TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol. 15, No. 2, August 201
5, pp. 352 ~
360
DOI: 10.115
9
1
/telkomni
ka.
v
15i2.808
1
352
Re
cei
v
ed Ma
y 13, 201
5; Revi
sed
Jul
y
4, 2015; Accept
ed Jul
y
20, 2
015
Three-Dimensional Application-Specific Protocol
Architecture for Wireless Sensor Networks
Mosta
f
a Bag
houri
1
, Abde
rrahmane
Ha
jraoui
2
, Saad Chakk
or
3
Dep
a
rtment of Ph
y
s
ics, Com
m
unic
a
tion a
n
d
detection s
y
st
ems lab
o
rator
y
,
F
a
cult
y
of Scie
nces, Univ
ersit
y
of Abd
e
lma
l
e
k
Essaâdi, T
e
toua
n, Morocco
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: bagh
ouri.mo
stafa@gma
il.c
o
m
1
, ad_ha
jra
o
u
i@h
o
tmai
l.co
m
2
,
saadc
hakk
o
r@
gmail.c
o
m
3
A
b
st
r
a
ct
Many rese
archers assu
me that
the distributio
n of the nod
es
is don
e in a tw
o-dime
nsi
o
n
a
l
envir
on
me
nt in
the r
easo
n
th
at the
hei
ght
o
f
this n
e
tw
ork i
s
ne
gli
g
ib
le
co
mp
are
d
to
its
w
i
dth an
d l
e
n
g
t
h.
How
e
ver, in t
he re
ality, thr
ee-d
i
mens
ion
a
l
(3D) de
pl
oyment of the w
i
r
e
less
s
ensor
netw
o
rks is u
s
ed.
T
herefore,
ma
ny ap
plic
atio
n
s
r
equ
ire
3D
architectur
e
such as
un
de
rw
ater, space
commun
i
cati
ons,
atmos
p
h
e
ric, forest or b
u
i
l
di
ng. Unfort
unat
ely,
the
ener
gy cons
u
m
pti
on a
nd thr
o
u
ghp
ut in th
e
3
D
envir
on
me
nt d
e
creas
es cons
i
dera
b
ly co
mpa
r
ed to 2D
in
w
h
ich w
e
can
’
t n
e
g
lect the
m
i
n
s
o
me a
ppl
icatio
ns.
In his pap
er we app
lie
d the 3
D
architectur
e
in LEACH
prot
o
c
ol an
d w
e
prove by compute
r
simu
latio
n
ho
w
this 2
D
a
ppr
ox
imatio
n is
n
o
t r
easo
n
a
b
le
sinc
e the
lif
eti
m
e
o
f
3D W
S
N
d
e
c
r
ease
by
ab
out
21
%
over th
a
n
2D W
S
N.
Ke
y
w
ords
:
w
i
reless se
nso
r
netw
o
rks, LEACH pr
otoco
l
,
Energy-
e
ffici
ency, 2D
an
d
3D W
S
N, ne
tw
ork
lifeti
m
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
In the reality the physical
world
we liv
e in
, is a 3
D
envi-ronm
ent. Therefo
r
e, many
appli
c
ation
s
, su
ch as und
e
r
wate
r,
u
nde
rgrou
nd,
ai
rbo
r
ne, spa
c
e
co
mmuni
cation
s,
atmo
sph
e
ri
c,
forest, b
ody o
r
buil
d
ing, of
WSN
depl
oyed in th
ree
-
di
mensi
onal
sp
ace
(see
Figu
re1
)
. A wi
rele
ss
sen
s
o
r
netwo
rk (WS
N
) is
con
s
id
ere
d
a
s
thre
e-dim
e
nsio
nal (3
D) whe
n
the hei
ght of deploy
ed
sen
s
o
r
no
de
s field is n
o
t n
egligible
a
s
com-p
a
re
d to l
ength a
nd b
r
eadth of n
e
twork [1]. Ho
we
ver,
with the com
p
lexity of
the design a
nd
analysi
s
of
the 3D WS
N, wirel
e
ss sen
s
or network in
2D
plane a
r
e mo
re studi
ed tha
n
in 3D spa
c
e
.
A 3D wi
rele
ss sensor n
e
twork i
s
a set wirel
e
ss
sen
s
or no
de
s dist
ributed i
n
a 3
D
plan
e.
