Int
ern
at
i
onal
Journ
al of
R
obot
ic
s
and
Autom
ati
on (I
JRA)
Vo
l.
8
,
No.
4
,
D
ece
m
ber
201
9
,
pp.
293
~
300
IS
S
N:
20
89
-
4856
,
DOI: 10
.11
591/
i
jra
.
v
8
i
4
.
pp
293
-
300
293
Journ
al h
om
e
page
:
http:
//
ia
escore.c
om/j
ourn
als/i
ndex.
ph
p/IJRA
Robust a
daptiv
e contr
oller desi
gn fo
r excavato
r ar
m
Nga
Thi
-
Thu
y Vu
School
of El
ec
tr
i
ca
l
Engi
n
ee
r
ing,
Hanoi
Univer
si
t
y
of
Sci
ence and
Technol
og
y
,
V
i
et
n
am
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
A
pr
26
, 201
9
Re
vised
A
ug
27, 2
019
Accepte
d
Oct
6
, 2
01
9
Thi
s
pap
er
pr
ese
nts
a
robust
a
dapt
iv
e
con
trol
l
er
th
at
does
not
depe
nd
o
n
the
s
y
stem
par
amete
rs
for
an
ex
ca
va
tor
ar
m
.
Firstl
y
,
th
e
m
odel
of
the
excavator
ar
m
is
demons
trat
ed
in
the
Eu
le
r
-
La
gra
ng
e
form
conside
ring
with
over
a
ll
excava
tor
s
y
stem.
Next,
a
robust
a
dapt
iv
e
controll
er
has
be
e
n
construc
t
ed
fro
m
informati
on
of
sta
te
err
or
.
I
n
thi
s
pap
er,
th
e
st
abi
l
ity
of
over
all
s
y
s
te
m
is
m
at
hematica
l
l
y
prove
n
b
y
using
L
y
apunov
stabil
ity
th
eor
y
.
Also,
the
propos
ed
cont
ro
ll
er
is
m
odel
fre
e
the
n
the
c
losed
loop
s
y
stem
is
not
aff
ecte
d
b
y
distu
rba
nce
s a
nd
unc
ert
a
int
i
es.
Fina
lly
,
th
e
sim
ula
t
ion
is
execut
ed
in
Matlab/Sim
ul
ink
for
both
p
re
sente
d
sch
eme
a
nd
the
PD
contr
oll
er
und
er
som
e
condi
ti
on
s
to
ensure
th
a
t
th
e
proposed
al
gor
it
hm
given
the
good
per
form
anc
es
fo
r
all
ca
ses
.
Ke
yw
or
d
s
:
A
da
ptive
Euler
-
l
agran
ge
syst
e
m
Exca
vator
R
obus
t c
on
t
ro
l
Copyright
©
201
9
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
Ng
a
T
hi
-
T
hu
y
Vu
,
School
of Elec
tric
al
Engineer
ing
,
Hanoi
Un
i
ver
si
ty
o
f Sci
ence a
nd Tec
hnology
,
1 Dai C
o Viet
, Ha
No
i,
V
ie
tn
a
m
.
Em
a
il
:
ng
a.vut
hithu
y
@hust.e
du.vn
1.
INTROD
U
CTION
Nowa
days,
r
obots
a
re
us
e
d
c
omm
ently
in
the
i
ndust
ries
because
of
their
versat
il
ity
and
eff
ic
ie
ncy
[1
-
3]
.
The
aut
om
atic
rem
ote
con
t
ro
l
of
the
r
obot
play
s
a
si
gn
i
ficant
r
ole
i
n
r
eal
-
li
fe
ap
plica
ti
on
,
su
c
h
as
nucl
ear
fiel
d,
c
onstructio
n,
a
nd
r
escue
m
issi
on
s
.
For
the
e
xc
avato
r
r
obot,
i
n
orde
r
to
perform
a
sp
eci
fic
duty
,
it
nee
d
s
to
com
plete
at
le
ast
t
wo
ta
s
ks
:
determ
ining
a
feas
ible
pat
h
f
ro
m
it
s
init
ia
l
loca
ti
on
t
o
the
desti
nation
an
d
the
n
e
xec
uting
the
ta
s
k
thr
ough
co
ntr
ol
al
gorithm
that
has
to
be
des
ign
e
d.
Acc
ordi
ng
to
these
re
qu
i
rem
ents,
th
e
t
rack
i
ng
c
ontr
ol
pro
blem
fo
r
the
e
xcav
at
or
rob
ot
syst
e
m
is
con
sta
ntly
receiving
t
he
interest
ing
of
sci
entist
s.
The
earli
er
resea
rc
h
w
ork
m
ai
nly
fo
c
us
e
d
on
m
od
el
li
ng
w
ork
includi
ng
ki
ne
m
at
ic
and
dynam
ic
m
od
el
,
m
od
el
li
ng
of
interact
ion
betwee
n
t
he
m
achine
a
nd
t
he
e
nvir
onm
ent,
an
d
pa
ram
et
er
identific
at
ion
[
4
-
10
]
.
