
.
It
means
that
the
CNF
cannot
stand
alone
to
o
v
ercome
the
e
xistence
of
disturbance
.
This
issue
leads
to
the
decreasing
of
the
v
ehicle
handling
perf
or
mance
.
Regarding
the
issue
,
an
integ
r
at
ion
of
CNF
controller
with
obser
v
er
such
as
reduced-
order
and
e
xtended
state
obser
v
ers
is
one
of
the
solution
to
enhance
the
rob
ustness
of
the
control
system
[8,
9,
10].
Thus
,
this
paper
proposes
the
combinat
ion
of
CNF
and
Disturbance
Obser
v
er
(DOB)
in
AFS
system
to
impro
v
e
the
tr
ansient
perf
or
mance
of
y
a
w
r
ate
tr
ac
king
control
and
at
the
same
time
sim
ultaneously
reject
the
element
of
disturbance
.
DOB
is
chosen
because
it
is
widely
used
in
SBW
system
due
to
its
efficiency
and
design
simplicity
[11,
12].
This
paper
is
organiz
ed
as
f
ollo
ws
.
In
the
ne
xt
section,
the
mathematical
model
of
SBW’
s
front
wheel
system
and
v
ehicle
model
are
presented.
Then,
the
f
ollo
wing
section
descr
ibes
the
Receiv
ed
March
26,
2017;
Re
vised
J
uly
12,
2017;
Accepted
J
uly
23,
2017
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4752
435
design
of
the
control
str
ucture
.
The
f
our
th
section
sho
ws
the
sim
ulation
results
and
discussion.
Finally
,
the
conclusion
and
future
w
or
ks
are
being
concluded
and
suggested
in
the
last
section.
2.
System
Modeling
2.1.
SBW’
s
Fr
ont
Wheel
System
(a)
(b)
Figure
1.
SBW’
s
front
wheel
system
a)
Mechanism,
b)
Bloc
k
diag
r
am
T
ab
le
1.
F
ront
wheel
system
par
ameter
v
alue
P
ar
ameter
Definition
V
alue
Unit
P
ar
ameter
Definition
V
alue
Unit
sw
Steer
ing
angle
-
r
ad
B
m
Motor
damping
coefficient
0.007
N
ms=r
ad
f
w
F
ront
wheel
angle
-
r
ad
J
f
F
ront
wheel
iner
tia
1.36
k
g
m
2
m
Motor
angle
-
r
ad
R
Motor
electr
ical
resistance
3.1124
O
hm
V
V
oltage
-
v
L
Motor
electr
ical
inductance
1.6663
H
enr
y
i
Motor
current
-
A
y
r
Rac
k
displacement
-
m
m
r
Rac
k
Mass
2
k
g
B
r
Rac
k
damping
coefficient
25
N
ms=r
ad
J
m
Motor
iner
tia
0.0012
k
g
m
2
B
k
Kingpin
damping
coefficient
70
N
ms=r
ad
K
s
Lumped
torque
stiffness
0.257
N
m=r
ad
r
p
Pinion
gear
r
adius
0.1
m
K
b
Motor
emf
constant
0.2319
V
s=r
ad
r
l
Offset
of
king
pin
axis
0.69
m
K
l
Steer
ing
linkage
stiffness
26e2
N
m=r
ad
Figure
1
sho
ws
the
mechanism
and
b
loc
k
diag
r
am
of
front
wheel
system
while
T
ab
le
1
lists
the
par
ameter’
s
details
[13].
Basically
,
this
system
consiste
d
of
steer
ing,
motor
,
r
ac
k,
pinion
and
front
wheels
,
as
sho
wn
in
Figure
1(a).
The
equations
f
or
the
system
can
be
e
xpressed
as
,
Motor
displacement
and
current:
•
m
=
1
J
m
(
B
m
_
m
+
K
s
i
)
;
_
i
=
1
L
(
R
i
+
K
b
_
m
+
V
)
(1)
Rac
k
and
pinion
:
•
y
r
=
1
m
r
2
K
l
r
l
2
y
r
K
s
r
p
2
y
r
B
r
_
y
r
K
s
r
p
m
(2)
F
ront
wheel
displacement:
•
f
w
=
1
J
f
K
l
f
w
y
r
r
l
B
k
_
f
w
(3)
Based
on
Figure
1(b),
these
equations
can
be
represented
in
the
f
ollo
wing
tr
ansf
er
function
f
or
ms
,
Composite
Nonlinear
F
eedbac
k
With
Disturbance
Obser
v
er
f
or
...