Each sen
s
or
node ha
s em
issi
on to sen
s
e the event
s detection, su
ch a
s
temperature, pressu
re
or vib
r
ation
a
nd
sen
d
thei
r mea
s
u
r
eme
n
ts to
wa
rd
a
pro
c
e
ssi
ng
center called
sink [1,
2]. Du
e to
the limitation
in their
battery capa
city wh
ich thei
r repla
c
eme
n
t is
im
possibl
e, opti
m
ization
of th
is
uniqu
e re
so
u
r
ce
ha
s be
co
me an im
port
ant issue.
No
de cl
uste
ring
is an
effective tech
niqu
e for
improvin
g the
ene
rgy effici
ency
and
prol
ongin
g
t
he
n
e
twork lifetim
e of a
WS
N [
3
] and
ha
s b
een
widely stu
d
ie
d in 2D WS
Ns.
LEACH [3, 4]
is o
ne
of the
first
proto
c
ol
s
whi
c
h
use thi
s
te
chni
que
a
nd h
a
s be
en
applie
d
into
the und
erwater
envi
r
on
-ment by
doin
g
so
m
e
chan
ge
s [
5
-8]. All of
these
literatu
r
e
s
con
s
id
ere
d
that the node
s are di
strib
u
te
d in tow-di
me
nsio
nal area.
In this pap
er,
we sho
w
tha
t
approximate
t
he 3D field i
n
the 2D
envi
r
onm
ent depl
oyment
is not negli
g
ib
le whe
n
a hei
ght of network is g
r
eate
r
.
The re
st of the pape
r org
a
n
izatio
n is do
ne as follo
ws:
Section II summari
ze
s the
related
work. T
h
ree-dimensional
wire-les
s sensor
net
work model
i
s
pr
ovided in
secti
on III. The Si
mu-
la
tio
n
r
e
s
u
lts a
r
e
ca
rr
ie
d
ou
t
in
s
e
c
t
io
n
IV.
F
i
n
a
lly w
e
c
o
nc
lud
e
o
u
r
r
e
se
ar
ch
w
o
rk
an
d
g
i
ve
s
o
me
per
spe
c
t
i
v
e
s i
n
se
ct
ion V
.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Thre
e-Dim
e
n
s
ion
a
l Application-Sp
ecifi
c
Protocol
Architecture for
Wirel
e
ss… (Mostafa
Bag
houri
)
353
a)
b)
c)
d)
Figure 1. Exa
m
ples of three-dimensiona
l wi
reless sen
s
or networks:
a) underwate
r, b)
underground,
c) airborne, and d) body
2. R
e
lated
Work
Some works
try to use th
e existing
WSN cl
uste
ring
proto
c
ol
for
WSN i
n
u
n
d
e
rwater
environ
ment.
Refe
re
nce
[6] assum
e
d
UASNs
are
less dynam
ic tha
n
n
o
rmal
WSNs
and
prop
osed
th
e LEACH-L, wh
ich upd
ates
its state
lo
cally
, and
red
u
ced
the ove
r
he
ad
of LEACH. In
[9], a clusteri
ng schem
e is propo
se
d in the contex
t o
f
routing sch
e
me to extend the lifetime of
UASN. Refe
rence [10] de
sign
ed a cl
uster stru
ct
u
r
e
without con
s
i
derin
g ene
rg
y problem. G
u
et
al [11] have pre
s
ente
d
a feasi
b
le ro
utin
g prot
o
c
ol for undergro
und
WSN in coa
l
mine, called
LEACH-min
e
. In the algo
rithm, all nod
e
s
are lo
cated
in three
sid
e
s
of the XY p
r
oje
c
tion pl
an
e
and in the in
ternal of the
rectan
gula
r
of the XZ
projectio
n
plan
e. Zhou et al [12, 13] have
deploye
d
a
2
D
WSN fo
r
coal min
e
, co
mpari
ng
to th
e ra
ndom
no
de de
ployme
nt strate
gy, the
strategy prop
ose
d
in this work can prolong t
he life by two times. Ho
weve
r, they have
not
con
s
id
ere
d
the influen
ce of
height of the netwo
rk.
