TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol. 13, No. 1, Janua
ry 201
5, pp. 91 ~ 1
0
0
DOI: 10.115
9
1
/telkomni
ka.
v
13i1.688
9
91
Re
cei
v
ed O
c
t
ober 2
1
, 201
4; Revi
se
d Novem
b
e
r
24, 2014; Accept
ed De
cem
b
e
r
15, 2014
Optimal Disturbance Rejection Control of
Underactuated Autonomous Underwater in Vertical
Plane
Yang Qing
1
, Gao De-xin*
2
1
Colle
ge of Info
rmation Sci
enc
e and En
gi
neer
ing, Ocean U
n
i
v
ersit
y
of C
h
in
a,
Qingd
ao C
h
in
a
2
Colle
ge of Aut
o
matio
n
an
d El
ectronic e
ngi
ne
er,
Qingda
o Un
iversit
y
of Scie
nce & T
e
chnol
og
y,
Qingd
ao C
h
in
a
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: qdga
od
e
x
in
@12
6
.com
A
b
st
r
a
ct
T
o
reali
z
e t
he
opti
m
a
l
contro
l
of under
actuat
ed aut
ono
mou
s
underw
a
ter v
ehicl
e (AUV) in
vehicl
e
pla
ne w
i
th
exte
rnal
distur
banc
es, a o
p
ti
ma
l d
i
sturba
nce
r
e
je
ction co
ntrol
l
er
is pr
opos
ed
w
i
th resp
ect to th
e
qua
dratic
perf
o
rmanc
e in
dex
es. Firs
tly, the depth c
ontro
l
mo
de
l of
u
n
d
e
ractuate
d
AU
V system
and
th
e
w
a
ve mo
del
is
propos
ed; T
h
en bas
ed o
n
the the
o
ry of
the qua
dratic o
p
t
i
mal co
ntrol a
n
d
stabil
i
ty degr
ee
constrai
nt, a f
eedforw
a
rd
a
n
d
fee
d
b
a
ck o
p
t
ima
l
distur
b
a
n
c
e re
jectio
n c
ontrol
law
w
i
th
a
hig
her
mea
n
-
squar
e co
nver
genc
e rate
is d
e
rive
d fro
m
the
Riccati
eq
uati
on a
nd th
e Syl
v
ester eq
uati
o
n, w
h
ich ca
n re
jec
t
the d
i
sturba
nc
e i
n
flue
nce
to
AUV. F
i
na
lly, t
he c
ontrol
l
er
is
ap
pli
ed t
o
the
div
e
p
l
an
e c
o
ntrol
of AUV w
i
th
w
a
ve force disturba
nces, an
d the results de
mo
nstrat
e
the effectiveness a
nd
rob
u
stness
of the controll
e
r
.
Ke
y
w
ords
:
un
deractu
ated A
U
V Systems, o
p
timal co
ntrol, disturb
ances, s
t
ability d
egre
e
Copy
right
©
2015 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
Autonomo
u
s Und
e
rwate
r
Vehicle
s
(AUV)
have
been
an a
r
ea
of a
c
tive re
se
arch for the
last few decades since t
hese v
ehicl
es have vari
ous applications
in military
,
commerci
a
l
an
d
sci
entific mi
ssion
s
. The
d
epth-kee
p
ing
contro
l for
AUV system
s is
a co
mm
on an
d impo
rtant
navigation
co
ntrol problem,
whic
h ca
n test stability of AUV, underwater depth a
n
d variable d
e
p
th
perfo
rman
ce.
The
r
e
are m
any go
od
me
thods to
solv
e this p
r
oble
m
, su
ch
a
s
P
I
D an
d im
pro
v
ed
PID method [
1
], sliding m
o
de control [2
-3], adaptiv
e
control [4], predictive
Cont
rol [5], optim
a
l
control [6-7] and ba
ckste
pping
cont
rol
[8], etc. In
vertical pl
ane
motion, the
AUV syste
m
is
inevitably infl
uen
ced
by
wi
nd, wave, flo
w
a
nd
othe
r compl
e
x
envi
r
onm
ental disturban
ce
force.
