Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
8
,
No.
6
,
D
ece
m
ber
201
8,
pp.
5
09
8~51
06
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v8
i
6
.
pp5098
-
51
06
5098
Journ
al h
om
e
page
:
http:
//
ia
es
core
.c
om/
journa
ls
/i
ndex.
ph
p/IJECE
A Multi
-
r
ob
ot
S
ystem
C
oordina
tion Desi
gn
and An
alys
is
o
n
Wall F
oll
ower R
obot G
roup
Agun
g Nu
groho J
at
i
,
R
an
d
y
Er
f
a Sa
putr
a
,
M. Gh
oz
y Nurc
ahyadi
, Nas
y’a
n Taufi
q A
l
Ghif
ari
Depa
rtment
o
f
C
om
pute
r
Engi
n
e
eri
ng,
School
of C
om
pute
r
Engi
n
ee
ring
Te
lkom Unive
rsi
t
y
,
Indon
esia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Feb
12
, 201
8
Re
vised
Jun
2
8
, 201
8
Accepte
d
J
ul
24
, 2
01
8
In
thi
s
rese
arc
h
,
m
ult
i
-
robot
form
at
ion
ca
n
be
e
stabl
ished
a
cc
or
ding
to
th
e
envi
ronm
ent
or
works
pac
e.
Gro
up
of
robots
will
m
ove
seque
ntly
if
th
ere
is
no
spac
e
for
rob
ots
to
stand
sid
e
b
y
sid
e.
Leade
r
r
obot
will
b
e
on
t
he
front
of
al
l
robots
and
foll
ow
the
right
wall.
On
the
othe
r
hand,
robots
will
m
ove
side
b
y
sid
e
if
th
ere
i
s
a
l
arg
e
sp
ac
e
bet
wee
n
the
m
.
Le
ad
er
robot
wi
ll
be
tr
ac
k
ed
the
wall
on
it
s
ri
ght
side
and
follow
on
it
while
eve
r
y
fol
lower
m
oves
side
b
y
side.
Th
e
l
ea
d
er
robot
hav
e
to
b
roa
dca
st
the
inf
orm
at
ion
to
al
l
r
obots
in
the
group
in
r
adi
us
9
m
et
ers.
Ne
v
ert
he
le
ss
,
eve
r
y
robot
should
be
re
ce
iv
ed
informati
on
fro
m
le
ade
r
robo
t
to
def
in
e
th
ei
r
m
ovements
in
the
area.
Th
e
err
or
provide
d
b
y
fuz
z
y
output
proc
ess
which
i
s
ca
used
b
y
re
a
d
dat
a
from
ult
rasound
sensor
will
drive
to
m
ore
ti
m
e
proc
ess.
More
sam
pli
ng
ca
n
red
uc
e
the
err
or
but
it
will
drive
m
ore
exe
cution
ti
m
e
.
Furthermore,
c
oordina
ti
on
ti
m
e
will
n
ee
d
l
onger
ti
m
e
and
del
a
y
.
Form
at
io
n
will
not
b
e
est
abl
iseh
ed
if
pac
ke
t
e
rror
ha
ppene
d
in
th
e
comm
unic
at
ion
proc
ess
because
robot
wi
l
l
exe
cu
te wrong
c
om
m
and.
Ke
yw
or
d:
Coordi
nation S
chem
e
Fu
zzy
L
ogic
Con
t
ro
l
Mob
il
e Ro
bot
Mult
i
-
Robot S
yst
e
m
Copyright
©
201
8
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
:
Agu
ng Nu
groho
Jat
i
,
Dep
a
rtm
ent o
f C
om
pu
te
r
E
ng
i
neer
i
ng,
Sc
hool
o
f
Elec
tric
al
En
gin
eeri
ng,
Tel
ko
m
Un
i
versi
ty
,
Jl. Tele
kom
un
ikasi
No.
1
Te
r
us
a
n
Bua
h
Ba
t
u,
Ba
ndun
g,
W
est
Jav
a,
40
257
,
I
ndonesi
a
.
Em
a
il
:
agu
ngnj
@tel
ko
m
un
ive
rsity
.ac.id
1.
INTROD
U
CTION
Mostl
y,
resear
ch
in
rob
ot
ap
plica
ti
on
s
is
f
oc
us
e
d
on
hu
m
an
w
ork
as
sist
ance
wet
her
it
’s
co
ntr
olled
m
anu
al
ly
or
m
ov
e
auto
no
m
ou
sly
.
T
o
s
up
port
that
j
obs,
there
a
re
s
o
m
any
com
plex
pro
blem
s
need
to
be
so
lve
d,
suc
h
a
s
co
ntr
ol
m
echan
ism
,
AI
a
nd
decisi
on
m
aki
ng
syst
em
,
path
plan
ning
a
nd
m
ob
il
e
na
vi
gation
syst
e
m
.
Ho
we
ver,
m
os
t
of
th
e
m
are
fo
c
us
e
d
on
sin
gle
r
obot
beh
a
viou
r
on
ly
.
S
o,
i
n
ou
r
cu
rr
e
nt
resea
rch
we
trie
d
to
a
pp
ly
desig
n
a
nd
a
na
ly
ze
prob
le
m
in
m
ulti
-
robo
t
.
Cu
rr
e
ntly
,
w
e
f
ocu
s
on
ho
w
rob
ot
com
m
un
ic
at
e
each
othe
r
in
order
to
s
hare
retrieve
d
sen
s
or
data
to
m
ake
a
coo
r
din
a
te
d
m
ov
e
m
ent.
In
the
wide
r
as
pects,
rob
ots can m
ov
e in
unif
or
m
m
ov
e
m
ent an
d ac
com
plish task
m
or
e e
ff
ic
ie
nt.
Mult
i
-
rob
ot
co
ordinati
on
is
pur
posed
t
o
m
a
ke
r
obots
ca
n
sh
are
a
ny
in
form
ation
bet
we
en
them
.
Fo
r
exam
ple,
a
rob
ot
posit
ion
ca
n
be
s
har
e
d
to
oth
e
rs
in
order
to
de
fine
m
ore
pr
eci
se
oth
e
r
rob
ot
posit
ion
and
avo
i
d
colli
sion.
In
the
ot
her
ha
nd,
a
robo
t
ca
n
fin
d
eff
ic
ie
nt
ro
ute
or
path
diff
e
re
nt
from
each
oth
e
r.
S
o,
it
can
wide
n
ope
rati
on
al
a
rea
of
r
obots
e
nv
ir
on
m
ent.
Be
sides,
it
can
be
us
e
d
in
rob
ot
ap
plica
ti
on
s
wh
i
ch
ne
ed
form
ation
, a
s e
xam
ple in robo
t soccer
team
.
Robot
abili
ty
us
ua
ll
y
has
li
m
it
a
ti
on
de
pe
nds
on
it
s
pro
gr
a
m
.
In
s
om
e
c
ases,
a
gr
oup
of
rob
ots
can
finish
ta
sk
fa
ste
r
tha
n
a
si
ng
l
e
r
obot.
Mo
reover
,
by
us
in
g
m
ul
ti
-
ro
bot,
it
will
wide
n
t
he
w
orkin
g
area
wethe
r
for
sea
rch
i
ng,
m
on
it
or
ing,
or
oth
e
r
j
obs
.