Mo
delli
ng
a
nd
par
am
et
er
identific
at
ion
duri
ng
t
he
op
e
rati
on
of
m
achine
is
ver
y
help
fu
l
for
the
r
eal
-
ti
m
e
m
on
it
or
in
g
a
nd r
em
ote control.
Abo
ut
the
co
nt
ro
l
desig
n
,
duri
ng
the
ea
rlie
r
sta
ge
of
stud
y
on
e
xca
va
tor,
im
ped
an
ce
co
ntr
ol
is
consi
der
e
d
a
s
a
po
pu
la
r
co
ntr
ol.
I
n
[
11
]
,
a
posit
ion
-
ba
sed
im
ped
anc
e
co
ntr
oller
is
pre
sente
d
on
m
ini
-
excav
at
or.
I
n
[
12
-
1
3
]
,
aut
hors
prese
nt
det
ai
l
of
r
obust
im
ped
ance
co
nt
ro
l
f
or
hy
dr
a
ulic
exca
vato
r.
Th
e
i
m
ped
ance
co
nt
ro
l
s
uit
s
to
a
pply
f
or
exc
avat
or
beca
us
e
it
c
an
deal
with
both
f
ree
a
nd
c
onstrai
n
m
otion
[
1
3
]
.
Howe
ver,
the
al
gorithm
is
qu
it
e
c
om
pli
cat
ed.
Re
centl
y,
m
any
m
odern
c
ontr
ol
te
chn
i
qu
e
s
are
us
e
d
in
trajecto
ry
c
ontrol
of
exca
vato
r
a
rm
.
In
[1
4
]
,
an
ada
ptive
co
ntr
olle
r
is
pr
es
ented
in
c
on
t
rol
li
ng
e
xca
vato
r
arm
.
The
sta
bili
ty
of
syst
em
is
ens
ur
e
d
thr
ough
m
at
he
m
at
ic
a
l
pro
of
an
d
ve
ri
fied
by
sim
ula
ti
on
res
ults.
H
ow
e
ve
r,
the
sim
ulatio
n
as
well
as
the
exp
la
natio
n
of
si
m
ulati
o
n
res
ults
is
qu
it
e
po
or.
I
n
[
1
5
-
1
6
]
,
the
fuzzy
co
n
tr
ollers
are
em
plo
ye
d
to
s
olv
e
t
he
trackin
g
c
ontr
ol
pr
ob
le
m
of
the
e
xcav
at
or.
I
n
th
ese
ty
pe
of
c
on
tr
ollers,
t
he
inf
or
m
at
ion
ab
ou
t
the
syst
em
do
e
s
not
re
quire.
Howe
ver
,
t
he
sta
bili
ty
of
the
overall
sys
tem
is
no
t
sho
wn
i
n
the
m
at
he
m
at
i
c.
Nowa
days,
t
he
PID
c
o
ntr
oller
sti
ll
bein
g
us
e
d
widely
in
the
pract
ic
e
be
cause
of
it
s
sim
pl
ify
.
Howe
ver,
tu
nin
g
of
the
PID
gains
t
o
a
dap
t
with
the
c
ha
nge
of
w
orkin
g
c
onditi
ons
is
di
f
ficult
an
d
dep
e
nd
i
ng
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2089
-
4856
I
nt
J
R
ob
&
A
uto
m
,
Vo
l.
8
, No
.
4
,
Decem
ber
201
9
:
293
–
300
294
on
the
pe
rs
on
al
ex
per
ie
nces.
I
n
recent
ti
m
es,
in
or
der
to d
eal
with
t
his p
r
oblem
,
so
m
e
op
ti
m
iz
at
ion
te
chni
qu
es
su
c
h
as
a
rtific
ia
l
ne
ur
al
ne
tw
ork
(
ANN
),
an
t
col
on
y
opti
m
iz
at
ion
(A
C
O),
et
c.,
hav
e
bee
n
a
ppli
ed
to
op
tim
iz
e
the
P
ID
par
am
et
ers.
In
[1
7
]
,
an
ge
netic
al
gorithm
(GA)
i
s
use
d
to
dete
rm
ine
the
P
ID
gai
ns
for
traj
ect
or
y
trackin
g
c
on
tr
ol
of
r
obotic
ex
cavato
r.
T
he
pr
esented
sc
hem
e
gav
e
t
he
go
od
pe
rfor
m
ance
s
in
com
par
iso
n
with
so
m
e g
iven
m
e
thods.
Howe
ve
r,
t
he go
tt
en
r
e
su
lt
s sti
ll
h
ave
pro
blem
s n
eeded
to
b
e
d
isc
usse
d.
In
this
pa
pe
r,
a
r
obus
t
ada
pti
ve
c
ontr
oller
i
s
pr
opos
e
d
f
or
exca
v
at
or
ar
m
syst
e
m
.
The
str
uctu
re
of
con
t
ro
ll
er
c
ons
ist
s
of
t
wo
pa
r
ts:
the
first
pa
r
t
is
respo
ns
ibl
e
for
kee
ping
t
he
sta
bili
ty
of
the
syst
em
and
the
seco
nd
pa
rt
is
us
e
d
f
or
a
dap
ti
ng
with
the
un
known
pa
ram
e
te
rs.