Wave fo
rce i
s
o
n
e
of
the
main
distu
r
b
ances,
whi
c
h
is treated a
s
a disturban
ce for AUV, and
modele
d
by the exo
s
yste
m. The
motio
n
situ
ation
of
AUV i
s
more compl
e
x u
nder
wave fo
rce
disturban
ce
s,
but al
so
whi
c
h
can
affect
its moti
on
control
accu
ra
cy, even ma
ke the
control
of
unsta
ble. If control failu
re, AUV will so
o
n
surfa
c
e
o
r
depth in
cre
a
ses sharpl
y, so the disturba
nce
reje
ction
prob
lems
of AUV
s
have im
po
rta
n
t sig
n
if
ica
n
ce in th
eory
an
d practi
ce. A
nother for AUV
control, the converg
e
n
c
e
speed of
syste
m
s state
s
i
s
also
an imp
o
r
tant
facto
r
which can not be
ignored, be
cause of the
faster d
e
cay, the better stability, so that
we draw i
n
to the stabil
i
ty
degree in de
sign of the control la
w for AUV sy
stem
s, based on
the linear q
u
adrati
c
optim
al
control theo
ry.
In this pape
r, a optimal disturb
a
n
c
e
reje
ctio
n co
ntrol with a
higher me
a
n
-squa
re
conve
r
ge
nce rate is propo
sed for und
era
c
tuated
AUV
with re
spe
c
t to the quad
rat
i
c perfo
rma
n
ce
indexe
s
. First
l
y, we introd
uce a mo
del
of under
a
c
t
uated AUV system in vertical plan
e, then
based
on th
e theo
ry of
linear qu
ad
ratic o
p
tima
l control
a
nd stability
de
gree con
s
trai
nt,
a
feedforwa
rd and feedb
ack optimal di
sturba
nce reje
ction controll
er
with a hig
her me
an-sq
ua
re
conve
r
ge
nce
rate
is de
riv
ed from th
e
Riccati
equ
ation a
nd th
e
Sylvester e
q
uation,
whi
c
h
is
robu
st for the
disturb
a
n
c
e i
n
fluen
ce to AUV.
The org
ani
za
tion of the paper is a
s
follo
ws. Sectio
n 2 presents the
AUV model and the
wave force
d
i
sturb
a
n
c
e
s
model. Th
e
optimal di
st
u
r
ban
ce
rej
e
ct
ion control l
a
w is
derive
d
in
Section
s
3. Then sim
u
latio
n
results a
r
e
pre
s
ente
d
in Section 4.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 13, No. 1, Janua
ry 2015 : 91 – 1
0
0
92
2.
Under
actu
a
ted
AUV Sy
stems an
d Disturba
nce
Model
2.1. Depth Control Mod
e
l and Linearization
A
sch
ematic
of the
AUV model with its body-f
ixed co
ordin
a
te syst
em is sho
w
n
in Figure
1, which is a compl
e
x non-linear sy
stem
, and has
stro
ng cou
p
ling b
e
twee
n state variable
s
, it’
s
a
very dif
f
icult
probl
em to desig
n an optimal c
ontrol law for
AUV kinem
atics system, so the
kinem
atics model of
AUV
is transfo
rme
d
into
a simple one from
the six degrees of freed
om
model p
r
op
o
s
ed by Fo
ssen, and then
the novel
model ha
s fou
r
degree
s of freed
om, and
four
indep
ende
nt input variabl
e
s
. In orde
r to facilitat
e the analysi
s
and
synthesi
s
of control sy
ste
m
,
the cou
p
ling
ef
fect betwe
e
n
the roll su
rface mo
ve
me
nt and two ca
se of plane
motion is u
s
u
a
lly
ignored, then
the vehicle motion is divided into
hori
z
ontal and vertical moveme
nt. In
this paper
,
we con
s
id
er the
vertical
movement, and assumi
ng that the
axial veloci
ty is constant,
all
transve
rse pa
ramete
r is ze
ro, and only a
AUV tail
rudder p
r
op
elle
r
,
so the
AUV
system has
only
one control i
nput
s
, and there a
r
e two
degre
e
s of
freedo
m of motion.