D
espite
to
a
pply
the
m
ulti
-
ro
bot
syst
em
,
there
are
m
any
aspects
wh
ic
h
ha
ve
to
be
co
ns
ide
red.
Ma
in
prob
le
m
from
tho
se
is
ho
w
to
de
sig
n
and
im
ple
m
ent
an
intel
li
gen
t
syst
e
m
wh
ic
h
can
de
fine
sim
ultanou
s
pa
t
h
by
c
omm
un
ic
at
ion
and
c
oor
din
a
ti
on
syst
em
b
ased
on
eac
h
robot
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
& C
om
p
Eng
IS
S
N: 20
88
-
8708
A
Multi
-
r
obot
System
Co
ordina
ti
on
Desi
gn
and An
alysis
on W
all Foll
ower Rob
ot
…
(
A
gung N
ug
r
oho
Jati
)
5099
inf
or
m
at
ion
.
F
ur
t
her
m
or
e,
a
gro
up
of
a
uton
om
ou
s
r
obots
has
to
a
vo
i
d
c
olli
sion
f
ro
m
each
ot
her
by
de
fining
each
path
base
d
on
sh
a
red
da
ta
[1
]
.
It
is
dif
f
eren
t
to
s
om
e
oth
e
rs
re
searc
h
w
hich
on
ly
a
ppli
ed
a
si
ng
le
r
obot,
especial
ly
in
def
i
ning
a
pat
h.
F
or
exam
ple,
in
T
at
i
y
a
Pa
dang
Tunggal
e
t
al
.
t
hey
only
app
li
ed
f
uzzy
cel
l
-
deco
m
po
sit
io
n t
o
de
fine
a
path
in
a si
ng
le
robo
t
[
2]
.
Mult
i
-
rob
ot
syst
e
m
m
us
t
be
desig
ne
d
to
ha
ve
abili
ty
in
colle
ct
ing
an
d
i
n
te
grat
ing
data
from
ro
bo
ts
wethe
r
it
’s
un
i
form
or
no
t
[
3]
.
W
e
ca
n
ex
plo
re
f
r
om
ants
wh
ic
h
li
ve
in
colo
ny.
They
ha
ve
a
kind
of
s
yst
e
m
wh
e
n
t
hey
tra
ve
l,
an
ant
will
le
ave
am
m
on
ia
to
ease
oth
e
r
a
nts
f
ollo
w
t
he
path
[
4].
Ba
se
d
on
that
m
ov
e
m
ent,
ant
colo
ny
m
o
st
ly
si
m
i
la
r
to
le
ader
-
f
ollow
e
r
m
echan
is
m
i
n
m
ult
i
-
robo
t
.
It’s
im
po
rtant
of
a
gro
up
r
obot
to
hav
e
a lea
der
t
o
ease
tha
data
acqu
isi
ti
on sys
tem
an
d
s
olv
e
the
j
obs
[5].
Re
search
in
M
ulti
-
Robot,
m
os
tl
y
dev
el
ops
c
on
t
ro
l
syst
em
base
d
on
com
pu
te
r
process
[
6].
Be
si
des
,
there
are
s
ome
kin
d
of
m
ul
ti
-
robo
t
aut
onom
ou
s
co
ntro
l
syst
e
m
,
wh
ic
h
rob
ot
will
ov
e
a
utono
m
os
afte
r
act
ivate
d
with
ou
t
a
ny
oth
e
r
com
m
and
s
[
7].
Moreove
r,dist
rib
uted
co
ntr
ol
syst
e
m
is
on
e
of
po
pu
la
r
res
earch
area
r
el
at
ed
to
m
ul
ti
-
ro
bot
c
ontr
olli
ng
.
O
ne
of
them
was
f
oc
us
e
d
on
colli
s
ion
a
voida
nce
betwee
n
rob
ots
w
hen
try
ing
to
acc
om
pl
ish
the
gi
ve
n
m
issi
on
[8
]
.
In
a
dva
nce,
rob
ot
can
a
vo
i
d
colli
sio
n
wit
hout
sto
ppin
g
their
m
ov
e
m
ent [
9].
On
the
ot
her
hand,
the
re
w
ere
m
ulti
-
ro
bo
t
ag
or
it
hm
ev
al
uated
t
o
ass
e
m
ble
rob
ots
in
a
sim
il
ar
locat
ion
[
10
]
,
[
11
]
.
It
ca
n
e
va
luate
le
ade
r
-
fol
lower
al
gorith
m
con
te
xt
an
d
al
so
one
kind
of
te
st
to
de
fi
ne
th
e
reab
il
it
y
of
the
m
echan
ism
.
Anothe
r
m
echan
ism
of
that,
fo
ll
ower
rob
ot
will
m
ov
e
by
fo
ll
owin
g
the
le
ader
track
[12].
T
he
n,
it
we
re
de
ve
lop
e
d
by
a
pp
l
yi
ng
distrib
ution
co
ntr
ol
a
nd
inf
or
m
at
ion
sha
rin
g
s
o
th
at
f
ol
lower
rob
ot can m
ov
e m
or
e p
recisel
y t
o
ad
just the
sp
ee
d,
path
, a
nd
or
ie
ntati
on
[
13
]
.
Netw
ork
c
onne
ct
ivit
y i
s also one of
im
po
rta
nt
p
art in m
ulti
-
r
obot. It sh
ould
be
reli
able i
n m
ul
ti
-
ro
bot
to
av
oid
in
f
orm
at
ion
m
issunderta
nd
i
ng
bet
ween
r
obots
be
cause
r
obots
will
al
ways
com
m
un
ic
at
e
du
ri
ng
op
e
rati
on.T
here
we
re
tw
o
ki
nd
of
c
omm
un
ic
at
ion
,
dece
ntrali
zed
a
nd
ce
ntrali
zed
m
et
ho
d.
I
n
dece
ntr
al
iz
e
or
distrib
uted
m
et
hod,
co
nnect
iv
it
y
can
h
a
nd
le
l
arg
e
am
ou
nt
of
r
obots
[
14
]
.
Wh
il
e
in
ce
ntr
al
iz
ed
m
et
ho
d,
it
has
to
be
de
fine
a
rob
ot
as
a
le
a
der
w
hich
will
ha
nd
le
data
from
al
l
of
r
obot
s.
I
n
netw
ork,
it
al
so
sim
il
ar
as
a
router.
E
ver
y
s
ing
le
r
obot
will
m
ov
e
based
on
data
com
e
f
ro
m
le
ader
[15
]
,
[16
]
.
It
is
ef
fecti
ve
an
d
effi
ci
ent
for
sm
al
l a
m
ou
nt of
rob
ots.
In
this
resea
rc
h,
we
will
app
ly
a
s
m
all
gr
oup
of
r
obots
consi
st
of
f
our
wh
ic
h
on
e
robo
t
will
be
def
i
ned as a le
ader. It
will
b
e
an
e
valuati
on t
o determ
ine the r
eabil
it
y of m
ulti
-
rob
ot m
ec
han
ism
w
hic
h m
ai
nly
pur
po
se
d
to
m
ai
ntain
the
for
m
at
ion
of
r
obot
s
with
si
m
ples
t
po
ssible
al
go
rithm
.