T
heref
or
e
,
the
pro
posed
con
t
ro
ll
er
has
abili
ty
t
o
ca
ncel
t
he
e
ff
ect
of
t
he
un
certai
nties
as
well
as
to
ke
e
p
t
he
tracki
ng
error
go
to
ze
r
o.
Als
o,
the
presented
con
t
ro
ll
er
is
si
m
ple
so
it
is
e
asy
to
i
m
ple
m
ent.
The
feasib
le
of
the
al
gori
thm
s
is
de
m
onstrat
ed
by
Ly
a
punov
sta
bili
ty
theor
y
an
d
ve
rifie
d
t
hro
ugh
sim
ulatio
n
m
od
el
s.
T
he
sim
ulatio
n
is
exec
uted
in
MATLAB/Si
m
ulink
for
both
prese
nt
ed
schem
e
an
d
the PD
c
ontr
oller
un
der
s
om
e
conditi
on
s to
ens
ure
that
t
he
pr
opos
e
d
al
gorith
m
giv
e
n
the
go
od p
e
rfor
m
ances
for
al
l case
s
.
2.
CONTR
OLL
ER D
E
SIG
N FOR
E
X
C
A
V
ATOR A
R
M
2
.
1.
M
od
el
li
n
g of exc
avator
a
rm
Con
si
der
a
n
e
xc
avato
r
syst
em
w
it
h
str
uctu
re
as Fig
ur
e
1. Th
e syst
e
m
co
ns
i
sts of tw
o
s
ub
s
yst
e
m
s: t
he
base
is
us
e
d
to
m
ov
e
t
he
enti
re
syst
em
on
t
he
x
0
O
0
y
0
,
plan
e
a
nd
the
arm
is
us
ed
for
m
ov
e
m
ent
in
the
z
0
O
0
y
0
and
z
0
O
0
x
0
plan
es
.
This
pa
per
c
on
ce
r
n
m
ai
nly
on
the
m
otion
con
t
ro
l
of
e
xca
vato
r
a
rm
,
so
t
he
base
pa
rt
a
nd
th
e
ro
ta
ti
on ar
ou
nd
O
0
z
0
axis
are c
on
si
der
ed
unch
ang
e
d.
Figure
1. Sc
he
m
at
ic
d
ia
gr
am
of an
ex
ca
vato
r
The
E
uler
-
Lag
range
m
od
el
of
e
xcav
at
or
arm
du
rin
g
th
e
diggin
g
pro
cess
corres
pondin
g
to
the
coor
din
at
es
of
each
j
oi
n
as
shown i
n
Fi
g
ure
1
is as
foll
ow [18
]
:
(
)
(
,
)
(
)
(
)
L
D
C
G
B
F
+
+
+
=
−
(1)
wh
e
re
2
3
4
=
is
the
posit
ion
of
e
ach
j
oi
nt
in
t
he
j
oi
nt
sp
ace
,
D
(
)
re
pr
ese
nts
ine
rtia
l
pa
rt,
(
,
)
C
is
the
Cori
olis
and
ce
ntri
petal
eff
ect
s,
G
(
)
is
the
gr
a
vity
par
t,
()
B
represe
nts
fr
ic
ti
ons;
i
s
the
corres
pondin
g
in
pu
t
m
at
rix,
1
2
3
=
sp
eci
fies
t
he
to
rques
act
ing
on
t
he
s
ha
ft
of
3
joints,
F
L
represe
nts the
i
nteracti
ve
t
orq
ues. The
for
m
ulas o
f
abo
ve pa
rts are
g
i
ven by
the foll
owin
g ex
pr
es
sio
ns
:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
R
ob
&
A
uto
m
IS
S
N:
20
89
-
4856
Ro
bu
st
adapti
ve co
nroll
er d
es
ign
f
or
… (
Ng
a Thi
-
T
huy V
u)
295
1
1
1
2
1
3
2
1
2
2
2
3
3
1
3
2
3
3
()
D
D
D
D
D
D
D
D
D
D
=
,
1
1
1
2
1
3
2
1
2
2
2
3
3
1
3
2
3
3
(
,
)
C
C
C
C
C
C
C
C
C
C
=
,
1
1
0
0
1
1
0
0
1
−
=
−
,
(2)
2
3
4
()
G
G
G
G
=
,
2
3
4
()
b
o
s
t
b
u
B
B
B
B
=
,
(3)
wh
e
re:
(
)
(
)
(
)
2
33
4
22
22
33
3
3
3
4
4
4
22
11
22
2
2
2
3
3
3
2
2
2
3
3
2
4
34
4
2
c
os
2
c
os
2
2
a
r
c
os
bu
bu
s
t
s