The
kinem
atic an
d
dynamic e
q
u
a
tions
[9]
can b
e
expre
s
sed
as follo
ws:
Figure 1. AUV model
2
2
0
2
00
[]
()
c
o
s
[(
)
]
()
c
o
s
)
s
i
n
co
s
s
i
n
GG
q
w
u
q
uw
uu
s
ww
q
q
yy
G
G
q
w
uq
uw
ww
q
q
GB
G
B
u
u
s
mw
u
q
x
q
z
q
Z
q
Z
w
Z
u
q
Zu
w
Z
w
w
Z
q
q
W
B
u
Z
Iq
m
x
u
q
w
z
w
q
M
q
M
w
Mu
q
M
u
w
M
w
w
M
q
q
xW
x
B
z
W
z
B
u
M
zw
u
q
(
(1)
Whe
r
e
is th
e pitch angl
e
,
w
is the heav
e velocity
,
s
is the cont
rol
fin angle,
yy
I
is the
moment of inertia of the
vehicl
e about the pitch axis,
u
is
the forward veloc
i
ty
,
W
denote
s
th
e
vehicle’
s wei
ght and
0
B
is th
e vehicle buo
yancy
.
The p
h
ys
ical meani
ng of other param
eters in
referen
c
e [10], the nonlinear sy
stem
(1) is
not convenie
n
t to control syst
em analysi
s
and
synthe
sis, so
the model is
linea
rzed b
a
se
d on the small pe
rturb
a
tion method
. Suppose th
e
referen
c
e mo
tion as the axial direct motion, not bow to the motion
and roll motio
n
, and seco
n
d
orde
r coef
fici
ent is relatively
small, wh
ich
can b
e
n
egle
c
ted,
(
,
,
)(,
,
)
0
GG
G
B
B
B
xy
z
x
y
z
,
then the linea
r equatio
n group is avail
a
b
l
e, as follow:
2
2
00
0
0
00
0
0
()
00
1
0
0
1
0
0
0
00
0
1
1
0
0
0
wq
u
w
u
q
uu
wy
y
q
u
w
u
q
uu
s
mZ
Z
w
Z
Z
w
Zu
MI
M
q
M
M
u
q
Mu
t
zu
z
(2)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Optim
a
l Distu
r
ban
ce
Reje
ction Control
o
f
Underactu
ated Autonom
o
u
s…
(Yang Q
i
ng)
93
2.2. Disturb
a
nces Mod
e
l of Wav
e
Force
The extern
al disturban
ce
s
for
AUV’
s
are
compl
e
x, and wave force
is one of th
e main
disturban
ce
s.
In ord
e
r to
st
udy
co
nvenie
n
tly
,
the irreg
u
lar lo
ng
storm wave
s i
s
si
mplified a
s
p
o
int
long cre
s
ted
wave
s, as foll
ow:
11
()
()
c
o
s
ll
jj
j
jj
tt
L
(3)
Whe
r
e,
l
is the numb
e
r of
comp
one
nt wave,
2(
)
jj
j
LS
,
j
jj
t
,
j
i
s
a
rand
om vari
a
b
le, by wave
theory,
it is uniform
dist
ribution
between
0
-
2
,
j
is the
j
comp
one
nt wave freque
ncy,
()
S
is the Ocea
n wave spe
c
trum de
nsity functio
n
.
We
co
nst
r
u
c
t a
system
mo
del to
de
scrib
e
the
irregul
a
r
wave fo
rces for th
e A
U
V
in two
-
dimen
s
ion
a
l hori
z
ontal pl
a
ne.
Define
cos
(
)
jj
j
vL
is th
e ho
ri
zontal
velocity of
water
parti
cle
orbital
motion
. Let
1
()
T
l
vt
v
v
,
is the
j
v
frequ
ency. By
2
jj
j
vv
,
1
,
2
,
...,
j
l
, we have:
vv
(4)
Whe
r
e
22
2
12
dia
g
{
,
,
...,
}
l
.
Define
()
()
,
(
)
T
TT
w
t
vt
vt
, then:
0
()
()
()
0
()
0
(
)
I
wt
wt
G
w
t
vt
I
w
t
(5)
Wwhere
I
is
the
l
dimen
s
ion
a
l
unit matrix,
and
0
is
the
l
dimensi
onal
ze
ro matrix.
Acco
rdi
ng to
the li
nea
r
wave th
eo
ry, the
re
sulta
n
t force fo
r the A
U
V
system i
s
1
()
(
)
()
l
jj
j
j
Ft
T
v
t
, where
()
j
j
T
is the st
re
ss
co
efficient, whi
c
h i
s
dete
r
mi
ned by th
e
freque
ncy of the co
rrespon
ding wave.