It
m
eans
that
the
co
m
pu
ta
ti
on
exp
ect
e
d
is
as
m
il
d
as
possible.
I
n
our
previo
us
researc
h,
it
is
pr
ov
e
d
that
desi
gn
e
d
al
gorithm
ca
n
r
un
pro
per
ly
in
a
rduin
o
based
co
ntr
oller
w
her
e
fo
ll
ower
r
obot
can
m
ov
e
th
r
ough
le
ade
r
tr
ack
[
17]
.
W
e
app
li
ed
le
ss
com
plexit
y
than
ei
ther
s
war
m
syst
e
m
or
local
iz
at
ion
m
et
ho
ds
p
res
e
nted
in [
18]
an
d
[
19
]
in o
r
der
to
m
ake
a quick
res
pond syst
em
in
on
l
y a si
m
ple r
obots.
In
this
pa
pe
r,
it
will
be
pr
esented
the
de
sig
ned
m
echan
is
m
of
com
m
un
i
cat
ion
an
d
ho
w
the
rob
ots
m
ake
a
pr
oper
coor
din
at
io
n
be
tween
them
.
This
pa
per
wil
l
be
arr
an
ge
d
as
fo
ll
ows:
in
sect
ion
1,
it
is
a
lready
pr
ese
nted
a
n
in
tro
du
ct
io
n
of
the
resear
ch
an
d
relat
ed
w
ork
s
of
that;
wh
il
e
th
e
syst
e
m
and
m
e
tho
d
will
be
put
on
t
he
sect
ion
2;
in
sect
ion
3,
te
sti
ng
scen
arios,
res
ults,
and
e
valuati
on
of
the
syst
em
are
giv
e
n;
as
a
la
st
sect
ion
,
in
sect
ion
4
will
be
presente
d
a
co
nc
lusio
n
an
d
the
descr
i
ption
of
our
f
uture
w
orks
relat
ed
to
c
urren
t
pu
blica
ti
on.
2.
SY
STE
M DESIGN
Gen
e
rall
y,
syst
e
m
con
sist
s
of
four
unif
orm
ro
bo
ts
whic
h
can
c
omm
un
ic
at
e
each
oth
er
f
or
coor
din
at
io
n.
Each
rob
ot
ha
s
em
bed
ded
pr
oc
esso
r
as
m
ai
n
co
ntr
oller,
tw
o
dc
-
m
oto
rs
f
or
act
uat
or
,
an
d
sens
or
syst
e
m
fo
r
local
iz
at
ion
input
in
def
i
ning
po
sit
io
n
su
c
h
as
ultrasoun
d
and
com
pass
sensor.
Be
sid
es,
for
com
m
un
ic
at
ion
each
of
the
m
has
an
RF
base
d
trancei
ve
r
-
receive
r
w
hi
ch
wor
ks
on
f
reque
ncy
channel
433
MHz.
F
ro
m
the
m
,
a
robo
t
ha
s
a
ro
le
as
le
a
der
w
hile
oth
e
rs
will
be
fo
ll
ow
e
r.
Eac
h
R
obot
Hard
war
e
Bl
ock
Syst
e
m
as sho
wn in Fi
gure
1.
Figure
1.
Each
rob
ot h
a
rdwa
re
b
loc
k sy
stem
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
87
08
In
t J
Elec
&
C
om
p
En
g,
V
ol.
8
, N
o.
6
,
Dece
m
ber
2
01
8
:
5098
-
5106
51
00
Ultraso
und
se
nsors
are
ne
ede
d
to
def
ine
d
th
e
posit
ion
on
a
m
aze
base
on
ra
ng
e
bet
wee
n
rob
ot
an
d
fron
t
-
side
wall
s.
Wh
il
e
c
ompass
s
ens
or
wi
ll
determ
ine
t
he
rob
ot
ori
en
ta
ti
on
.
Be
si
des
,
ultras
ond
ca
n
hel
p
rob
ot
to
m
ov
e
arr
ou
nd
a
nd
avo
i
d
colli
sion
wethe
r
with
or
without
data
f
ro
m
oth
er
r
obot.
A
fter
each
rob
ot
def
i
nes
it
s
posi
ti
on
a
nd
or
ie
nt
at
ion
,
t
hey
will
send
t
hat
in
f
or
m
at
ion
to
le
ader
r
obot
in
orde
r
to
be
proc
essed
.
Leader
r
obot
will
co
m
par
e
ever
y
receive
d
inf
or
m
at
ion
with
it
s
ow
n
to
de
te
rm
ine
the
nex
t
ste
p
for
each
rob
ot
m
ov
e
m
ent.
The
com
m
un
ic
at
i
on
sc
hem
e
us
ed
in
this
resear
ch
was
a
broa
dc
ast
or
m
esh
network
s
o
that
ever
y
rob
ot
can
c
omm
un
ic
at
e
directl
y
to
oth
ers
e
v
en
the
decisi
on
will
be
m
ade
by
le
ader
only
.
It
helpe
d
wide
ning
the
ra
nge
beca
us
e
it
was
de
fined
that
eve
ry
fo
ll
ower
r
obot
will
al
ways
resen
d
data
com
e
from
oth
er
fo
ll
ow
e
r
ti
ll
it
r
ecei
ved
by leade
r.
Use
d
C
omm
un
ic
ation
Netw
ork
S
chem
e o
f
Mult
i
-
Robot
as
sho
w
n i
n Fi
gure
2.
Figure
2
.
Use
d C
omm
un
ic
at
i
on N
et
wor
k
Sc
hem
e o
f
Mult
i
-
Robot
Fo
r
com
m
un
ic
at
ion
a
nd
na
vi
gation
sh
a
rin
g
necessit
y,
the
r
e
wer
e
create
d
com
e
pr
eced
ure
an
d
data
form
at
to
be
s
ent
by
r
obots
to
gro
up.
Si
nce
there
are
tw
o
kind
of
f
orm
ation
s
desi
red,
proce
dure
creat
ed
al
s
o
hav
e
d
i
ff
e
ren
c
es b
et
ween t
he
m
.
2
.
1.
Par
alel R
obot
For
ma
tio
n
In
pa
ralel
f
or
m
at
ion
,
f
ollow
e
r
r
obots
will
m
ov
e
f
orwa
rd
by
fo
ll
owin
g
le
ader’s
t
rack
be
hind.
Paralel
form
ation
is
eq
ual
to
se
qu
e
ntial
form
at
ion
w
her
e
r
obots
m
ov
e
i
n
a
strai
ght
li
ne.
This
f
orm
at
ion
is
ver
y
us
ef
ul
wh
e
n r
obots
find a
narrow la
ne.
The
re
wer
e
so
m
e p
r
ocedu
res
to
d
e
fine
th
e p
a
ralel
for
m
a
ti
on
as
foll
ows
:
1.
At
first
,
r
obot
m
ov
e
by
us
ing
wall
fo
ll
ow
i
ng
al
gorithm
wh
ic
h
f
ollo
wing
a
righ
t
wall
ba
sed
on
efi
ned
range
(3
cm
f
r
om
the w
al
l).
2.