t
bu
bo
bo
s
t
bu
D
I
M
r
D
D
I
M
r
M
a
a
r
D
D
I
M
r
M
a
a
r
M
a
a
a
c
=+
=
+
+
+
+
+
=
+
+
+
+
+
+
+
+
+
(
)
(
)
(
)
(
)
2
3
3
2
3
3
3
4
4
4
1
3
3
1
2
3
2
4
3
4
4
2
1
2
2
1
1
3
3
2
3
3
3
2
3
2
3
3
3
4
4
4
c
o
s
c
o
s
c
o
s
c
o
s
bu
bu
s
t
s
t
bu
D
D
D
M
a
r
D
D
D
M
a
r
D
D
D
I
M
r
a
r
M
a
a
a
c
a
r
=
=
+
+
=
=
+
+
=
=
+
+
+
+
+
+
+
+
(
)
(
)
(
)
(
)
(
)
(
)
1
1
2
3
2
3
3
3
2
3
2
3
3
2
4
2
3
4
3
4
4
1
2
2
3
2
3
3
3
2
3
2
3
3
2
4
2
3
4
3
4
4
1
3
2
4
2
3
4
3
4
4
2
1
2
2
3
3
3
3
3
3
4
2
3
4
4
si
n
si
n
si
n
si
n
si
n
si
n
si
n
s
t
b
u
b
u
s
t
b
u
b
u
bu
b
u
s
t
b
u
C
M
a
r
M
a
a
s
M
a
r
C
M
a
r
M
a
a
s
M
a
r
C
M
a
r
C
a
M
a
s
M
r
M
a
r
=
−
+
−
−
+
=
−
+
−
−
+
=
−
+
=
+
+
−
+
(
)
(
)
(
)
(
)
(
)
(
)
(
)
4
2
2
3
4
2
3
4
4
4
2
3
3
4
2
3
4
4
4
3
1
4
2
2
3
4
4
3
4
4
3
4
3
4
4
3
2
3
4
4
4
33
si
n
si
n
si
n
si
n
si
n
si
n
0
bu
bu
b
u
b
u
bu
C
M
a
r
C
M
a
r
C
M
r
a
a
M
a
r
C
M
a
r
C
=
−
+
=
−
+
=
+
+
+
+
+
=+
=
(
)
(
)
(
)
(
)
2
2
2
2
2
2
3
3
2
3
3
2
3
3
4
4
2
3
4
4
c
o
s
c
o
s
c
o
s
b
u
s
t
b
o
b
u
s
t
bu
G
M
M
g
a
c
M
g
r
G
M
g
a
c
M
g
r
G
M
g
r
=
+
+
+
=
+
+
=+
2
.
2.
R
obust
adaptive
contr
oller
desig
n
Def
i
ne
the
er
ror
sig
nal:
=
−
(
4)
wh
e
re
is de
sir
ed value
of
.
Th
e filt
ered er
ror sur
face is c
hos
en
as
=
−
+
=
−
+
d
d
xe
xe
(
5
)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2089
-
4856
I
nt
J
R
ob
&
A
uto
m
,
Vo
l.
8
, No
.
4
,
Decem
ber
201
9
:
293
–
300
296
wh
e
re
1
2
3
(
,
,
)
0
d
i
a
g
=
Ba
sed on
(
5),
(
1) can
b
e
writ
te
n
as
(
)
(
)
(
)
(
)
(
)
(
)
(
)
,,
,
L
T
mm
D
x
C
x
M
y
C
y
G
B
F
Cx
=
−
−
−
−
−
−
+
=
−
−
+
(6
)
wh
e
re
d
ye
=
−
,
(
)
(
)
(
)
(
)
,
1
1
1
,
T
T
m
L
m
D
C
G
F
y
y
B
=
=
. A
s
umi
ng
t
hat
(
)
(
)
(
)
(
)
1
2
3
5
46
,
T
T
m
L
D
C
B
G
F
=
+
an
d
i
(
i
= 1
,
2,3,4,
6
)
are
unko
wn posit
ive c
onsta
ns
.
Con
si
der the
f
ollow
i
ng the
ore
m
.
Theorem
:
If
t
her
e
ex
ist
t
he
scala
rs
i
(
i
=
1
,
2,3,4,
6
)
s
o
t
hat
(
)
(
)
(
)
(
)
1
2
3
5
46
,
T
T
m
L
D
C
B
G
F
=
+
then
the
fo
ll
owin
g
c
ontrolle
r
a
nd a
da
pt
at
ion
law
can
m
ake th
e
dyna
m
ic
err
or
go to
zer
o.
2
1
˙
2
'
6
2
ˆ
,
,
,
6
ˆ
1
ˆ
kk
k
kk
kk
k
k
k
kk
x
Kx
x
k
x
x
=
=
−
−
+
=
−
=
+
(7)
wh
e
re
K
=
dia
g(
k
1
,
k
2
,
k
3
)
w
it
h
k
i
>
0
(
i
=
1,
2,
3),
'
0
,
0
kk
are
const
ant,
=
diag
(
1
,
2
,
3
)
>
0
,
11
T
=
Pro
of:
Ch
oose
the Lya
punove
fun
ct
io
n:
(
)
2
1
6
11
22
T
k
k
k
V
x
D
x
=
=+
(8)
wh
e
re
ˆ
k
k
k
=−
,
ˆ
k
are esti
m
at
ed
values
of
k
,
k
is
po
sit
i
ve
c
onsta
nt.