1
1
(
)
()
()
(
)
()
()
0
(
)
()
l
l
Ft
T
T
v
t
TT
I
w
t
Hw
t
(6)
So the effect on AUV of the total wave distur
ban
ce
can be de
scrib
ed by the followin
g
system
:
()
(
)
()
()
wt
G
w
t
F
tH
w
t
(7)
AUV in the proce
s
s of operation, the wa
ve di
sturb
a
n
c
e can b
e
dire
ctly put into the AUV
dynamics mo
de l as extern
al distur
bing f
o
rce, so
we h
a
ve the vertical motion mo
del for co
nsta
nt
spe
ed AUV system, as foll
ow:
()
()
()
()
s
x
tA
x
t
B
t
F
t
(8)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 13, No. 1, Janua
ry 2015 : 91 – 1
0
0
94
Whe
r
e,
1
00
0
0
00
0
0
,
00
1
0
0
1
0
0
00
0
1
1
0
0
w
q
uw
uq
wy
y
q
u
w
u
q
wm
Z
Z
Z
Z
qM
I
M
M
M
u
xA
zu
,
1
2
2
00
00
00
1
0
0
00
0
1
0
wq
uu
wy
y
q
uu
mZ
Z
Z
u
MI
M
M
u
B
3. Design Op
timal Contr
o
ller
3.1. Design
of the
Con
t
r
o
l La
w
In AUV
operation, the
state of
the sy
stem converges
faster,
it
s stability is better. In order
to enhan
ce th
e stability of AUV, we ca
n cho
o
se the following
quad
ratic perfo
rma
n
ce in
dex:
2
0
1
[(
)
(
)
(
)
(
)
]
2
tT
T
ss
J
ex
t
Q
x
t
t
R
t
d
t
(9)
W
h
er
e
Q
and
R
re
sp
ectively
is
semi
-defini
t
e and
p
o
sitiv
e
-definite
ma
trix.
0
is a
k
now
n
scalar fun
c
tio
n
. The
optim
al control
pro
b
le
m i
s
to
se
arch the
opti
m
al control la
w
*
()
s
t
, which
make
s the va
lue of perfo
rmance index
(12
)
minimum
.
Theorem 1
:
Con
s
id
er the
LQR p
r
o
b
le
m of the syst
em (8
) with t
he pe
rform
a
n
c
e ind
e
x
(9), the optim
al control LQ
R is existe
nt and uni
que, a
nd its form a
s
follows:
*1
1
()
[
(
)
(
)
]
ˆ
[(
)
(
)
]
T
sw
T
w
tR
B
P
x
t
P
w
t
RB
P
x
t
P
w
t
(10)
Whe
r
e
P
is the unique
soluti
on the
Riccati
matrix equation.
()
(
)
0
T
AI
P
P
A
I
P
S
P
Q
(11)
v
P
is the uniqu
e
solution of th
e matrix differential equ
atio
n.
[(
)
]
(
)
T
ww
A
IP
S
P
P
I
G
P
H
(12)
Whe
r
e
1
T
SB
R
B
.
Proof:
Orde
r
()
(
)
()
()
()
()
t
t
s
t
x
te
x
t
ut
e
t
wt
e
w
t
(13)
Takin
g
(1
3) to (8) a
nd (9
),
after s
i
mplification we get:
0
()
()
()
()
(0
)
x
tA
x
t
B
u
t
H
w
t
xx
(14)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Optim
a
l Distu
r
ban
ce
Reje
ction Control
o
f
Underactu
ated Autonom
o
u
s…
(Yang Q
i
ng)
95
Whe
r
e
()
A
AI
.