If
le
a
der
r
obot
fin
d
a
n
ob
sta
cl
e
in
fro
nt
of
it
,
it
will
sen
d
E
I
M
data
to
ot
he
rs.
The
n,
it
will
tur
n
le
ft
a
nd
m
ov
e forw
a
r
d un
ti
l data
B is
receive
d.
3.
Fo
ll
ower
r
obot
at
the
sec
ond
po
sit
io
n
will
stop
w
he
n
it
rec
ei
ve
an
E
IM
da
ta
for
a
m
o
m
ent
(
delay
set
)
,
and m
ov
e agai
n by wall
fo
ll
owin
g on the
rig
ht w
al
l.
4.
The
sec
ond
r
ob
ot
will
send
I
M
data
to
rob
ot
s
beh
in
d
it
w
he
n
fin
d
a
n
obst
acl
e
in
the
fro
nt
(obstacl
e
ca
n
be
a lea
der al
s
o) an
d
m
ake a t
urn
to
left t
he
fo
ll
ow t
he
ri
ght wall
.
5.
If
t
he
thir
d
rob
ot
fin
d
a
n
obsta
cl
e,
oth
e
rs
wi
ll
be
sto
pp
e
d.
I
t
will
send
dat
a
AEM
to
f
our
th
r
obo
t.
Othe
r
beh
a
viou
r
is e
qual
to
le
a
der n
seco
nd robot.
6.
Fo
r
the
la
st
robo
t,
it
will
send
AE
I
data
if
it
find
an
obsta
cl
e
and
the
m
ak
e
a
le
ft
turn
a
nd
sen
d
BFJ
dat
a
to all
robots.
T
he next
proce
dures
are
b
ac
k
t
o
the
f
ir
st.
To
determ
ine
m
ov
e
m
ent
ste
ps
of
m
ulti
-
ro
bo
t
above,
data
f
or
m
at
wh
ic
h
know
n
by
each
rob
ot
ha
ve
t
o
be
desi
gn
e
d.
T
o
ease
the
proc
edures,
data
form
at
create
d
as
si
m
ple
as
po
ss
ible
so
that
it
on
ly
us
es
a
strin
g
or
char
act
e
r
to
de
fine
com
m
and
s.
H
oweve
r,
com
m
and
s
are
def
i
ned
by
sensor
input
conditi
o
n
ba
sed
on
env
i
ronm
ent.
As
desc
ribe
d
be
fore,
se
ns
or
use
d
co
ns
ist
of
ultraso
und
to
def
i
ne
an
obst
acl
e
and
ra
ng
e
to
the
wall
,
a
nd
com
pass
us
e
to
fi
nd
rob
ot
or
ie
nta
ti
on
.
O
n
t
he
T
able
1,
sho
wn
m
essages
c
reted
t
o
be
se
nt
from
on
e
to o
t
her r
obots.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
& C
om
p
Eng
IS
S
N: 20
88
-
8708
A
Multi
-
r
obot
System
Co
ordina
ti
on
Desi
gn
and An
alysis
on W
all Foll
ower Rob
ot
…
(
A
gung N
ug
r
oho
Jati
)
5101
Table
1.
C
omm
un
ic
at
ion
and
Nav
i
gation P
ro
ce
dure
of M
ulti
-
Robot Pa
ra
le
l For
m
at
ion
ROB
OT
1
ROB
OT
2
ROB
OT
3
ROB
OT
4
MOV
EM
ENT
DATA
MOV
EM
ENT
DATA
MOV
EM
ENT
DATA
MOV
EM
ENT
DATA
RWF
RWF
RWF
RWF
AP4
FUZ
ZY
STOP
SEND
EIM
RECEIV
E
EIM
RECEIV
E
EIM
RECEIV
E
EIM
TURN
LE
F
T
STOP
STOP
STOP
DEL
A
Y
DEL
A
Y
DEL
A
Y
FORWAR
D
RWF
RWF
RWF
STOP
AP4
FUZ
ZY
ROB
OT
1
S
TOP
UNTIL
RECEIV
E
D
ATA
‘B’
STOP
SEND
IM
RECEIV
E
IM
RECEIV
E
IM
TURN
LE
F
T
STOP
STOP
DEL
A
Y
DEL
A
Y
FORWAR
D
RWF
RWF
RECEIV
E
B
SEND
B
AP4
FUZ
ZY
RWF
RWF
STOP
RECEIV
E
AEM
RECEIV
E
AEM
SEND
AEM
RECEIV
E
AEM
STOP
STOP
TURN
LE
F
T
STOP
ROB
OT
1
D
AN 2
STOP
UNT
IL
RE
CEIVE
DA
TA
‘B
F’
DEL
A
Y
FORWAR
D
RWF
RECEIV
E
BF
RECEIV
E
BF
SEND
BF
AP4
FUZ
ZY
RWF
RWF
RWF
STOP
RECEIV
E
AEI
RECEIV
E
AEI
RECEIV
E
AEI
SEND
AEI
STOP
STOP
STOP
TURN
LE
F
T
ROB
OT
1
,
2, DA
N 3 STO
P H
ING
G
A M
ENE
R
IM
A D
ATA
‘BFJ’
FORWAR
D
RECEIV
E
BFJ
RECEIV
E
BFJ
RECEIV
E
BFJ
SEND
BFJ
RWF
RWF
RWF
RWF
2
.
2
.
Seri
al
Robo
t
F
orm
at
io
n
Wh
e
n
r
obots
de
te
ct
and
def
in
e
la
rg
er
area
wh
ic
h
is
fit
to
pu
t
r
obots
toge
ther
in
a
row,
then
Mult
i
-
Robot
syst
e
m
will
be
entered
the
serial
fo
rm
at
ion
m
od
e.
In
this
m
od
e,
r
ob
ots
will
m
ov
e
f
orward
to
gethe
r
in
a
sam
e
ro
w
a
nd
in
eq
ual
s
pee
d.
The
dif
ficult
pro
blem
is
wh
en
rob
ots
fi
nd
ob
sta
cl
e
or
wal
l
in
f
ront
of
th
e
m
,
so
that
they
hav
e
to
m
ake
a
turn.
But,
it
will
b
e
def
ine
d
by
the
le
ader
wh
e
r
e
the
po
sti
on
is
near
est
to
th
e
right
wall
. Pro
ce
dur
es is dete
rm
ine
d
as
foll
ows:
1.
Wh
e
n
le
ade
r
r
obot
detect
s
a
n
obsta
cl
e
or
wa
ll
on
the
f
ront,
it
w
il
l
stop
m
ov
e
at
a
m
o
m
ent
an
d
c
heck
th
e
le
ft
side.
If
the
distance
is
m
or
e
tha
n
10cm
to
the
second
rob
ot,
it
will
s
end
the
GI
M
da
ta
to
com
m
and
the sec
ond rob
ot to st
op. Besi
des, i
t wil
l al
so sen
d
E
KM
dat
a to st
op the
ot
her f
ollow
e
r robo
t.
2.
Soon
a
fter
the se
co
nd
r
ob
ot
stop
s
,
it
will
al
so
check
t
he
le
ft side
an
d
re
peat
the
proce
dure u
se
d
by
le
ade
r
rob
ot.
St
op com
m
and
is d
e
fi
ned as
IO d
at
a.