The
ti
m
e d
eriv
at
ive of Lya
punov f
un
ct
io
n u
sing (
6)
is as
(
)
(
)
(
)
(
)
(
)
(
)
(
)
˙
1
1
˙
1
6
6
11
,
[
,
2
]
1
ˆ
TT
kk
k
k
TT
m
m
k
L
k
k
k
V
x
C
D
x
x
D
y
C
y
G
B
F
x
=
=
=
−
−
+
+
−
−
−
−
−
+
=
−
+
−
+
(9)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
R
ob
&
A
uto
m
IS
S
N:
20
89
-
4856
Ro
bu
st
adapti
ve co
nroll
er d
es
ign
f
or
… (
Ng
a Thi
-
T
huy V
u)
297
Substi
tuti
ng (7
)
int
o
(
9), it i
s
ob
ta
ine
d:
2
2
6
6
6
66
66
6
2
1
1
1
2
2
11
2
2
1
1
6
1
1
ˆ
ˆ
ˆ
ˆ
kk
T
T
T
T
k
k
k
m
m
k
k
k
k
k
k
k
k
k
k
kk
TT
k
m
m
k
k
kk
k
k
k
kk
TT
k
kk
kk
k
k
k
T
k
kk
ik
k
x
x
V
x
K
x
x
x
xx
x
x
K
x
x
x
x
x
K
x
x
x
x
K
x
=
=
=
==
==
==
=
−
−
−
−
+
++
−
+
−
+
+
−
+
−
+
+
−
++
(
)
2
2
11
2
6
2
1
6
6
()
22
2
ˆ
kk
T
k
k
k
k
k
k
ki
kk
k
m
in
k
k
k
x
K
x
Kx
V
==
=
−
−
+
+
−
−
+
−
+
(
10
)
wh
e
re
(
)
(
)
(
)
'
m
in
[
,
/
2
m
a
x
,
1
/
]
/
[
]
0
m
in
k
k
k
KD
=
´
2
k
1
2
3
4
5
k
6
6
1
,
(
)
2
T
k
kk
k
=
=
=
+
(11)
Mult
iply
ing
(1
0) b
y
t
e
gi
ves:
(
)
tt
d
Ve
e
dt
(
12
)
In
te
gr
at
in
g (
12
)
le
ads
to
t
he f
ollow
i
ng ine
qual
it
y:
0
(
)
(
0
)
(
0
)
t
V
t
V
e
V
−
−
+
+
(1
3
)
Ba
sed on ab
ov
e
res
ults an
d
B
arb
al
at
’s
lem
ma, all
erro
r
sig
na
ls wil
l g
o
to
z
ero w
he
n
ti
m
e
go
e
s to
infin
it
e
-
ti
m
e.
3.
SIMULATI
O
N RESULTS
To
e
valuate
t
he
correct
ne
ss
a
nd
su
it
a
bili
ty
of
the
pro
posed
rob
us
t
ada
ptiv
e
co
ntr
oller,
th
e
al
gorithm
was
set
a
nd si
m
ula
te
d
in
Sim
ulink
s
of
twa
re
with a
n
e
xcava
tor
syst
em
g
ive
n param
et
ers
as the
fo
ll
owin
g [
1
8
]:
Mbo
=156
6;
M
st=73
5;
Mbu
=432;
M
loa
d
=5
00;
Ib
o=1425
0.6;
I
st=72
7.7;
Ib
u=224.6;
a2=5.
16; a3=
2.59;
r
2=2.7
1; r3=0.6
4; r
4=0.6
5;
Bbo
=0.02; Bst
=0.02; Bb
u
=0.
02;
The
pa
ram
et
er
f
or
the
co
ntr
ol
l
aw
is
ch
os
e
n
by
pr
act
ic
al
m
et
ho
d
(trial
–
and
–
e
rror):
=
di
ag
(1
),
K
=
dia
g
(2
x1
0
6
, 1.5
x1
0
6
, 10
5
),
k
=
1,
'
k
=
0.
01
,
k
=
0.1
W
it
h
the
pur
pose
of
c
om
par
ison,
the
sim
ul
at
ion
is
exe
rcut
ed
f
or
bo
t
h
pr
opos
e
d
c
on
tr
ol
le
r
an
d
PD
con
t
ro
ll
er
[
15
]
.
The
sim
ulati
on h
as
b
ee
n
c
onduct
ed f
or
t
hr
ee
cases
:
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2089
-
4856
I
nt
J
R
ob
&
A
uto
m
,
Vo
l.
8
, No
.
4
,
Decem
ber
201
9
:
293
–
300
298
-
Ca
se
1
:
Th
e
p
a
ram
et
ers
of
sys
tem
are
rated
, t
he
m
achine wo
rk
s
w
it
ho
ut p
a
yl
oad
.
-
Ca
se
2
:
Th
e
p
a
ram
et
ers
of
sys
tem
are
rated
, t
he
m
achine wo
rk
s
w
it
h f
ull
pa
yl
oad
.
-
Ca
se
3
:
Th
e
p
a
ram
et
ers
of
m
od
el
ch
a
nge, t
he
m
a
chine
wor
ks wil
l f
ull payl
oad.
In
eac
h
case,
t
he
res
ults
were
com
par
ed
wi
th
the
res
ponse
of
P
D
co
ntr
oller
w
hich
w
as
m
entione
d
in
[18].