Gene
rally, in the pra
c
tical
appli
c
ation
s
of
WSN, the sen
s
o
r
nod
es need to be deployed
and
comm
uni
cate in th
e th
ree
-
dime
nsi
o
nal area in t
he orde
r to
monitori
ng th
e ho
stile re
gi
ons
su
ch a
s
un
d
e
rwater, un
d
e
rg
rou
nd min
e
, airbo
r
ne,
a
nd body e
n
vironm
ents. Th
erefo
r
e, to more
approa
ch to the reality situ
ations, a 3
D
WSN d
eploy
ment is stu
d
ied dentally in
this pape
r.
Based
on th
e analy
s
is
a
bove, we fin
d
t
hat few
works
on 3
D
deployme
nt have be
en
studied for WSNs. Driv
en by this observation;
we will show by simul
a
tion that these
assumptio
n
s
and ap
proxim
ations a
r
e not
reas
ona
ble i
n
some a
ppli
c
ation
s
of WSN.
3.
Three
Dimen
s
ional Wirel
ess Senso
r
Net
w
o
r
k M
o
dels
3.1.
Energ
y
Mod
e
l
This study assume
s a simple model
for
the radio
hardware where the transmitter
dissipates en
ergy for runni
ng the radio electroni
cs to transmit and amplify
the
signals, and the
receiver runs the radio elec
tronics for recept
ion of signals [7]. Mu
ltip
ath fading
model (
d
power
loss) for larg
e distance transmissions
and the free
space mode
l (
d
power loss) for proxima
l
transmissio
n
s are conside
r
ed. Thus to tr
ansmit an
b
its
message over a
distance
d
, the
radio
expends:
,
,
(
1
)
(
2
)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 15, No. 2, August 2015 : 352 –
360
354
,
,
,
(
3
)
Where
d
o
is the distance threshold for swapping ampl
if
ication mode
ls, which can be calculated as
To receive an
message the receiver expends:
(
4
)
To aggregate
n
data signals of length
l
b
its
,
the energy con
s
umption wa
s calculated
as:
(
5
)
3.2. Net
w
o
r
k
M
o
del
This section describ
es the network mod
e
l and other b
a
sic assumptions:
1)
N sen
s
ors are uniformly
distributed within a square 3
D
rectangula
r
field of area
. The Base Station is positioned at the c
enter of the square
regio
n
. The number o
f
senso
r
nodes N to be deployed depends spec
ifically on the application.
2)
All nodes are deployed randomly.
3)
Each senso
r
can sen
s
e the environm
ent in
the 3D sphere of radius r.
4)
All sensors are homogeneo
us, i.e.,
they
have the same capacities.
5)
All the senso
r
nodes have
a particular id
enti
fier (ID) al
located to the
m
. Each clu
s
ter
head coordin
a
tes the MAC
and routing o
f
packets
within their clusters. (see Figure 2)
Figure 2. Three-dimension
a
l Wi
reless Sensor Network model
3.3.
Optimal Nu
mber of Clus
ter
We assume there are
no
des distributed uniformly in
3D region. If there
are
clusters, there are on
average
/
nodes per cluster. Each cluster-head dissipa
tes energy
receiving sig
nals from the nodes and
transmitti
ng
the aggrega
te signal to the base station.
Therefore, the energy dissipated in the
cl
uster-head n
ode during a single frame is:
(
6
)
Where
is the number of bits in each data messag
e
,
is the distance from the cluster head
node to the
BS, and we ha
ve assumed perfect data aggregation
.
The expressi
on for the ene
rgy spends b
y
a non-cluster head is given by:
(
7
)
Where
is the
distance from the node to
the cluster he
ad.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Thre
e-Dim
e
n
s
ion
a
l Application-Sp
ecifi
c
Protocol
Architecture for
Wirel
e
ss… (Mostafa
Bag
houri
)
355
Let
Ed
be the Expected distance of cluster h
ead from the base station. Assuming
that
the nodes are uniformly distributed,
so it is calculated as follows:
,
,
(
8
)
Where
f
x,
y,
z
is the
probability d
ensity
function of
three di
mensions ran
dom variable
,
,
which is uniform and given
by:
(
9
)
If we assume that base
station is the
cent
er of th
e network we can passin
g
in the
spheri
c
al coo
r
dinates:
,
,
sin
(
1
0
)
The area of network is asp
heric with radius
3/4
.