3.
Me
anwhil
e, ro
bo
t
3 wil
l al
so
check i
ts l
eft si
de
a
nd se
nd M
data to
r
obot
4.
Data
flo
w
f
rom
pr
oce
dure
a
bove
al
so
ca
n
be
sho
wn
by
ta
ble
2
bel
ow
.
Me
anwhil
e,
F
igure
3
s
how
s
the f
l
ow pro
ce
ss and c
omm
u
nicat
ion
bet
we
en
le
ade
r
a
nd
f
ollow
e
r rob
ots.
Table
2
. C
omm
un
ic
at
ion
and
Nav
i
gation P
ro
ce
dure
of M
ulti
-
Robot Se
ri
al
Form
ation
ROB
OT
1
ROB
OT
2
ROB
OT
3
ROB
OT
4
MOV
EM
ENT
DATA
GERKA
N
DATA
MOV
EM
ENT
DATA
MOV
EM
ENT
DATA
RWF
RWF
RWF
RWF
AP4
FUZ
ZY
STOP
SEND
GIM
RECEIV
E
GIM
RECEIV
E
GIM
RECEIV
E
GIM
STOP
RWF
STOP
STOP
LE
FT
SE
NSOR
CHECK <
10
?
SEND
EKM
RECEIV
E
EKM
RECEIV
E
EKM
RECEIV
E
EKM
STOP
STOP
RWF
STOP
CEK
S
ENSOR
KIRI
< 10
?
SEND
IO
RECEIV
E
IO
RECEIV
E
IO
STOP
STOP
RWF
CEK
S
ENSOR
KIRI
< 10
?
SEND
M
RECEIV
E
M
STOP
STOP
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
87
08
In
t J
Elec
&
C
om
p
En
g,
V
ol.
8
, N
o.
6
,
Dece
m
ber
2
01
8
:
5098
-
5106
5102
S
t
a
r
t
R
a
n
g
e
S
e
n
s
o
r
D
e
t
e
c
t
i
o
n
(
F
r
o
n
t
+
R
i
g
h
t
)
F
r
o
n
t
O
b
j
e
c
t
?
F
u
z
z
y
C
o
n
t
r
o
l
P
r
o
c
e
s
s
M
o
v
e
F
o
r
w
a
r
d
(
R
i
g
h
t
W
a
l
l
F
o
l
l
o
w
i
n
g
)
I
n
f
o
r
m
a
t
i
o
n
S
e
n
t
t
o
O
t
h
e
r
R
o
b
o
t
s
S
t
a
r
t
A
n
y
I
n
f
o
R
e
c
e
i
v
e
d
?
T
u
r
n
L
e
f
t
(
9
0
d
e
g
)
M
o
v
e
F
o
r
w
a
r
d
(
R
i
g
h
t
W
a
l
l
F
o
l
l
o
w
i
n
g
)
F
r
o
n
t
O
b
j
e
c
t
?
I
n
f
o
r
m
a
t
i
o
n
S
e
n
t
t
o
L
e
a
d
e
r
A
n
y
R
e
p
l
i
e
s
?
F
u
z
z
y
C
o
n
t
r
o
l
P
r
o
c
e
s
s
R
a
n
g
e
S
e
n
s
o
r
D
e
t
e
c
t
i
o
n
(
F
r
o
n
t
+
R
i
g
h
t
)
E
n
d
E
n
d
(a)
(
b)
Figure
3.
Flo
w
Char
t
of F
or
m
at
ion
Sc
hem
e (a)
Lea
der Ro
bot,
(b)
F
ollo
we
r
Ro
bot
2
.
3
.
M
ov
e
men
t
R
obo
t Contr
ol
Gen
e
rall
y,
rob
ot
m
ov
es
by
grab
bing
on
the
righ
t
wall
or
com
m
on
ly
kn
own
as
rig
ht
wa
ll
fo
ll
ow
in
g
m
et
ho
d.
It
is
a
kind
of
sim
plest
m
et
ho
d
t
o
m
ov
e
a
nd
c
omm
on
ly
us
ed
by
blind
m
obil
e
robo
t
w
hich
on
ly
dep
e
nd
on
range
sens
or.
To
def
i
ne
the
decisi
on
to
m
ov
e,
fu
zzy
log
ic
sc
hem
e
is
designed
us
in
g
tw
o
input
by
us
in
g
range
se
ns
or.
I
t ca
n be
sh
ow
n
on the
F
igure
4
belo
w.
Figure
4. F
uzz
y Lo
gic D
esi
gn
(16)
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
& C
om
p
Eng
IS
S
N: 20
88
-
8708
A
Multi
-
r
obot
System
Co
ordina
ti
on
Desi
gn
and An
alysis
on W
all Foll
ower Rob
ot
…
(
A
gung N
ug
r
oho
Jati
)
5103
F
f
ar
(x)
=
{
1
,
x
≥
18
−
8
18
−
8
,
8
<
<
18
0
,
x
≤
8
(1)
F
near
(x)
=
{
1
,
x
≤
8
18
−
18
−
8
,
8
<
<
18
0
,
x
≥
18
(2)
Fo
rm
ulati
on
s
a
bove
is
def
in
ed
for
f
ront
se
ns
or.
It
is
us
ed
t
o
m
ake
a
decisi
on
to
m
ov
e
or
st
op.
O
n
th
e
oth
e
r
ha
nd,
t
he
rig
ht
sens
or
is
us
e
d
f
or
de
fin
ing
t
he
sp
e
ed
of
m
oto
rs
a
nd
m
ake
a
r
obot
m
ov
ing
by
gra
bb
i
ng
the
rig
ht
wall
.
Fu
rt
her
m
or
e,
i
n
the
fo
ll
owe
r
rob
ot,
it
can
be
us
ed
f
or
detect
ing
the
oth
e
r
ro
bot
on
their
righ
t
side.
All val
ue
s ar
e
de
fine
d
a
s a r
a
nge in
ce
nti
m
et
ers.
R
f
ar
(
x)
=
{
1
,
x
≥
40
−
15
25
−
15
,
20
<
<
40
0
,
x
≤
20
(3)
R
mid
(x)
=
{
1
,
=
15
−
5
15
−
5
,
5
<
<
15
25
−
25
−
15
,
15
<
<
25
0
,
5
≥
≥
25
(4)
R
near
(x)
=
{
1
,
x
≤
5
15
−
15
−
5
,
5
<
<
15
0
,
x
>
15
(5)
Figure
5
ab
ov
e
sh
ows
the
de
fu
zzy
ficat
io
n
resu
lt
of
the
proces
s.
It
giv
e
s
the
m
oto
r
spe
ed
betwee
n
rig
ht an
d
le
ft in
orde
r
to m
ak
e a r
obot m
ov
e fo
r
ward b
y f
ol
lowing the r
i
ght wall
. I
n
the r
e
su
lt
, it still
p
rodu
ce
s
the
osc
il
la
ti
on
w
hich
is
c
use
d
by
di
ff
e
rent
dc
m
oto
rs
pro
blem
.
Def
uz
zy
ficat
ion
pro
cess
is
ba
sed
on
th
e
fo
ll
owin
g form
ula.