And the new
perfo
rman
ce i
ndex as follo
ws:
0
1
[(
)
(
)
(
)
(
)
]
2
J
xt
Q
x
t
u
t
R
u
t
d
t
(15)
Acco
rdi
ng to
Pontryagin
m
a
ximum p
r
in
ciple, the
opt
i
m
al control p
r
oblem
of sy
stem (14) with
the
quadratic perf
ormance index (18)
leads t
he followi
ng
TPBV problems:
0
(
)
()
()
(
)
()
()
(
)
(0
)
,
(
)
0
T
tQ
x
t
A
t
x
tA
x
t
S
t
H
w
t
xx
(16)
And the optimal cont
rol la
w ca
n be exp
r
esse
d as:
1
()
()
T
ut
R
B
t
(17)
In order to s
o
lve the TPBV
problems
(17), let:
()
()
()
w
tP
x
t
P
w
t
(18)
Whe
r
e
,
w
PP
are
pendi
ng m
a
trixes, de
rivate
two
side
s
of (21
)
, an
d
su
bstituting the
se
co
nd
type of (16), we get:
()
()
(
)
{
(
)
[
()
()
]
()
}
[
()
()
]
()
()
()
(
)
()
()
(
)
()
(
)
()
w
w
tt
w
t
w
tt
ww
w
w
tP
x
t
P
w
t
P
A
xt
SP
xt
P
w
t
H
w
t
P
ew
t
G
ew
t
PA
P
S
P
x
t
P
S
P
e
w
t
PH
e
w
t
P
P
G
e
w
t
PA
P
S
P
x
t
P
S
P
w
t
PH
w
t
P
I
G
w
t
(19)
By adding (1
9) into the first expression
of (16), it follows:
()
(
)
[(
)
(
)
]
(
)
0
T
T
ww
AP
P
A
P
S
P
Q
x
t
AP
S
P
P
I
G
P
H
w
t
(20)
Becau
s
e
of
selectin
g eithe
r
()
x
t
,
()
wt
and, e
q
u
a
tion (20
)
is
all hold,
so
we
ca
n get
matrix
differential eq
uation
s
of
P
,
w
P
.So we can get
)
(
t
, then from (1
8):
1
()
[
(
)
(
)
]
T
w
ut
R
B
P
x
t
P
w
t
(21)
Referen
c
e to (7) a
nd (1
3), the feedforwa
rd-feed
ba
ck o
p
timal disturb
ance reje
ctio
n control law
of
system (8)
ca
n be uniq
ue confirme
d.
*
1
1
()
()
*
{
[
(
)
(
)]}
=
[
()
()
]
t
s
tT
t
t
w
T
w
te
u
t
eR
B
P
e
x
t
P
e
w
t
RB
P
x
t
P
w
t
(22)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 13, No. 1, Janua
ry 2015 : 91 – 1
0
0
96
Theo
rem 1 is
proved.
Notic
e
1.
Co
mpared
with the cla
s
sical feedba
ck opt
i
m
al co
ntrol l
a
w, the fee
d
forward-
feedba
ck opti
m
al distu
r
ba
nce
reje
ction
control la
w
(22
)
ha
s the
feed-fo
rward
items. So for a
system with
a disturban
ce
its c
ontrol p
e
rform
a
n
c
e i
s
clea
rly sup
e
rio
r
to the classical feedb
ack
optimal co
ntrol law.
Lemma 1
If
(,
)
A
B
is co
mpletely controllabl
e, then
(,
)
A
B
is compl
e
tely controll
able
Proof.
If
(,
)
A
B
is completely cont
rollabl
e, so:
1
[,
,
,
]
n
ra
nk
B
A
B
A
B
n
(23)
Let,
()
(
)
KK
K
K
A
AI
A
f
A
(24)
Whe
r
e,
()
K
f
A
is Low-level
sub
-
K
expressio
n
, so
:
21
21
21
21
,,
,
,
,,
(
)
,
,
(
)
,,
,
,
n
n
n
n
r
a
n
k
BA
BA
B
A
B
r
a
n
k
B
A
B
B
AB
f
A
B
A
B
f
AB
r
a
n
k
BA
BA
B
A
B
n
(25)
The proof is complete. Similarl
y the observability can prove.
3.2. Design
of the
Distu
r
bance
s
Obs
e
rv
er
In fac
t,
)
(
t
w
in (22) i
s
un
kn
o
w
n for it i
s
the
state ve
ctor of exo
system (2
). T
h
e
feedforwa
rd
control term i
n
(2
2) i
s
p
h
ysically
unreal
izabl
e in the
pra
c
tical
engi
neeri
ng. In th
is
se
ction, we in
trodu
ce a di
sturba
nce ob
se
rver to ma
ke i
t
realiza
b
le.
Suppo
se th
a
t
exosystem
(2
) is ob
servabl
e
com
p
letely. Con
s
tru
c
t a
dist
urba
nce
observe
r as f
o
llows:
00
ˆ
ˆˆ
()
()
[
(
)
(
)
]
ˆˆ
()
w
t
Gw
t
K
F
t
Hw
t
wt
w
(26)
Whe
r
e
)
(
ˆ
t
w
is the
output ve
cto
r
of
(26
)
,
K
is t
he o
b
serve
r
matrix of a
p
p
r
op
riate
dime
nsio
ns.