(Sar
ah
’Atif
ah
Sar
uchi)
Evaluation Warning : The document was created with Spire.PDF for Python.
436
ISSN:
2502-4752
G
1
(
s
)
=
m
(
s
)
sw
(
s
)
=
(
2
:
274
10
13
)
s
2
+
(2
:
365
10
11
)
s
s
3
+
506
:
2
s
2
+
11360
s
+
34390
s
3
+
506
:
2
s
2
+
11360
s
(4)
G
2
(
s
)
=
y
r
(
s
)
m
(
s
)
=
(
3
:
553
10
15
)
s
+
(1
:
75
10
4
)
s
2
+
12
:
5
s
+
(1
:
805
10
5
)
(5)
G
3
(
s
)
=
f
w
(
s
)
y
r
(
s
)
=
(1
:
421
10
14
)
s
+
1015
s
2
+
51
:
47
s
+
1471
(6)
2.2.
V
ehic
le
Model
Figure
2.
Linear
v
ehicle
model
T
ab
le
2.
V
ehicle
model
par
ameter
P
ar
ameter
Definition
V
alue
Unit
P
ar
ameter
Definition
V
alue
Unit
l
F
Distance
of
F
ront
Wheel
to
COG
1.14
m
F
!
Cross
wind
F
orce
2000
N
m
l
R
Distance
of
Rear
Wheel
to
COG
1.64
m
F
y
F
F
ront
wheel
later
al
f
orce
-
N
m
f
w
F
ront
wheel
angle
-
r
ad
Y
a
w
r
ate
-
r
ad
l
!
Exter
nal
F
orce
P
oint
0.5
m
F
y
R
Rear
wheel
later
al
f
orce
-
N
m
V
ehicle
body
slip
angle
-
r
ad
m
V
ehicle
Mass
1529.98
k
g
C
f
F
ront
Wheel
Cor
ner
ing
Stiffness
54500
N
=r
ad
I
z
Y
a
w
Iner
tia
4000
k
g
m
2
C
R
Rear
Wheel
Cor
ner
ing
Stiffness
42600
N
=r
ad
v
V
elocity
22.22
ms
2
Figure
2
sho
ws
the
2-DOF
linear
v
ehicle
model
while
T
ab
le
2
lists
the
par
ameter’
s
details
.
The
v
ehicle
model
is
used
as
the
v
ehicle
plant
to
in
v
estigate
the
handling
perf
or
mance
.
It
consid-
ers
2-DOF
of
the
v
eh
icle
body
in
later
al
and
y
a
w
motion.
Assuming
the
v
elocity
to
be
constant,
the
linear
v
ehicle
model
can
be
presented
in
the
f
ollo
wing
state
space
f
or
m
[14,
15,
16]:
_
x
b
=
A
b
x
b
+
B
b
u
b
+
E
b
F
!
(7)
where
,
A
b
=
2
6
6
4
2(
C
F
+
C
R
)
mv
1
2(
l
R
C
R
l
F
C
F
)
mv
2
2(
l
R
C
R
l
F
C
F
)
I
z
2(
l
2
F
C
F
+
l
2
R
C
R
)
I
z
v
3
7
7
5
B
b
=
2
6
6
4
2
C
F
mv
2
l
F
C
F
I
z
3
7
7
5
,
E
b
=
2
6
6
4
1
mv
l
!
I
z
3
7
7
5
x
b
=
2
4
3
5
,
u
b
=
f
w
.
IJEECS
V
ol.
7,
No
.
2,
A
ugust
2017
:
434
441
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4752
437
3.
Contr
oller
Design
Figure
3.
Proposed
control
str
ucture
T
ab
le
3.
P
ar
ameter
of
the
control
str
ucture
P
ar
ameter
Definition
P
ar
ameter
Definition
c
Correctiv
e
steer
ing
angle
G
4
(
s
)
T
r
ansf
er
function
of
nominal
y
a
w
Actual
y
a
w
r
ate
C
2
(
s
)
Y
a
w
r
ate
tr
ac
king
controller
r
ef
Desired
y
a
w
r
ate
C
1
(
s
)
Wheel
synchronization
controller
e
Y
a
w
r
ate
error
D
Disturbance
compensator
control
input
R
2
(
s
)
Ref
erence
model
G
1
4
(
s
)
In
v
erse
tr
ansf
er
function
of
nominal
y
a
w
!