If the density of senso
r
nod
es is uniform
throughout the area
then b
e
comes inde
pendent
of
r
,
θ
and
then:
0
.
5312
(
1
1
)
The expected
square
d
distance from th
e nodes to the cluster hea
d (assumed to be at
the center of
mass of t
he cluster) is given by:
,
,
sin
(
1
2
)
If we assum
e
this area is a sphe
re
with radius
3/4
and
,
,
is
constant for
r
,
θ
and, (10) simplifies to
:
sin
/
(
1
3
)
If
the density
of nodes is uniform throughout the cluster area, then
/
and:
(
1
4
)
Therefore, the total energy dissi
pated in
the network per round,
, is expresse
d
by:
(
1
5
)
Where
is the
energy dissip
a
ted in cluster which giving by:
1
(
1
6
)
This can be calculated by:
(17)
Therefore, the total energy dissipated
in
the network is simplified b
y
:
2
(18)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 15, No. 2, August 2015 : 352 –
360
356
We can find the
optimu
m
number of clus
ters
by
s
e
tting the
d
e
rivative
of
with
res
p
ec
t to c
t
o
z
e
ro.
0
(
1
9
)
0
.2147
(
2
0
)
The optimal p
r
obability
for
becoming a cl
uster-hea
d ca
n also be computed as:
(
2
1
)
Figure 3. Variation of energ
y
consumption
for different
values of clusters number
c
In Figure 3, we
s
h
ow the
average energy c
o
ns
umption per round
by
each s
e
nsor node
as a fun
c
tion
of the numbe
r of cluste
rs for two ty
pes
of model, 3D
and 2
D
WSN.
Firstly, graph
of
the LEACH 3
D
mo
del follo
w the
sam
e
o
f
LEACH
2D.
Secon
d
ly, the
gra
ph of LE
ACH
3D
mod
e
l
sho
w
s that
th
e si
mulation
agre
e
s
well
with the
anal
ysis
re
sults. In
t
he oth
e
r ha
n
d
, the
3D mo
del
con
s
um
es m
o
re
ene
rgy th
an the 2
D
m
o
del which de
pend
s e
s
sent
ially to the no
negligi
b
le val
u
e
of the net
work hei
ght. Ho
wever, thi
s
model h
a
s
a
n
optimal
nu
mber
of clu
s
t
e
rs le
ss th
an
the
other
mod
e
l
whi
c
h
ca
n e
x
ploit advant
ageo
us to m
i
nimize
the li
fetime of the
network.In this
se
ction, it is
explained th
e
result
s of research a
nd at
the sam
e
time is given the
comp
reh
e
n
s
i
v
e
discu
ssi
on.
Result
s
can
be
prese
n
ted i
n
figure
s
,
g
r
ap
h
s
, table
s
and
others th
at m
a
ke
the
re
ad
er
unde
rsta
nd e
a
sily [2, 5]. T
he discu
s
sion
can be ma
de
in several
su
b-chapte
r
s.
4. Simulation
Results
4.1.
Parameter Setting
s
In this secti
on, we
stud
y the pe
rform
ance of
L
EACH 3D
p
r
ot
ocol unde
r
different
scena
rio
s
usi
ng MATLAB. We con
s
id
er a
model illustrate in the Figure 2 with
N
100
nodes
rand
omly an
d uniformly
distrib
u
ted
in a
100m
100m
100m
field. To comp
are th
e
perfo
rman
ce
of LEACH 3
D
with
LEACH 2
D
p
r
oto
c
ol, we i
gno
re
the effect
caused by
sig
nal
colli
sion a
nd i
n
terferen
ce i
n
the wi
rele
ss chan
nel. Th
e radi
o pa
ra
meters u
s
ed i
n
our
simul
a
tions
are sho
w
n in
Table 1.
2
4
6
8
10
12
14
0.
0
4
5
0.
04
55
0.
0
4
6
0.
04
65
0.
0
4
7
0.
04
75
0.
0
4
8
0.
04
85
0.
0
4
9
0.
04
95
0.
0
5
X
:
3.
59
1
Y
:
0.