Sim
ul
at
ion
res
ults
f
or
Ca
se
1,
Ca
se
2,
a
nd
Ca
se
3
are
s
how
n
i
n
Fig
ure
2,
3
an
d
4,
r
especti
ve
ly
.
I
n
these fi
gures,
t
he
s
olid line
(T
heta
d
)
re
presen
ts t
he
re
fer
e
nce
v
al
ue o
f
a
nd
das
hed
li
ne
(
T
heta) is
for
real
v
al
ue
of
.
Fig
ure
2 Sy
ste
m
r
espo
nse
s
unde
r
c
onditi
on
of w
it
ho
ut p
ay
l
oad an
d rate
d s
yst
e
m
p
aram
eter
s
Fig
ure
3. Syst
em
r
espo
nse
s
unde
r
c
onditi
on
of full
p
ay
loa
d and rate
d
syst
e
m
p
ara
m
et
ers
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
R
ob
&
A
uto
m
IS
S
N:
20
89
-
4856
Ro
bu
st
adapti
ve co
nroll
er d
es
ign
f
or
… (
Ng
a Thi
-
T
huy V
u)
299
Fig
ure
4. Syst
em
r
espo
nse
s
unde
r
c
onditi
on
of full
p
ay
loa
d and sy
ste
m
p
ara
m
et
ers
var
ia
ti
on
In
F
ig
ure
2,
w
hen
syst
em
wo
r
ks
with
ou
t
load
a
nd
the
pa
ram
et
ers
of
m
od
el
is
r
at
ed
,
the
posit
io
n
respo
ns
e
of
joi
nts
by
us
i
ng
pr
opos
e
d
co
ntr
ol
le
r
and
PD
c
on
trolle
r
is
abs
ol
utely
tracked
t
o
desi
red
valu
e,
an
d
the
tracki
ng
e
r
ror
is
trivia
l.
I
n
the
sec
ond
c
ase
(
F
ig
ur
e
3),
the
syst
e
m
par
am
et
ers
re
m
ain
unc
hange
d
but
the
excav
at
or
w
or
k
wit
h
fu
ll
pa
yl
oad
.
The
a
da
ptive
c
ontroll
er
giv
es
the
posit
ion
res
ponse
of
joints
al
m
os
t
no
dev
ia
ti
on
from
desire
d
tra
j
ec
tory,
w
hile
th
e
PD
c
on
tr
oller
giv
es
a
trac
king
e
rror
ab
out
0.0
5
ra
d.
I
n
case
syst
e
m
wo
r
ks
with
f
ull
payl
oa
d
a
nd
t
he
para
m
et
er
of
syst
e
m
ch
an
ge
(
F
igure
4),
the
re
su
lt
s
obta
ine
d
for
the
adap
ti
ve
c
on
tr
oller are
sti
ll
r
e
la
ti
vely
g
ood.
The PD
contr
ol
le
r
giv
e
s
a m
axim
a
l t
rack
in
g err
or ab
out 0
.15 rad
.
Fr
om
the
sim
ulati
on
res
ults,
i
t
can
be
see
n
t
hat
the
PD
co
ntr
oller
can
m
ake
the
syst
em
only
w
ork
w
el
l
w
hen
the
el
e
m
ents
of
syst
e
m
are
deter
m
ined.
When
the
syst
em
has
load
distu
r
bance,
the
tracki
ng
e
rror
app
ea
rs
(
0.05
rad)
a
nd
will
increase
to
0.1
5
rad
if
t
he
sy
stem
is
aff
ect
e
d
by
a
dd
it
io
n
disturba
nces.
This
is
pro
ve
d
that
t
he
PD
c
on
t
ro
ll
er
will
increas
e
tracki
ng
er
r
or
w
he
n
syst
e
m
has
distu
rbance.
Me
an
whil
e,
for
adap
ti
ve
c
ontr
oller,
the
res
po
ns
es
of
syst
em
un
der
c
onditi
on
of
syst
em
uncertai
nties
as
well
as
payl
oa
d
noise
are s
o goo
d.
4.
CONCL
US
I
O
N
This
pap
e
r has
pr
ese
nted
the
r
obus
t a
dap
ti
ve
con
t
ro
ll
er
t
hat
do
e
s not
dep
e
nd on t
he
syst
e
m
m
od
el
f
or
excav
at
or
arm
.
The
c
on
t
ro
ll
er
has
sim
ple
structu
re,
easy
to
i
m
ple
m
ent
bu
t
sti
ll
gu
ara
ntees
go
od
pe
r
form
ance
s
to
the
un
ce
rta
inty
of
the
sy
stem
.
The
sta
bili
ty
and
su
it
abili
ty
of
the
co
ntr
oller
we
re
dem
on
strat
ed
by
Ly
apun
ov
sta
bi
li
t
y
theo
ry
a
nd
exam
ined
t
hro
ugh
sim
ulatio
n.