And the obse
r
ver erro
r is d
enoted a
s
:
)
(
ˆ
)
(
)
(
~
t
w
t
w
t
w
(27)
Then we hav
e:
)
(
~
)
(
)
(
~
t
w
KH
G
t
w
(28)
Bec
a
us
e
)
,
(
H
G
is obse
r
vable, ei
genvalu
e
s of
KH
G
can be cho
s
en to ma
ke the
observe
r e
r
ror ve
ctor
)
(
~
t
w
conve
r
ge
s t
o
ze
ro
at a
n
app
ointed
spe
ed
of expone
ntial
attenuation, that is:
0
l
i
m
(
)l
i
m
e
x
p
(
(
)
)
(
)0
tt
wt
G
K
H
t
w
t
(29)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Optim
a
l Distu
r
ban
ce
Reje
ction Control
o
f
Underactu
ated Autonom
o
u
s…
(Yang Q
i
ng)
97
So the phy
si
cally realization
of the
co
ntrol la
w (22
)
can b
e
gu
arante
ed
and
(22
)
i
s
rewritten as
follows
:
*1
ˆ
()
[
(
)
(
)
]
T
sw
tR
B
P
x
t
P
w
t
(30)
4. Simulatio
n
Example
The hyd
r
ody
namic
co
efficients of a fo
reign typical AUV to the n
o
minal m
odel
[10] as
follows
:
-
1
.
040
0.
865
-
0
.
020
0
-
0.
072
6.
000
-
0
.
681
0.
708
0
-
0.
722
()
01
0
0
0
10
2
0
0
s
ww
qq
t
zz
(31)
The initial
state of AUV is
0
050
T
, taking the
axial velocity 2
m
/s, and the
para
m
eters o
f
wave force
disturban
ce
s as follows,
12
3
4
()
1
.
1
,
()
0
.
8
,
()
1
.
3
,
()
1
.
5
TT
T
T
, The
para
m
eters o
f
the quadrati
c
perfo
rma
n
ce index
,1
QI
R
,
0.5
.
(1) Sele
ct
01
10
G
,
(0
)
1
0
T
w
The
wave f
o
rce di
sturb
ances
are sinusoidal
sig
nal. Using
LQR
co
ntroll
er a
nd o
p
timal
disturban
ce rejectio
n
contro
ller
(O
DRC), and the
sim
u
lati
on
comp
arative curve
s
of
()
,
(
)
x
tu
t
,
as
follows
:
Figure 2. State vector
1
()
x
t
Figure 3. State vector
2
()
x
t
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 13, No. 1, Janua
ry 2015 : 91 – 1
0
0
98
Figure 4. State vector
3
()
x
t
Figure 5. State vector
4
()
x
t
Figure 6. Con
t
rol vector
()
ut
(2) Sele
ct
-0
.2
2
.
3
2.
1
0
.
4
G
,
(0
)
1
0
T
w
The wave fo
rce di
sturban
ce
s are the
conv
e
r
ge
nt signal, Usi
ng
LQR
controll
er and
optimal di
stu
r
ban
ce
rej
e
ct
ion
controller (O
DRC)
, an
d the
simul
a
tion compa
r
a
t
ive curve
s
of
()
,
(
)
x
tu
t
, as
follows
:
Figure 7. State vector
1
()
x
t
Figure 8. State vector
2
()
x
t
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Optim
a
l Distu
r
ban
ce
Reje
ction Control
o
f
Underactu
ated Autonom
o
u
s…
(Yang Q
i
ng)
99
Figure 9. State vector
3
()
x
t
Figure 10. State vector
4
()
x
t
Figure 11. Co
ntrol vecto
r
()
ut
From th
e si
mulation
cu
rves, it ca
n b
e
se
en that
the presente
d
optimal disturban
ce
reje
ction
co
n
t
roller i
s
effective, and it
is mo
re
rob
u
st abo
ut exte
rnal
distu
r
ba
nce
s
tha
n
L
Q
R
controlle
r.
5. Conclusio
n
This pap
er concentrate
s on
u
nde
ra
ctu
a
ted
AU
V
sy
stem
s contro
l pro
b
lem i
n
vertical
plane affe
cte
d
by the wa
ve force di
st
urba
nc
es, a
n
d
based on t
he qua
drati
c
optimal co
ntrol
theory a
nd
st
ability deg
ree
co
nstraint, the o
p
timal di
sturb
a
n
c
e
rej
e
ction
control
l
er
with a
hig
h
e
r
mean
-squa
re
conve
r
ge
nce
rate i
s
de
rive
d from
the
Ri
ccati equ
atio
n and th
e Sylvester
equati
on,
and
we
intro
d
u
ce
a
distu
r
b
ance o
b
serve
r
to m
a
ke
it realizable.