04
56
4
N
u
m
b
er o
f
cl
u
s
t
e
r
s
E
n
e
r
gy c
o
n
s
um
pt
i
o
n(
J
)
T
o
t
al
energy
c
o
n
s
um
p
t
i
o
n i
n
t
h
e
net
w
o
r
k
p
e
r ro
un
d
X:
2
.
9
Y:
0
.
04
57
8
L
E
A
CH 2D
L
E
A
CH 3D
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Thre
e-Dim
e
n
s
ion
a
l Application-Sp
ecifi
c
Protocol
Architecture for
Wirel
e
ss… (Mostafa
Bag
houri
)
357
Table 1. Energy Model Parameters
Parameter Value
Initial Node Ener
g
y
0.5J
N 100
P 0.05
E
50
nJ
/bit
E
5
pJ/
b
it
ϵ
10
pJ/bit/m
ϵ
0.0013
pJ/bit/m
d
100 m
500 B
y
tes
Rounds
2000
4.1.1. Simula
tion metrics
Perform
a
n
c
e
metrics u
s
ed
in the simulati
on study are:
1)
Energy con
s
u
m
ption analy
s
is
2)
Numb
er of ali
v
e node
s per
roun
d.
3)
Percentag
e o
f
Node death
4) Thro
ugh
put
5) De
cre
a
se:
100
(
2
2
)
4.2. Simulation
Results
4.2.1.
Energ
y
consumption ana
l
y
s
is
The perfo
rma
n
ce of LEACH 3D is
comp
ared
with
that of the original LEACH in term
s of
energy a
nd i
s
sho
w
n
in
Figure 4.
Wi
th the
us
e o
f
3D de
ploy
ment of
nod
es, the
e
nergy
con
s
um
ption
of the net
work is de
crea
se
d. This i
s
du
e to the g
a
in
of the en
erg
y
dissi
pated
by
height of net
work. From the graph it is clea
r t
hat LE
ACH 3
D
de
crease twice th
e ene
rgy savi
ngs
than LEACH proto
c
ol.
Figure 4. Energy analysi
s
compa
r
ison of LEACH 3
D
a
nd LEACH 2
D
4.2.2. Net
w
o
r
k
Li
fe
time
The
numb
e
r
of nod
es aliv
e for ea
ch
round
of d
a
ta
tran
smi
ssi
on
is
ob
se
rved
for th
e
LEACH 2
D
a
nd 3
D
protocols to
eval
uat
e the
lifet
ime
of the
network. Fig
u
re
5
a
nd Fi
gure
6
show
the pe
rform
a
nce
of LEACH 3
D
com
pared to LEA
C
H 2D. It is ob
served th
at th
e LEACH 3
D
is
less perfo
rm
than LEACH 2D du
e to energy
dissipation of individual nod
e throug
hout
the
netwo
rk
whi
c
h depe
nd e
s
sentially on t
he distan
ce b
e
t
ween n
ode
s
and si
nk.
0
10
20
30
40
50
60
0
250
500
750
1000
1250
1500
1750
2000
Tota
l
Ene
r
g
y
of
the
Network
(J)
R
o
unds
LEACH_3D
LEACH_2D
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 15, No. 2, August 2015 : 352 –
360
358
Figure 5. Nu
mber of de
ad
node
s per
ro
und
compa
r
i
s
on of LEACH
3D an
d LEACH 2D
Figure 6. Nu
mber of alive
node
s pe
r ro
und
compa
r
i
s
on of LEACH
3D an
d LEACH 2D
4.2.3. Through
put
Referre
d
to Figure 7, it
sho
w
cl
early
that LEACH 3D p
r
ovid
e a poor through
put
comp
ared
to LEACH 2D p
r
otocol,
this decrea
s
e
i
s
j
u
stified by th
e low lifetime
whi
c
h give t
he
three dim
e
n
s
i
onal de
ploym
ent of the nod
es in the net
work.
Figure 7. Performa
nce of the proto
c
ol
s
4.2.4. Decr
eas
e
Generally, we can illustrat
e
t
he decrease of the LEACH 3D
in the Figure 8. It’s noted
that the throu
ghput d
e
cre
a
s
e
s
21%
as
much
than
L
EACH 2
D
du
e to its le
ss e
nergy.