Side
wind
disturbance
^
!
Estimated
side
wind
disturbance
Figure
3
sho
ws
the
o
v
er
all
vie
w
of
the
proposed
control
str
ucture
,
which
includes
wheel
synchronisation
and
AFS
control
systems
.
Meanwhile
,
T
ab
le
3
sets
out
the
par
ameters
used
in
the
proposed
control
str
ucture
.
F
ront
wheel
system
needs
a
controller
to
control
the
motor
position
in
order
to
steer
the
front
wheels
.
Theref
ore
,
CNF
is
also
implemented
as
the
wheel
synchronisation
controller
to
ensure
that
the
front
wheel
angle
is
ab
le
to
synchronise
with
the
steer
ing
wheel
angle
commanded
b
y
the
dr
iv
er
,
as
discussed
in
detail
in
[17].
3.1.
Composite
Nonlinear
Feedbac
k
According
to
Figure
3,
in
y
a
w
r
ate
tr
ac
king
control
str
ategy
,
an
additional
correcte
d
angle
c
obtained
fro
m
the
CNF
controller
C
2
(
s
)
is
added
to
the
steer
ing
input
sw
to
k
eep
the
actual
y
a
w
r
ate
response
closely
tr
ac
k
the
desired
y
a
w
r
ate
response
r
ef
gener
ated
b
y
ref
erence
model
R
2(
s
)
.
The
CNF
controller
is
designed
based
on
SBW’
s
front
wheel
and
linear
v
ehicle
model
systems
.
Consequently
,
the
equations
of
the
systems
is
represented
in
t
he
f
ollo
wing
state
space
f
or
m
where
the
input
u
=
sw
and
output
y
=
.
_
x
=
Ax
+
B
u
y
=
C
x
(8)
Composite
Nonlinear
F
eedbac
k
With
Disturbance
Obser
v
er
f
or
...
(Sar
ah
’Atif
ah
Sar
uchi)
Evaluation Warning : The document was created with Spire.PDF for Python.
438
ISSN:
2502-4752
where
,
A
=
2
6
6
6
6
6
6
6
6
6
6
6
6
6
6
4
B
m
J
m
0
K
s
J
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
K
b
L
0
R
L
0
0
0
0
0
0
0
K
s
m
r
r
p
0
B
r
m
r
K
l
m
r
r
2
l
+
K
s
m
r
r
2
p
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
K
l
J
f
r
l
B
k
J
f
K
l
J
f
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
2
C
F
mv
2(
C
F
+
C
R
)
mv
1
2(
l
R
C
R
l
F
C
F
)
mv
2
0
0
0
0
0
0
2
l
F
C
F
I
z
2(
l
R
C
R
l
F
C
F
)
I
z
2(
l
2
F
C
F
+
l
2
R
C
R
)
I
z
v
3
7
7
7
7
7
7
7
7
7
7
7
7
7
7
5
B
T
=
0
0
1
L
0
0
0
0
0
0
;
x
T
=
_
m
m
i
_
y
r
y
r
_
f
w
f
w
CNF
consists
of
linear
u
L
and
nonlinear
u
N
f
eedbac
k
la
ws
.
The
o
v
er
all
of
the
la
ws
u
o
is
e
xpressed
as
[10],
u
o
=
u
L
+
u
N
(9)
In
linear
f
eedbac
k
par
t,
the
damping
r
atio
is
set
to
be
small
so
that
the
system
could
achie
v
e
the
ref
erence
input
closely
and
ha
v
e
a
f
ast
r
ising
time
.
While
in
the
nonlinear
f
eedbac
k
par
t,
the
damping
r
atio
w
as
increased
to
reduce
the
o
v
ershoot
resulted
from
the
linear
par
t.
The
details
of
linear
f
eedbac
k
la
w
is
as
f
ollo
ws:
u
L
=
F
x
+
Gr
(10)
Here
,
r
is
the
step
input
command
or
ref
erence
.
F
is
the
controller
tuning
par
ameter
to
gener
ate
responses
with
f
ast
r
ising
time
.