Figure
5. F
uzz
y Lo
gic Outp
ut
Surf
ace
ba
sed
on T
wo Se
ns
or
Inputs
∗
=
Σ
(
)
×
(
)
(6)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
87
08
In
t J
Elec
&
C
om
p
En
g,
V
ol.
8
, N
o.
6
,
Dece
m
ber
2
01
8
:
5098
-
5106
5104
3.
RESU
LT
S
A
ND AN
ALYSIS
It
is
i
m
po
rtant
to
te
st
the
design
e
d
syst
em
i
n
orde
r
to
kn
ow
how
reli
able
it
is.
Moreo
ve
r,
it
can
be
us
e
d
to
ve
rify
the
syst
e
m
perform
ance.
The
r
are
s
om
e
par
a
m
et
ers
us
ed
t
o
m
easur
e
the
perform
ance
of
the
syst
e
m
.
First
of
al
l,
w
e
ha
ve
te
st
how
the
com
m
un
ic
at
ion
can
c
ov
e
r
c
oor
din
at
ion
betwee
n
r
obots.
Delay
par
am
et
er
was
us
e
d
t
o
know
how
good
t
he
us
e
d
c
omm
un
i
cat
ion
syst
em
is.
Re
su
lt
s
of
t
he
m
easur
em
e
nt
ca
n
be
seen
belo
w.
Fo
r
i
nfor
m
at
i
on,
data
w
as
s
ent
al
l
fr
om
lead
er
rob
ot.
Ba
sed
on
the
r
es
ult
sh
ow
n
in
T
able
3
,
delay
can
be
m
ini
m
iz
ed
by
us
in
g
tra
ns
fe
r
rate
4000bps
on
the
c
omm
u
nicat
ion
syst
e
m
.
Ho
we
ve
r,
distance
betwee
n
rob
ot
al
so
a
ff
ect
s
t
he
delay
.
Ma
xim
u
m
distance
of
data
tra
ns
fe
r
is
9
m
et
ers
f
or
this
ki
nd
of
transceive
r.
O
n
the
oth
e
r
ha
nd
,
if
we
ap
plie
d
20
00bps
of
t
ran
s
fer
rate,
da
ta
transf
e
r
pr
oc
ess
is
pro
du
c
e
m
or
e
sta
bili
ty
and
the
var
ia
nce
of
de
la
ys
is
le
ss.
At
so
m
e
po
int,
it
can
be
us
ed
m
or
e
tha
n
9
m
e
t
ers.
Re
cei
ve
d
s
ign
al
stren
gth
is
sti
ll
in
the
ra
nge
to
be
acce
pte
d
by
recei
ver.
Re
gar
ding
to
the
e
valuati
on
of
com
m
un
ic
at
ion
syst
e
m
,
we
m
a
ke
a
bo
unda
ry
of
t
he
r
obots
w
orks
pace
in
a
9
m
et
ers
rad
iu
s.
It
is
us
e
d
beca
us
e
the
le
a
der
r
obot
hav
e
to
broad
c
ast
the
info
rm
at
ion
to
al
l
ro
bo
ts
in
the
gr
oup.
Ne
ver
th
el
ess,
ever
y
r
obot
sh
ould
be
re
c
ei
ved
inf
or
m
at
ion
from
le
ader
ro
bo
t
to d
e
fine
their
m
ov
e
m
ents in t
he
area
.
The
n,
it
has
be
en
te
ste
d
how
r
obots
e
xec
ute
their
c
omm
and
f
ro
m
le
ader
rob
ot.
It
is
s
how
n
on
t
he
T
able
4
,
that
e
ver
y
r
obot
in
the
gro
up
m
os
tl
y
execu
te
s
co
m
m
and
accura
te
ly
based
on
r
ecei
ved
data.
It
can
be
con
cl
ud
e
d
that
al
l
of
com
m
an
d
from
le
ader
can
be
receive
d
by
fo
ll
ower
r
obots
with
out
err
or,
so
that
f
ollow
e
r
rob
ots
can
e
xe
cute
them
accuratel
y.
Mult
i
-
rob
ot
f
or
m
at
ion
ca
n
be
est
a
bl
ished
acc
ordi
ng
to
t
he
e
nvir
on
m
ent
or
work
s
pace
.
Group
of
r
obots
will
m
ov
e
seq
uen
tl
y
if
there
is
no
s
pac
e
fo
r
rob
ots
to
sta
nd
side
by
side.
Leader
rob
ot
will
be
on
th
e
fron
t
of
al
l
rob
ots
an
d
f
ollow
the
rig
ht
wall
.
On
the
oth
e
r
ha
nd,
r
obots
wil
l
m
ov
e
side
by
side
if
there
is
a
la
rg
e
sp
ace
betwe
en
them
.
Leader
rob
ot
will
be
t
r
acked
the
wall
on
it
s
rig
ht
sid
e
an
d
fo
ll
ow
on it
wh
il
e ever
y f
ollo
wer m
ov
es sid
e b
y si
de.
Table
3
.
Delay
Testi
ng Result
No
.
Sa
m
p
lin
g
Rate
Tr
an
s
m
itt
er
Ro
b
o
t
1
Ro
b
o
t
2
Ro
b
o
t
3
Sen
t data
Receiv
ed
data
Delay
(s)
Receiv
ed
data
Delay
(s)
Receiv
ed
data
Delay
(s)
1
4
0
0
0
b
p
s
ABC
ABC
0
.15
ABC
0
.15
ABC
0
.15
2
4
0
0
0
b
p
s
ABC
ABC
0
.13
ABC
0
.13
ABC
0
.13
6
4
0
0
0
b
p
s
AB
AB
0
.13
AB
0
.13
AB
0
.13
7
4
0
0
0
b
p
s
AB
AB
0
.14
AB
0
.14
AB
0
.14
11
4
0
0
0
b
p
s
A
A
0
.12
A
0
.12
A
0
.12
12
4
0
0
0
b
p
s
A
A
0
.13
A
0
.13
A
0
.13
16
2
0
0
0
b
p
s
ABC
ABC
0
.19
ABC
0
.19
ABC
0
.19
17
2
0
0
0
b
p
s
ABC
ABC
0
.16
ABC
0
.16
ABC
0
.16
21
2
0
0
0
b
p
s
AB
AB
0
.2
AB
0
.23
AB
0
.23
22
2
0
0
0
b
p
s
AB
AB
0
.22
AB
0
.22
AB
0
.22
26
2
0
0
0
b
p
s
A
A
0
.22
A
0
.22
A
0
.22
27
2
0
0
0
b
p
s
A
A
0
.17
A
0
.17
A
0
.17
31
1
0
0
0
b
p
s
ABC
ABC
0
.7
ABC
0
.49
ABC
0
.49
32
1
0
0
0
b
p
s
ABC
Failed
Pack
et
Los
s
Failed
Pack
et
Los
s
Failed
Pack
et
Los
s
38
1
0
0
0
b
p
s
AB
Failed
Pack
et
Los
s
Failed
Pack
et
Los
s
Failed
Pack
et
Los
s
39
1
0
0
0
b
p
s
AB
Failed
Pack
et
Los
s
Failed
Pack
et
Los
s
Failed
Pack
et
Los
s
41
1
0
0
0
b
p
s
A
A
0
.2
A
0
.2
A
0
.2
42
1
0
0
0
b
p
s
A
A
0
.2
A
0
.2
A
0
.2
Table
4
.