Whe
r
e
a
s,
LEACH 2
D
o
u
tperfo
rms th
e FND of LE
ACH 2
D
by
21% and by 28% for LND. In the other h
and,
LEACH 3
D
consume
s
32
% more ene
rgy than LEACH 2
D
.
0
10
20
30
40
50
60
70
80
90
100
110
0
250
500
750
1000
1250
1500
1750
2000
Number
of
de
a
d
node
s
R
o
unds
LEACH_3D
LEACH_2D
0
10
20
30
40
50
60
70
80
90
100
110
0
250
500
750
1000
1250
1500
1750
2000
Number
of
al
i
v
en
o
d
es
R
o
unds
LEACH_3D
LEACH_2D
0
1000
2000
3000
4000
5000
6000
0
250
500
750
1000
1250
1500
1750
2000
Number
of
pa
cke
t
s
R
o
unds
LEACH_3D
LEACH_2D
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Thre
e-Dim
e
n
s
ion
a
l Application-Sp
ecifi
c
Protocol
Architecture
for Wirel
e
ss… (Mostafa
Bag
houri
)
359
Figure 8. De
crea
se of LEACH 3
D
comp
ared to LEACH 2D.
4.3. Resul
t
An
aly
s
is
From ou
r si
mulation
s, we obse
r
ved
that LEACH 3D con
s
um
es more en
ergy and
delivers less
packet
s
to the base
statio
n. These re
sults ca
n be in
terpreted by the differen
c
e
of
distan
ce b
e
twee
n nod
es i
n
both situati
ons
whi
c
h na
turally cau
s
e
s
by the ran
d
o
m deploym
e
nt of
node
s.
5.
Conclu
sion and Futu
r
e
Work
R
e
c
e
n
t
ly,
3D
w
i
r
e
le
ss
s
e
n
s
o
r
ne
tw
ork
s
ha
ve
kno
w
n
a
great
prevale
n
t due
to thei
r la
rge
appli
c
ation
s
su
ch as
un
d
e
rwater, sp
ace
commu
ni
ca
tions, atmo
sp
heri
c
, forest
or b
u
ildin
g. T
he
analytic of 3
D
WSN i
s
more
compl
e
xity
than
the analytic in 2D WS
N. Therefo
r
e, ma
ny
resea
r
che
s
p
r
oje
c
t the 3D WSN in 2D
WSN. In th
is pape
r, we de
monst
r
ate by
simulation, that
this app
roxim
a
tion is n
o
t reasona
ble if the heig
h
t
of netwo
rk i
s
g
r
eater tha
n
le
ngth and
bre
adth
of this netwo
rk. We st
rongl
y persu
ade
d that proje
c
tion
of WSN in 2D enviro
nme
n
t is unju
s
tifiable
in reason that the 3D
WSN i
s
much
cl
oser to our physical
word.
As
future
work, we
will
work
to
optimize
the
energy con
s
u
m
ption of
this network,
sin
c
e th
e n
u
mb
er
of cl
uste
r
head
in
3D
WSN
gives mo
re re
sult than 2
D
WSN.
Referen
ces
[1]
HM Ammari, S
K
Das. C
o
ver
a
ge
and
con
nec
tivit
y
in thr
ee-
di
mensi
ona
l
w
i
re
less se
nsor
net
w
o
rks us
in
g
percolation theory
.
IEEE Trans. Parallel Distr
i
b. Syst. (IEEE
TPDS)
. 2009; 20(6).
[2] Kay
Romer,
Friedemann
Mattern. T
he Design Spac
e of W
i
reless Se
ns
or Net
w
orks.
IEEE Wireles
s
Co
mmun
icati
o
ns.
2004; 1
1
(6)
:
54-61.
[3]
W
endi
R H
e
in
zelma
n
, Ana
n
tha C
h
a
ndrak
a
s
an, Har
i
Ba
la
krishn
an.
En
er
gy efficie
n
t co
mmu
n
icati
o
n
protoco
l
for w
i
reless
micr
ose
n
sor netw
o
rks
. IEEE International Conference on S
y
stem
Sciences
.
200
0: 1-10.
[4]
W
Heinz
e
lma
n
,
A Cha
ndr
aka
s
an, H B
a
l
a
kri
s
hna
n. An
ap
plicati
o
n
spec
ific prot
ocol
arc
h
itecture
fo
r
w
i
rel
e
ss micr
o
s
ensor
net
w
o
r
ks.