Si
mulation
re
su
lts sho
w
that
the
desi
gne
d co
ntrol la
w ha
s a good
con
v
ergen
ce
effect and
effectively supp
ress the exte
rnal
disturban
ce
s.
Ackn
o
w
l
e
dg
ements
This work wa
s su
ppo
rted i
n
part by the Nation
al Natu
ral Scie
nce Found
ation of Chin
a (6
0804
005
),
and by the Nat
u
ral Sci
e
n
c
e
Found
ati
on of
Shandon
g Province (ZR20
11FQ0
06
),
b
y
the Natural Scien
c
e Fo
und
ation of
Qing
dao City(1
2-1
-
4-3-(17
)-j
ch
)
Referen
ces
[1]
Xi
ao JP. T
he researc
h
for AUV deep-s
e
t and
pitch control.
Dali
an: Da
lia
n Mariti
me Un
ive
r
sity
. 2009.
[2]
Bian
XQ, C
h
e
ng
XQn, Ji
a H
M
, etal. A b
o
ttom foll
o
w
i
ng c
ontrol
l
er for
un
deractu
ated A
U
V bas
ed
o
n
iterative sl
idi
n
g
and incr
ement
feedback.
C
o
n
t
rol and D
e
cisi
on
. 201
1; 26(2)
: 289–2
93.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 13, No. 1, Janua
ry 2015 : 91 – 1
0
0
100
[3]
BESSA WM, D
U
T
R
A MS, KR
EUZER E. Depth control of
remotely
operated under
w
a
t
e
r vehicles usi
ng
an ad
aptiv
e fuzz
y
sli
d
in
g mod
e
control
l
er.
Ro
bust and A
u
ton
o
mous Syste
m
s
. 2008; 56(
8): 670
–6
77.
[4]
Z
hang LJ, Qi X, Z
hao JM, Pang YJ. Dept
h-kee
p
in
g cont
rol for auton
o
m
ous un
der
w
a
ter vehicle i
n
vertical pl
an
e u
s
ing a
daptiv
e
o
u
tput feed
back
controll
er.
Co
ntrol T
heory &
Appl
icatio
ns
. 2
012; 29(
10):
137
1-13
76.
[5]
Du L, Yan W
S
, Gao J, Chen
YH. Model
ing
and mo
del
pr
edictiv
e contro
l
of AUV lon
g
it
udi
nal m
o
tion
und
er w
a
ve
dis
t
urbanc
es.
Jou
r
nal of Syste
m
Simulati
on
. 20
12; 24(6): 1
180
-118
4
[6]
Xi
ao JP. T
he researc
h
for AUV deep-s
e
t and
pitch control.
Dali
an: Da
lia
n Mariti
me Un
ive
r
sity
, 2009.
[7]
Moreira
L, S
o
ares CG.
H2
an
d H
∞
d
e
si
gns for
div
i
n
g
an
d co
urse
control
of a
n
Autonom
ous
und
er
w
a
ter ve
hicle i
n
pres
en
ce of
w
a
ves.
Journ
a
l of oce
a
n
ic en
gin
eer
ing
. 2008; 33(
2): 69-88.
[8]
Lin Y-Yi. Study
of ship course
nonlinear
control
bas
ed backstepping.
Dali
an: Dal
i
an Mariti
me
Univers
i
ty,
200
7.
[9]
Mugd
ha S
Nai
l
, Sah
jen
d
ra
N Sin
gh. Stat
e-de
pen
de
nt R
i
ccati e
q
u
a
tion
- base
d
ro
bust
drive
pl
a
n
e
control of AUV
w
i
t
h
contro
l co
nstraints.
Ocea
n Engi
ne
erin
g
. 200
7; 34: 171
1
-
1723.
[10]
Prestero T
.
Verificati
on
of a
si
x-de
gre
e
of
freed
om simu
latio
n
mo
del
for the
REMU
S auto
nomo
u
s
und
er
w
a
ter ve
hicles.
USA: M
I
T
, 2001.
Evaluation Warning : The document was created with Spire.PDF for Python.