R
obot
Ex
ec
utio
n
Tes
t
Testin
g
Nu
m
b
e
r
-
Ro
b
o
t L
eader
Ro
b
o
t Follo
wer
An
n
o
tatio
n
Sen
d
Data
Sen
t Data
Receiv
e
Data
Receiv
ed
Data
Mov
e as sent co
m
m
a
n
d
1
Yes
Yes
Yes
Yes
Yes
Accuratel
y
2
Yes
Yes
Yes
Yes
Yes
Accuratel
y
3
Yes
Yes
Yes
Yes
Yes
Accuratel
y
4
Yes
Yes
Yes
Yes
Yes
Accuratel
y
5
Yes
Yes
Yes
Yes
Yes
Accuratel
y
6
Yes
Yes
Yes
Yes
Yes
Accuratel
y
7
Yes
Yes
Yes
Yes
Yes
Accuratel
y
8
Yes
Yes
Yes
Yes
Yes
Accuratel
y
9
Yes
Yes
Yes
Yes
Yes
Accuratel
y
10
Yes
Yes
Yes
Yes
Yes
Accuratel
y
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
& C
om
p
Eng
IS
S
N: 20
88
-
8708
A
Multi
-
r
obot
System
Co
ordina
ti
on
Desi
gn
and An
alysis
on W
all Foll
ower Rob
ot
…
(
A
gung N
ug
r
oho
Jati
)
5105
4.
CONCL
US
I
O
N
On
t
he
m
ulti
-
ro
bot
cases,
e
ve
ry
rob
ot
in
th
e
group
ca
n
co
m
m
un
ic
at
e
in
order
t
o
sh
a
re
inf
or
m
at
ion
betwee
n
them
si
m
ultanoeousl
y. It is
purpose
d
to
m
ake a
go
od coor
di
natio
n
f
or e
nv
ir
ontm
ent ex
pl
or
at
i
on or in
a
si
m
plifie
d
case,
na
vig
at
io
n.
In
the
bigger
pur
pose,
a
gro
up
of
rob
ots
can
exp
lo
re
an
d
m
app
e
d
the
unknow
n
env
i
rontm
ent q
uicke
r
tha
n
si
ngle
m
ob
il
e rob
ot.
In
our
case,
m
ulti
-
rob
ot
f
or
m
at
ion
ca
n
be
e
sta
bli
sh
ed
acc
ordin
g
t
o
the
e
nv
i
ronm
ent
or
w
orks
pace.
Group
of
r
obot
s
will
m
ov
e
se
qu
e
ntly
if
there
is
no
sp
ace
f
or
rob
ots
to
sta
nd
side
by
side
.
Leader
r
obot
w
il
l
be
on
the
fron
t
of
al
l ro
bots an
d
f
ollow
the
rig
ht w
al
l. On
the
ot
her
h
a
nd, ro
bots wil
l
m
ov
e si
de
b
y si
de
if th
ere is
a
la
rg
e
s
pace
betwee
n
t
hem
.
Leade
r
rob
ot
will
be
t
rack
e
d
th
e
wall
on
it
s
rig
ht
side
a
nd
f
ollo
w
on
i
t
wh
il
e
ever
y
fo
ll
owe
r m
ov
es side
by
side.
Ba
sed
on
pe
rfor
m
ance
te
sti
ng
,
form
at
io
n
ca
n
be
es
ta
blished
acc
ur
at
el
y
with
out
er
ror
on
com
m
un
ic
at
ion
.
T
he
e
rror
is
pro
vid
e
d
by
fuzzy
ou
t
pu
t
pro
cess
w
hich
is
c
ause
d
by
read
data
f
ro
m
ultrasoun
d
sens
or
.
Mo
re
s
a
m
pling
ca
n
re
du
ce
the
er
ror
bu
t
it
will
dri
ve
m
or
e
exec
ution
ti
m
e.
Fu
rth
erm
or
e,
c
oord
i
natio
n
tim
e w
il
l need
longer
tim
e and
delay
.
Com
m
un
ic
at
io
n
proces
s
bet
ween
le
a
der
a
nd
fo
ll
ower
robo
t
use
d
RF
433M
Hz
trance
iver
m
od
ule
with
rate
4000
bp
s
in
order
t
o
m
ini
m
iz
e
tran
sm
issi
on
delay
.
Ba
sed
on
te
st
,
there
was
no
pack
et
l
os
s
un
t
il
th
e
m
axi
m
u
m
distance
9
m
et
ers.
Packet
l
os
s
or
error
tra
ns
m
iss
i
on
w
ou
l
d
m
ake
r
obot
e
xecu
te
w
ron
g
c
omm
a
nd
s
o
that t
he fo
rm
ation
c
ould
not b
e esta
blishe
d.
In
the
ne
ar
f
utu
re
,
we
plan
t
o
de
velo
p
m
or
e
pr
oc
ess
in
th
e
m
ulti
-
ro
bot
schem
es
su
ch
as
colli
sion
avo
i
dan
ce
an
d
ta
sk
al
locat
i
on.
It
will
be
pur
posed
t
o
gain
le
ss
tim
e
process
in
t
arg
et
or
desti
nation
accom
plish
m
e
nt.
For
the
exam
ple,
that
four
r
obots
w
il
l
deter
m
ine
their
ta
sk
inde
pende
ntly
based
on
inf
or
m
at
ion
shared
and t
heir po
sit
io
n
i
n
the
m
aze, and defi
ne
eac
h path
to
r
eac
h
the
tar
ge
t diff
e
re
ntly
.
REFERE
NCE
S
[1]
J.
Li
u
,
K.
Yu
a
n,
W
.
Zou
and
Q.
Yang,
"M
onte
Ca
rlo
Mult
i
-
Robot
Local
iz
a
ti
on
Based
on
Grid
Cel
ls
and
Chara
c
te
rist
ic Pa
rti
cles"
,
In
te
rnat
ional
Conf
ere
nc
e
on
Adv
an
ce
d
I
nte
lligent Mecha
tronic
s,
2005
.
[2]
Ta
tiy
a
Padang
Tungga
l
,
e
t
al,
“
Purs
uit
Algorit
hm
for
Robot
T
rash
Can
Based
on
Fuzz
y
-
Ce
ll
Dec
om
positi
on”
,
Inte
rnational
Jo
urnal
of
El
e
ct
ri
cal
and
Computer
Engi
n
ee
ring
(
IJE
CE)
,
Vol.
6,
No.
6,
De
c
ember
2016,
pp
.
2863~2869.
[3]
K.H.
L
ee,
First
Course
on
Fuz
z
y
The
or
y
and
Applicati
ons,
Ber
li
n:
Spinger, 2005.
[4]
Adam
s,
L.
Vig
and
J.
A.,
"M
ult
i
-
Robot
Coal
itio
n
Form
at
ion"
,
IEE
E
Tr
ansacti
o
ns
on
Robot
ic
s
,
vol.
22,
no.
4,
p
p.
637
-
649,
2006
.