While
G
is
a
scalar
which
is
giv
en
as:
G
=
h
C
(
A
+
B
F
)
1
B
i
1
(11)
On
the
other
hand,
the
nonlinear
f
eedbac
k
la
w
can
be
descr
ibed
as
,
u
N
=
(
r
;
y
)
B
0
P
(
x
x
e
)
(12)
Here
,
P
is
the
solution
of
the
f
ollo
wing
L
ypuno
v
equation,
(
A
+
B
F
)
0
P
+
P
(
A
+
B
F
)
=
W
(13)
where
W
is
a
positiv
e
definite
matr
ix.
In
this
study
,
W
is
simplified
as
I
.
The
equation
f
or
x
e
is
descr
ibed
as
,
x
e
=
(
A
+
B
F
)
1
B
Gr
(14)
Based
on
Equation
12
,
(
r
;
y
)
is
a
nonpositiv
e
function
locally
Lipschitz
in
x
which
can
be
e
x-
pressed
as
,
(
r
;
y
)
=
'e
0
j
h
r
j
(15)
where
,
>
0
and
'
>
0
are
the
tuning
par
ameters
to
minimise
o
v
ershoot.
IJEECS
V
ol.
7,
No
.
2,
A
ugust
2017
:
434
441
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4752
439
3.2.
Disturbance
Obser
ver
Based
on
Figure
3,
the
function
of
DOB
is
to
eliminate
the
side
wind
disturbance
!
eff
ect
without
aff
ecting
the
input
sw
from
the
steer
ing
wheel
b
y
pro
viding
the
estimated
disturbance
^
!
,
which
is
then
used
to
perf
or
m
the
compensation
through
negativ
e
f
eedbac
k
loop
.
Firstly
,
the
y
a
w
model
der
iv
ed
from
linear
v
ehicle
model
is
defined
as
the
nominal
model
G
4
(
s
)
.
F
rom
Equation
7,
the
tr
ansf
er
function
of
nominal
y
a
w
model
G
4
(
s
)
is
giv
en
as
f
ollo
ws
where
the
input
is
front
wheel
angle
,
f
w
and
the
output
is
y
a
w
r
ate
,
.
G
4
(
s
)
=
(
s
)
f
w
(
s
)
=
(2
C
F
l
F
mv
)
s
+
4
C
F
C
R
(
l
F
v
+
l
R
)
mI
z
v
s
2
+
(2(
C
F
+
C
R
)
I
z
+
2(
C
F
l
2
F
+
C
R
l
2
R
)
v
)
s
+
a
0
(16)
where
,
a
0
=
(4
C
F
C
R
(
l
F
+
l
R
)
2
+
2(
C
R
l
R
C
F
l
F
))
mv
2
In
DOB
design
procedure
,
f
w
is
calculated
based
on
G
2
(
s
)
and
G
3
(
s
)
.
F
or
an
easier
procedure’
s
e
xplaination,
firstly
let
G
2
(
s
)
and
G
3
(
s
)
assumed
as
1.
Based
on
Figure
3
pre
viously
,
the
linear
v
ehicle
model
can
be
e
xpressed
as
,
=
G
4
(
s
)(
f
w
+
!
)
(17)
where
!
is
the
side
wind
e
xter
nal
disturbance
.
The
estimated
disturbance
^
!
is
giv
en
b
y
,
^
!
=
G
4
(
s
))
1
(
G
4
(
s
)(
f
w
+
!
))
f
w
(18)
The
selection
of
Lo
w
pass
filter
L
p
(
s
)
deter
mined
the
disturbance
rejection
perf
or
mance
at
lo
w
frequencies
in
the
closed-loop
system.
The
L
p
(
s
)
order
is
chosen
to
be
at
least
equal
to
the
G
4
(
s
)
order
f
or
causality
of
L
p
(
s
)
=G
4
(
s
)
[18].
The
equation
of
^
!
m
ultiplied
b
y
L
p
(
s
)
is
as
f
ollo
ws
,
^
!
=
L
p
(
s
)[
G
4
(
s
))
1
(
G
4
(
s
)(
f
w
+
!
))
f
w
]
(19)
where
L
p
(
s
)
is
designed
as
second
order
tr
ansf
er
function.
4.
Sim
ulation
and
Result
4.1.
Sim
ulation
Set
up
0
1
2
3
4
5
6
7
8
9
10
−1
−0.5
0
0.5
1
1.5
time (s)
angle (rad)
J−curve
Lane change
(a)
0
1
2
3
4
5
6
7
8
9
10
0
500
1000
1500
2000
time (s)
Force (N)
(b)
Figure
4.