IEEE T
r
ansactions
on W
i
reless
Co
mmunic
a
tions
. 20
02;
1(4): 66
0-
670.
[5]
Xi
a
Li, Yo
ng
qi
an W
a
ng, Ji
ng
j
i
n Z
h
ou.
An
e
n
e
rgy-effici
ent cl
usterin
g
a
l
g
o
rit
h
m for
und
erw
a
ter ac
oustic
sensor n
e
tw
orks
. Control En
gin
eeri
ng a
nd
Commun
i
cati
o
n
T
e
chnolo
g
y
(
I
CCECT
), 2012 Internati
ona
l
Confer
ence. 2
012: 71
1-7
14.
[6]
X
Li, S
L
F
a
n
g
,
YC Z
han
g.
T
he study
on
clusteri
ng
alg
o
r
ithm of
the
u
nderw
a
ter
aco
u
stic se
nsor
netw
o
rks.
T
he 14th Intern
atio
nal C
onfer
enc
e
on Mech
atron
i
cs and Mac
h
i
n
e Visio
n
in Pr
a
c
tice (M2VIP
200
7). 200
7.
[7]
G
uangs
on
g Yang, Min
gbo
Xi
ao, En Ch
e
ng, Jing Z
h
a
n
g
.
A cluster-head se
lectio
n
scheme for
und
erw
a
ter a
c
oustic se
nso
r
netw
o
rks
. Commun
i
cati
o
n
s an
d Mob
i
l
e
Comp
utin
g
(CMC), 201
0
Internatio
na
l C
onfere
n
ce. 20
1
0
; 3: 188-1
91.
[8]
Liu G, Wei C.
A new
multi-
path ro
utin
g p
r
otocol
bas
ed
on
cl
uster for
und
erw
a
ter ac
oustic se
nso
r
netw
o
rks
. Internatio
nal C
onfer
ence o
n
Mult
im
edi
a T
e
chnolo
g
y
(ICMT
)
. 2011: 91-94.
[9]
P W
ang,
C
Li,
J Z
hen
g.
D
i
stri
buted
mi
ni
mu
m-cost
c
l
uster
i
ng pr
otoc
ol f
o
r un
derw
a
ter s
e
nsor
netw
o
rks
(UWSNs)
. IEEE Internatio
nal
Confer
ence
on
Co
mmunic
a
tio
n
s (ICC 200
7). 200
7: 351
0-35
15.
[10]
Salva-G
a
r
au F
,
Stojanovic M.
Multi-cluster p
r
otocol for ad
hoc mo
b
ile u
n
derw
a
ter acou
stic netw
o
rks
.
O
C
EANS 200
3
Proceed
in
gs. 200
3; 1(26): 91
-98.
FND
‐
21%
LND
‐
28%
Throug
hput
‐
21%
Energy
‐
32%
LEACH
3D
an
d
LEACH
2D
in
crease
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 15, No. 2, August 2015 : 352 –
360
360
[11]
X Gu, Y Jin, Y Sun, J Yan.
Maxi
mu
m lifeti
m
e ro
utin
g strategi
es fo
r wir
e
less sensor networks in coal
mi
ne
. In 201
0 Internatio
na
l C
onfere
n
ce o
n
Comp
uter Engi
neer
ing a
nd T
e
chno
log
y
. 2
010
: 341-34
4.
[12]
Z
C
Z
hu, GB
Z
hou, GZ
Chen. Ch
ain-t
y
p
e
w
i
re
less u
n
dergr
oun
d min
e
sens
or net
w
o
rks for ga
s
monitori
ng.
Ad
vance
d
Scie
nc
e Letters.
201
1
;
4(2): 391-39
9
.
[13]
GB Z
hou, Z
C
Z
hu, GZ
Chen, NN Hu.
Energy-efficie
n
t chain-typ
e
w
i
rele
ss sensor net
w
o
rk for gas
mo
nitori
ng
. In Internati
o
n
a
l Co
nferenc
e on Inf
o
rmatio
n
an
d Comp
uting Sci
ence. 20
09: 12
5-12
8.
Evaluation Warning : The document was created with Spire.PDF for Python.