The
sim
ulati
on
res
ults
s
how
that
al
l
j
oi
nts
of
excav
at
or
arm
abs
olu
te
ly
tra
ck
t
o
desire
d
value
eve
n
if
there
is
t
he
e
ffec
t
of
t
he
l
oad
distu
r
ban
ce
a
nd
the
i
m
pact o
f
t
he u
ncer
ta
inti
es
of
syst
e
m
m
od
el
.
REFERE
NCE
S
[1]
S
.
Kum
ar,
K
.
Rani
,
and
V.
K.
Bang
a
,
“
Robo
ti
c
a
rm
m
ovement
O
pti
m
izat
io
n
u
sing
S
oft
C
om
puti
ng
,”
I
AES
Inte
rnational
Jo
urnal
of Robot
i
c
s and
Aut
omat
io
n
,
vol
.
6
,
no
.
1
,
2
017.
[2]
A.
A.
Jabe
r
and
R.
B
ic
ker
,
“
W
ire
le
ss
Fau
lt
Det
e
ct
ion
S
y
st
em
for
an
Industri
al
R
obot
B
ase
d
on
S
ta
ti
st
ical
C
o
ntrol
C
har
t
,”
In
te
rnati
onal
Journal
of
El
e
ct
rica
l
and
C
omputer
Engi
n
e
ering
,
vo
l. 7, no. 6, pp. 3421
-
342
5,
2017
.
[3]
H
.
J
.
Azh
ar,
F
.
H
ad
ar
y
,
and
S
.
S
y
ai
fu
rra
hm
an
,
“
Designing
of
R
obot
G
a
m
el
an
M
usi
c
u
sing
ATm
ega
16
M
ic
roc
ontro
ll
er
,
”
IA
ES
In
te
rnati
onal
Journal
of
Robot
ic
s
and
Au
tomati
on
,
vol
.
6
,
no.
2,
2017.
[4]
H.
Yu,
Y.
Li
u
,
a
nd
M.
S.
Hasan
,
“
Revi
ew
of
M
odel
li
ng
and
Remo
te
Con
trol
for
E
x
ca
va
tors
,
”
Int
ernati
onal
Journal
of
Ad
vanc
ed
Me
chat
ronic
System
s
,
vol. 2, no. 1
–
2,
pp
.
68
–
80
,
2
009.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2089
-
4856
I
nt
J
R
ob
&
A
uto
m
,
Vo
l.
8
, No
.
4
,
Decem
ber
201
9
:
293
–
300
300
[5]
A.
J.
Koivo,
“
Kinemati
cs
of
exc
av
at
ors
(b
ackhoes)
for
tr
ansfe
rring
surfa
ce
m
at
eri
a
l,”
Journ
al
of
Ae
ros
pace
Engi
ne
ering
,
vol
.
7
,
no
.
1
,
pp
.
17
–
32,
1994
.
[6]
P.
K.
Vaha
and
M.
J.
Skibni
ewski,
“
D
y
n
amic
m
odel
of
ex
ca
va
to
r,
”
Journal
o
f
A
erospace
Eng
ineering
,
vo
l.
6,
no.
2,
pp
.
148
–
158
,
1993
.
[7]
A.
J.
Koivo
,
M.
Thoma,
E.
Koca
ogla
n,
and
J.
An
dra
de
-
Ce
tt
o,
“
Modell
ing
and
con
trol
of
ex
ca
va
tor
d
y
namics
dur
in
g
diggi
ng
op
era
t
io
n,
”
Journal
o
f A
erospace
Eng
ineering
,
vo
l. 9, no. 1, pp. 10
–
18,
1
9
96
.
[8]
S.
T
afazol
i
,
P.
D
.
La
wren
ce,
and
S.
E
.
Sal
cude
an
,
“
Ide
nt
ifi
c
at
ion
of
I
ne
rtial
and
F
ric
ti
on
Par
ameters
for
Excava
tor
Arm
s,”
IEE
E
Tr
ansacti
ons on R
oboti
cs
and
Aut
o
mation
,
vo
l. 15,
no.
5
,
pp
.
966
–
9
71,
1999
.
[9]
Y.
H.
Zwe
i
ri,
“
Ide
nti
ficati
on
S
che
m
es
for
Unm
anne
d
Excav
at
or
Arm
Para
m
et
ers,
”
In
te
rn
ati
onal
Journal
of
Aut
omation
and
Computing
,
vo
l.
5,
no
.
2
,
pp
.
185
–
192,
2008
.
[10]
C.
P
.
Ta
n
,
Y.
H.
Zwe
iri,
K.
Alth
oef
er,
and
L
.
D.
Senevi
ra
tne
,
“
O
nli
ne
Soil
Par
amete
r
E
stim
at
ion
Scheme
b
ase
d
o
n
Newton
-
Raphson
M
et
hod
for
Au
tonomous
Exc
av
at
ion
,
”
IEEE/A
S
ME
Tr
ansacti
on
s
on
Me
chat
roni
cs
,
vol
.
10,
no.
2
,
pp.
221
–
229
,
20
05.
[11]
S.
Ta
f
az
o
li
,
S.
E.
Sa
lc
ud
ea
n
,
K.
Hashtrudi
-
Zaad,
and
P.
D.