Manoeuvres
a)
J-cur
v
e
and
Lane
change
b)
Side
wind
disturbance
The
perf
or
mance
of
the
proposed
control
str
ucture
is
e
v
aluated
in
sim
ulation
using
Mat-
lab/Sim
ulink
softw
are
.
As
sho
wn
in
Figure
4(a),
the
sim
ulation
is
conducted
in
tw
o
types
of
ma-
noeuvres
which
are
J-cur
v
e
and
Lane
change
.
Then,
side
wind
disturbance
as
in
Figure
4(b)
is
pur
posely
added
to
the
manoeuvres
to
in
v
estigate
the
rob
ustness
of
the
proposed
controller
.
Moreo
v
er
,
compar
ison
betw
een
the
o
utput
responses
of
CNF
controller
and
DOB
(CNF-DOB),
Propor
tional
Integ
r
al
Der
iv
ativ
e
cont
roller
and
DOB
(PID-DOB),
and
Linear
Quadr
at
ic
Regulator
Composite
Nonlinear
F
eedbac
k
With
Disturbance
Obser
v
er
f
or
...
(Sar
ah
’Atif
ah
Sar
uchi)
Evaluation Warning : The document was created with Spire.PDF for Python.
440
ISSN:
2502-4752
controller
and
DOB
(LQR-DOB)
based
systems
are
carr
ied
out
to
compare
and
analyz
ed
the
ro-
b
ustness
and
eff
ectiv
eness
of
the
proposed
control
str
ucture
.
Dur
ing
sim
ulation,
the
v
ehicle
is
assumed
r
unning
in
constant
speed
80
k
m=h
at
nor
mal
road
condition.
The
y
a
w
r
ate
ref
erence
model
is
der
iv
ed
as
[15,
16,
19],
r
ef
=
nv
(1
+
T
s
s
)
L
(1
+
K
s
v
2
)(
T
f
w
s
+
1)
sw
(20)
Here
,
K
s
is
the
stability
f
actor
,
T
s
is
the
desired
time
constant,
n
is
the
steer
ing
r
atio
,
T
f
w
is
the
dela
y
time
and
sw
is
the
steer
ing
wheel
angle
.
4.2.
Sim
ulation
Results
and
Discussion
0
1
2
3
4
5
6
7
8
9
10
0
0.05
0.1
0.15
0.2
0.25
time (s)
yaw rate (rad/s)
Reference
CNF
PID
LQR
(a)
0
1
2
3
4
5
6
7
8
9
10
−0.2
−0.1
0
0.1
0.2
time (s)
yaw rate (rad/s)
Reference
CNF
PID
LQR
(b)
Figure
5.
Sim
ulation
results
a)
Y
a
w
r
ate
response
dur
ing
J-cur
v
e
,
b)
Y
a
w
r
ate
response
dur
ing
Lane
change
Figure
5
depicts
the
output
responses
of
y
a
w
r
ate
tr
ac
king
perf
or
mances
dur
i
ng
J-cur
v
e
and
Lane
change
manoeuvres
.
Based
on
the
figure
,
the
f
astest
r
ising
and
settling
time
with
the
smallest
margin
of
steady
state
error
of
y
a
w
r
ate
tr
ac
king
response
is
accomplished
b
y
CNF-DOB
control
method,
f
ollo
w
ed
b
y
PID-DOB
and
LQR-DOB
.
Moreo
v
er
,
the
influence
of
the
e
xter
nal
side
wind
di
s
t
urbance
in
CNF-DOB
based
system
is
reduced
b
y
87%
.
The
results
sho
w
that
the
combination
of
CNF
and
DOB
is
ab
le
to
o
v
ercome
the
disturbance
issue
compared
to
the
findings
in
[7],
where
the
tr
ac
king
perf
or
mance
is
major
ly
aff
ected
b
y
the
e
xistence
of
disturbance
.
5.
Conc
lusion
In
this
paper
,
the
combination
of
CNF
and
DOB
is
successfully
proposed
as
AFS
controller
in
SBW
v
ehicle
.
Based
on
the
sim
ulation
results
,
it
can
be
concluded
that
the
CNF-DOB
based
system
is
ab
le
to
enhance
the
SBW
v
ehicle
handling
p
erf
or
mance
b
y
producing
the
f
astest
y
a
w
r
ate
tr
ac
king
responses
with
least
disturbance
eff
ects
compared
to
PID-DOB
and
LQR-DOB
.