[5]
FR
Q
Aini,
AN
J
at
i
,
U
Sunar
y
a,
“
A
study
of
Mo
nte
Car
lo
Local
i
za
t
ion
on
Robot
Opera
ti
ng
S
y
s
tem
”,
Inte
rnationa
l
Confe
renc
e
on
I
nformation
Tech
nology
S
yste
m a
nd
Innov
at
ion
(
ICITSI)
,
Bandung,
2015
.
[6]
M.
Čáp,
P.
Nov
ák,
A.
Kl
ei
n
er
a
nd
M.
Selecký,
"P
riori
ti
z
ed
Plan
ning
Algorit
hm
s
for
Traje
c
tor
y
Coordin
at
ion
of
Multi
ple
Mobil
e
Robots
"
,
IEE
E
Tr
ansacti
ons
on
Aut
omation
Sc
ience
and
Engi
n
eer
ing
,
vol.
12,
no
.
3,
pp.
835
-
849,
2015.
[7]
A.M.
Fatha
n,
A.N.
Jati
and
R.
E.
Sapu
tra,
"M
appi
ng
Algori
thm
Us
ing
Ultr
asonic
and
Co
m
pass
Sensor
on
Autonom
ous
M
obil
e
Robo
t,
"
in
2016
Inte
rnat
io
nal
Confe
ren
ce
on
Control,
El
e
ct
ronics,
Re
n
ew
able
En
ergy
an
d
Comm
unic
ati
ons (
ICCER
EC)
,
Ba
ndung,
20
16
.
[8]
C.
H.
Hs
u
and
C.
F.
Juang,
"M
ult
i
-
obj
ective
C
onti
nuous
-
Ant
-
Colon
y
-
op
ti
m
ize
d
FC
f
or
Robot
W
al
l
-
Following
Control
"
,
I
EEE
Computati
onal I
nte
lligen
ce Mag
azine
,
pp.
28
-
40
,
2013
.
[9]
S.
Li
,
R.
Kong
and
Y.
Guo,
"Coopera
ti
v
e
Distribut
ed
Source
Seeki
ng
b
y
Multi
pl
e
Robots:
Algorit
hm
s
and
Expe
riments"
,
I
EE
E/ASME
Tr
ansacti
ons on
M
e
chat
ronics
,
vol
.
19,
no
.
6
,
pp
.
18
10
-
1820,
2014
[10]
L.
Luo,
N.
Chak
rab
ort
y
and
K.
S
y
ca
r
a,
"P
rova
bl
y
-
Good
Distribu
te
d
Algorit
hm
for
Constrai
ned
Multi
-
Robot
Ta
s
k
As
signm
ent
for
Grouped
Ta
sks
"
,
IEEE
Tr
ansactions
on
Robotics
,
vol. 31, no. 1, p
p.
19
-
30
,
2015
.
[11]
A.
Marino,
G.
Antonel
li,
A.P.
Aguiar
,
A.
Pasc
oal
and
S.
Chiaverini,
"A
Dec
ent
ral
i
ze
d
Stra
te
g
y
for
Multi
robo
t
Sam
pli
ng/Pa
trol
l
ing:
The
o
r
y
and
Expe
riments
"
,
I
EE
E
Tr
ansacti
o
ns
on
Control
S
yste
ms
Techn
olo
gy
,
vo
l.
23,
no.
1,
2015.
[12]
D.
P
ana
gou,
D.
M.
Stipa
novi
c
a
nd
P.
G.
Voulga
ris,
"D
istri
but
ed
Coordina
t
ion
C
ontrol
for
Mul
ti
-
Robot
Network
s
Us
ing
L
y
apunov
-
Li
ke
B
arr
ie
r
Fu
nct
ions"
,
IE
EE
Tr
ansacti
ons
on
Aut
omatic
Cont
rol
,
vol.
61
,
no.
3,
pp.
617
-
632
,
2016.
[13]
L.
Sa
battini,
C.
Secc
hi,
M.
Co
ce
t
ti
,
A.
L
evr
atti
and
C.
Fantu
zz
i
,
"Im
ple
m
ent
at
ion
of
Coordi
nat
ed
Com
ple
x
D
y
namic
Beh
aviors
in
Mult
irobo
t
S
y
st
ems
"
,
IEEE
Tr
ansacti
ons
on
Robotics
,
vol. 31, no. 4, pp. 10
18
-
1032,
2015
.
[14]
A.P.
Suparno,
H
.
W
id
y
antara
an
d
Hari
ant
o
,
"S
i
m
ula
si
Tra
j
ector
y
Pl
anni
ng
dan
Pem
bent
ukan
Form
asi
pada
Rob
ot
Obs
ta
cl
e
Avoida
nce
"
,
Journal
of Control
and
N
etwor
k
Syste
ms
,
v
ol.
4
,
no
.
1
,
pp
.
3
1
-
38,
2015
.
[15]
L.
Vig
and
J.A.
Adam
s,
"M
ult
i
-
Robot
Coaliti
on
Form
at
ion"
,
IE
EE
Tr
ansacti
ons
on
Robotics,
v
ol.
22
,
no.
4,
p
p
.
637
-
649,
2006
.
[16]
W
.
B.
Xu,
X.P.
Li
u,
X
.
Chen
a
nd
J.
Zha
o
,
"Im
prove
d
Artificial
Mom
ent
Method
for
Dec
en
tralized
Local
Pa
t
h
Planni
ng
of
Mul
ti
robots"
,
I
EEE
Tr
ansacti
ons on Control Sy
st
ems Tec
hnology
,
vo
l. 23, no. 6, pp. 23
83
-
2390,
2015
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
87
08
In
t J
Elec
&
C
om
p
En
g,
V
ol.
8
, N
o.
6
,
Dece
m
ber
2
01
8
:
5098
-
5106
5106
[17]
NT
Al
-
Ghifa
r
y
,
AN
J
at
i,
RE
Saputra
,
“
Coordina
ti
on
Contro
l
of
Sim
ple
Aut
onom
ous
Mobile
Robot”
,
IE
EE
Inte
rnational
C
onfe
renc
e
on
Instrum
ent
ati
on
Control
and
Aut
omation
,
Yog
y
ak
art
a
–
Indone
sia,
201
7.
10.
1109/ICA.
20
17.
8068420.
[18]
Siti
Nurm
ai
ni,
Bambang
Tut
uko,
“
Inte
ll
ig
ent
R
oboti
cs
Navigati
on
Sy
st
em:
Pro
ble
m
s,Met
hods,
and
Algorit
hm
”
,
Inte
rnational
Jo
urnal
of
El
e
ct
ri
cal
and
Computer
Engi
n
ee
ring
(
IJE
CE)
,
Vol.
7,
No.
6,
De
c
ember
2017,
pp
.
3711~3726.
[19]
SM
Mirza
ei
,
M
H
Moatt
ar,
“
Opt
imize
d
PID
Control
ler
wi
th
Ba
cteri
a
l
Foraging
A
lgori
thm
”
,
In
te
r
nati
onal
Journal
of
E
le
c
tric
al
and
Computer
Eng
i
nee
ring (
IJE
C
E)
,
Vol
.
5
,
No.
6,
Dec
ember
2015
,
pp.
1372~1380.
Evaluation Warning : The document was created with Spire.PDF for Python.