La
wrenc
e,
“
Im
peda
nc
e
Con
trol
of
a
T
e
le
ope
rated
E
xca
v
at
or
,
”
IEEE
Tr
ansacti
ons
on
Control
S
ystem
s Tec
hnology
,
vol.
10
,
no
.
3
,
pp
.
355
–
367
,
2002
.
[12]
Z.
Lu
,
A.
A.
Goldenbe
rg,
“
R
obust
Im
peda
nc
e
Cont
rol
and
Force
R
egul
a
ti
o
n
:
Th
eor
y
and
E
xper
iment
,
”
T
he
Inte
rnational
Jo
urnal
of Robot
i
c
s R
ese
arch
,
vol
.
14,
no
.
3
,
pp
.
22
5
–
254,
1995
.
[13]
Q.
P.
Ha
,
Q.
H
.
Ngu
y
en
,
D.
C.
R
y
e
,
and
H.
F.
Durrant
-
W
h
y
t
e,
“
Im
peda
nce
C
o
ntrol
of
a
H
y
dra
uli
c
al
l
y
-
Ac
tua
t
e
d
Roboti
c
Excava
t
or
,”
Aut
omat
ion in
Constr
uct
ion
,
vol.
9
,
no
.
5
–
6
,
p
p.
421
–
435
,
200
0
.
[14]
P.
-
H.
Yang,
S.
-
P.
W
ang,
J
.
-
S.
Chia
ng,
and
C
.
-
H.
W
ang,
“
The
Adapti
ve
Contr
ol
Appli
ed
to
th
e
A
naly
sis
o
f
th
e
Exc
av
at
or
D
y
n
a
m
ic
s
,
” IE
EE Co
nfe
renc
e
on
Inno
vat
i
ve
Comput
in
g
Information
,
2
008
.
[15]
H.
M.
Yous
if,
and
K.
Gan
eshInte
rva
l
,
“
T
y
pe
-
2
Fuzz
y
Pos
i
ti
on
Control
of
E
lectr
o
-
H
y
dra
ul
ic
Actua
t
ed
Robo
tic
E
xca
v
a
tor
,”
Int.
J.
M
in. Sci.
Tech
nol
.
,
vol
.
22
,
pp
.
437
-
445
,
2012
.
[16]
D.
L.
Hanh
,
K.
K.
Ahn,
N.
B.
Kha,
and
W.
K
.
Jo
,
“
Traje
ct
or
y
C
o
ntrol
of
El
e
ct
ro
-
H
y
dra
ul
ic
Ex
ca
v
at
or
using
Fuz
z
y
Self
Tun
ing
Alg
orit
hm
w
it
h
Neu
ral
Ne
twork
”
.
J
.
Me
ch. Sci. Tec
h
nol
.
,
vol
.
23
,
pp
.
149
-
160
,
2009
.
[17]
H
.
Feng,
Ch
.
-
B
.
Yin,
W
.
-
W.
W
e
ng,
W
.
Ma
,
J
.
-
J
.
Zhou,
W
.
-
H.
Jia,
and
Z
.
-
L.
Zha
n
g,
“
Roboti
c
Excava
tor
Tr
aje
ct
o
r
y
C
ontrol
using
an
Im
prove
d
GA
base
d
PID
C
ontro
ll
er
,
”
M
ec
han
ic
a
l
Syste
ms
and
Si
gnal
Proce
ss
ing
,
vol.
105
,
no.
5,
pp.
153
-
168
,
20
18.
[18]
Y.
Li
u
and
M.
S.
Hasan
,
“
Modell
ing
and
Remote
Control
of
an
E
x
ca
v
at
or
,
”
Int
ernati
onal
journ
al
o
f
automati
on
and
computi
ng
,
vol.
7
,
no
.
3
,
pp
.
349
-
358,
2010
.
BIOGR
AP
H
I
ES
OF
A
UTH
ORS
Nga
Th
i
-
Th
uy
Vu
r
ecei
ved
th
e
B
.
S.
and
M
.
S.
degr
ee
s
in
el
e
c
tri
c
al
eng
ine
e
rin
g
from
H
anoi
Univer
sit
y
of
Sc
ie
nc
e
and
Techn
olog
y
,
Hano
i,
Vi
et
nam,
in
2005
a
nd
2008,
r
espe
ctively
,
and
the
Ph.D.
degr
ee
in
el
e
ct
roni
cs
and
el
e
ct
ri
ca
l
eng
ineeri
ng
from
Don
gguk
Univer
sit
y
,
Seoul,
Ko
rea,
in
2013.
She
is
cur
ren
tly
wi
th
the
Dep
art
m
en
t
of
Autom
at
i
c
Control
,
Hanoi
Univer
sit
y
of
Scie
nc
e
and
Technol
og
y
,
H
ano
i,
Vie
tna
m
,
as
a
Full
L
ecture
r
.
Her
rese
ar
ch
i
nte
rests
in
cl
ude
DSP
-
base
d
el
ectric
m
a
chi
ne
d
rive
s
and
cont
r
ol
of
distri
bu
ted
gene
r
at
ion
s
y
stems
using
ren
ewa
bl
e ene
rg
y
sour
ce
s.
Evaluation Warning : The document was created with Spire.PDF for Python.