As
f
or
the
fut
ure
w
or
k,
e
xper
iment
should
be
conducted
to
v
alidate
the
proposed
control
method
in
a
real
time
condition.
Ac
kno
wledg
ement
The
w
or
k
presented
in
this
study
is
funded
b
y
Ministr
y
of
Higher
Education,
Mala
ysia
under
Research
Univ
ersity
Gr
ant,
Univ
ersiti
T
eknologi
Mala
ysia
(v
ote
no:
13H73)
and
P
er
usahaan
Otomobil
Nasional
(PR
O
T
ON)
Sdn.
Bhd
(v
ote
no:
4C099).
Ref
erences
[1]
M.
Hosaka
and
T
.
Mur
akami,
“Y
a
w
Rate
Control
of
Electr
ic
V
ehicle
using
Steer-b
y-Wire
Sys-
tem,
”
Adv
anced
Motion
Control,
2004.
AMC
’04.
The
8th
IEEE
Inter
national
W
or
kshop
,
pp
.
31–34,
2004.
IJEECS
V
ol.
7,
No
.
2,
A
ugust
2017
:
434
441
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4752
441
[2]
I.
Mousa
vinejad,
R.
Kaz
emi,
and
M.
B
.
Khaknejad,
“Nonlinear
Controller
Design
f
or
Activ
e
F
ront
Steer
ing
System,
”
Inter
national
Jour
nal
of
Mechanical,
Industr
ial
Science
and
Engi-
neer
ing
V
ol:6
,
v
ol.
6,
no
.
1,
pp
.
1–6,
2012.
[3]
J
.
Tian,
Y
.
W
ang,
and
N.
Chen,
“Integ
r
ated
Control
of
Nonlinear
V
ehicle
Stability,
”
T
elk
ominka
,
no
.
6,
p
.
3020˜3027,
2013.
[4]
Z.
Y
onghui
and
C
.
Dingyue
,
“Research
on
Ste
er-b
y-Wire
System
in
Electr
ic
V
ehicle
,
”
TELK
OMNIKA
(T
elecomm
unication
Computing
Electronics
and
Control)
,
v
ol.
15,
no
.
1,
p
.
115,
2017.
[5]
S
.
A.
Sar
uchi,
H.
Zamzur
i,
S
.
A.
Mazlan
,
M.
Hatta,
M.
Ar
iff
,
and
M.
Af
andi,
“Activ
e
F
ront
Steer
ing
f
or
Steer-b
y-Wire
V
ehicle
via
Composite
Nonlinear
F
eedbac
k
Control,
”
Control
Con-
f
erence
(ASCC),2015
10th
Asian
,
pp
.
1–6,
2015.
[6]
M.
K.
Ar
ipin,
Y
.
M.
Sam,
A.
D
.
K
umeresan,
and
K.
P
eng,
“A
Y
a
w
Rate
T
r
ac
king
Control
of
Activ
e
F
ront
Steer
ing
System
Using
Composite
Nonlinear
F
eedbac
k,
”
Comm
unications
in
Computer
and
Inf
or
mation
Science
V
olume
(2013)
,
v
ol.
402,
pp
.
231–242,
2013.
[7]
M.
Che
Hasan,
Y
.
Sam,
K.
M.
P
e
ng,
M.
K.
Ar
ipin,
and
M.
F
.
Ismail,
“Composite
Nonlinear
F
eedbac
k
f
or
V
ehicle
Activ
e
F
ront
Steer
ing,
”
Applied
Mechanics
and
Mater
ials
,
v
ol.
663,
pp
.
127–134,
Oct.
2014.
[8]
M.
Che
Hasan,
“An
Activ
e
F
ront
Steer
ing
Control
based
on
Composite
Nonlinear
F
eedbac
k
f
or
V
ehicle
Y
a
w
Stability
System,
”
Univ
ersiti
T
eknologi
Mala
ysia
,
no
.
J
an
uar
y
,
2013.
[9]
Y
.
Huang
and
G.
Cheng,
“A
rob
ust
composite
no
nlinear
control
scheme
f
or
ser
v
omotor
speed
regulation,
”
Inter
national
Jour
nal
of
Control
,
v
ol.
88,
no
.
1,
pp
.
104–112,
A
ug.
2014.
[10]
G.
Cheng
and
K.
P
eng,
“Rob
ust
Composite
Nonlinear
F
eedbac
k
Control
With
Application
to
a
Ser
v
o
P
ositioning
System,
”
IEEE
T
r
ansactions
on
Industr
ial
Electronics
,
v
ol.
54,
no
.
2,
pp
.
1132–1140,
Apr
.
2007.
[11]
K.
Nam,
Y
.
Hor
i,
and
H.
Fujimoto
,
“Adv
anced
Motion
Control
f
or
Electr
ic
V
ehicles
Using
Later
al
Tire
F
orce
Sensors
Adv
anced
Motion
Control
f
or
Electr
ic
V
ehicles
Using
Later
al
Tire
F
orce
Sensors,
”
The
Univ
ersity
of
T
oky
o
,
no
.
J
une
,
2012.
[12]
A.
Ito
and
Y
.
Ha
y
aka
w
a,
“Design
of
F
ault
T
oler
ant
Control
System
f
or
Steer-b
y-Wire
De-
pending
on
Dr
iv
e
System,
”
T
r
ansactions
of
the
Society
of
Instr
ument
and
Control
Engineers
,
v
ol.
48,
no
.
12,
pp
.
872–881,
2012.
[13]
S
.
M.
H.
F
ahami,
H.
Zamzur
i,
S
.
A.
Mazlan,
and
M.
A.
Zakar
ia,
“Modeling
and
sim
ulation
of
v
ehicle
steer
b
y
wire
system,
”
in
2012
IEEE
Symposium
on
Humanitie
s
,
Science
and
Engineer
ing
Research
.
IEEE,
J
un.
2012,
pp
.
765–770.
[14]
B
.
L.
Boada,
M.
J
.
Boada,
and
V
.
D
´
ıaz,
“Fuzzy-logic
applied
to
y
a
w
moment
cont
rol
f
or
v
ehicle
stability,
”
V
ehicle
System
Dynamics
,
v
ol.
43,
no
.
10,
pp
.
753–770,
Oct.
2005.
[15]
M.
Nagai,
M.
Shino
,
and
F
.
Gao
,
“Study
on
Integ
r
ated
Control
of
Activ
e
F
ront
Steer
Angle
and
Direct
Y
a
w
Moment,
”
JSAE
Re
vie
w
,
v
ol.
23,
no
.
3,
pp
.
309–315,
2002.
[16]
M.
H.
Mohammed
Ar
iff,
H.
Zamzur
i,
N.
R.
Nik
Idr
is,
S
.
A.
Mazlan,
and
M.
A.
Mohamad
Nordin,
“Direct
Y
a
w
Moment
Control
of
Independent-Wheel-Dr
iv
e
Electr
ic
V
ehicle
(
IWD-
EV
)
Via
Composite
Nonlinear
F
eedbac
k
Controller,
”
2014
First
Inter
national
Conf
erence
on
Systems
Inf
or
matics
,
Modelling
and
Sim
ulation
,
pp
.
88–93.
[17]
S
.
A.
Sar
uchi,
H.
Zamzur
i,
S
.
A.
Mazlan,
S
.
M.
H.
F
ahami,
and
N.
Zulkar
nain,
“Wheel
Syn-
chronization
Control
in
Steer-b
y-Wire
Using
Composite
Nonlinear
F
eedbac
k,
”
Applied
Me-
chanics
and
Mater
ials
,
v
ol.
575,
pp
.
762–765,
J
un.
2014.
[18]
L.
G.
Tilman
Biinte
,
Dir
k
Odenthal
Bilin
Aksun-Giiv
en,
“Rob
ust
V
ehicle
Steer
ing
Control
Design
Based
On
The
Disturbance
Obser
v
er,
”
Ann
ual
Re
vie
ws
in
Control
,
v
ol.
26,
pp
.
139–
149,
2002.
[19]
M.
H.
Lee
,
S
.
Ki
Ha,
J
.
Y
.
Choi,
and
K.
S
.
Y
oon,
“Impro
v
ement
of
the
Steer
ing
F
eel
of
an
Electr
ic
P
o
w
er
Steer
ing
System
b
y
T
orque
Map
Modification,
”
Jour
nal
of
Mechanical
Science
and
T
echnology
,
v
ol.
19,
no
.
3,
pp
.
792–801,
2005.
Composite
Nonlinear
F
eedbac
k
With
Disturbance
Obser
v
er
f
or
...
(Sar
ah
’Atif
ah
Sar
uchi)
Evaluation Warning : The document was created with Spire.PDF for Python.