TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol. 12, No. 10, Octobe
r 20
14, pp. 7452
~ 745
8
DOI: 10.115
9
1
/telkomni
ka.
v
12i8.557
0
7452
Re
cei
v
ed
Jan
uary 11, 201
4
;
Revi
sed
Jul
y
22, 201
4; Accepted Aug
u
st 16, 2014
The Research of DV-HOP Positioning Algorithm Based
on RSSI Calibration
Jie Cao, Mu
Chen*
Schoo
l of Com
puter an
d Com
m
unic
a
tion, L
a
n
zho
u
Univ
ersi
t
y
of T
e
chnolo
g
y
, L
anzh
ou 7
300
50, P.R.Chi
n
a
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: chenmu
a
j@
1
63.com
A
b
st
r
a
ct
Becaus
e of the
DV-HOP alg
o
r
i
thm
in w
i
rel
e
s
s
sens
or network (WSN), the first jump range error,
w
h
ich affect t
he
positi
oni
ng
error
of the
w
hole
nod
e
n
e
tw
ork prob
le
ms. T
h
is
pa
pe
r puts forw
ard
a
n
improve
d
al
gor
ithm is
prop
ose
d
. T
he alg
o
rith
m is
mai
n
ly us
i
ng the RSSI ra
ngi
ng tech
nol
o
g
y, in the proc
ess
of DV-HOP first jum
p
range thro
u
g
h
corre
ction
of RSSI
ran
g
in
g tec
h
nol
ogy, to
red
u
ce th
e
distan
c
e
me
asuri
n
g
erro
r, in or
der t
o
i
m
prove
pos
itio
ni
ng
accura
cy. T
he si
mul
a
tion
r
e
sults s
how
th
at co
mp
ared
w
i
th
traditio
nal
DV-
H
OP alg
o
rith
m. T
he i
m
pr
ove
d
al
gorit
hm
ca
n w
i
thout a
d
d
i
tion
al h
a
rdw
a
re
and
co
mp
utati
o
n
und
er the pre
m
ise of impr
ov
in
g positi
oni
ng a
ccuracy.
Ke
y
w
ords
: location algorithm, DV-HOP
, RSSI, hybrid posit
ioni
ng
Co
p
y
rig
h
t
©
2014 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
W
i
r
e
le
ss se
ns
o
r
ne
tw
or
k
(
W
S
N
)
is a
k
i
n
d
o
f
b
r
and
-
n
ew in
fo
rma
t
io
n
ac
qu
is
itio
n
an
d
pro
c
e
ssi
ng m
e
thod. It is compo
s
ed
of many Light
q
uality, small
size, and the
price of che
a
p
sen
s
o
r
node
s
[1].
Equi
p
ped with a low-po
we
r
transceive
r
s a
nd limited
d
a
ta p
r
o
c
e
ssi
ng
cap
ability. No
des
are
ra
nd
omly deploye
d
in a a
r
ea
of intere
st. Ran
dom de
ploym
ent nod
e can
not
kno
w
in adva
n
ce thei
r position.
Many wirele
ss
sen
s
o
r
netwo
rk ap
plicatio
ns a
r
e
based o
n
se
lf-
positio
ning
sensor, such a
s
targ
et tracki
ng, env
iro
n
m
ental monito
ri
ng, etc. It all depe
nd
s on t
h
e
locatio
n
of the se
nsor n
ode itself inf
o
rmatio
n.
Positionin
g
tech
nology playe
d
a key role
in
wirel
e
ss
se
nsor n
e
two
r
k a
pplication
s
. Nowa
days,
th
e
r
e a
r
e
many
sin
c
e l
o
caliza
t
ion alg
o
rithm
,
gene
rally
can
be
divided
i
n
to two
cate
gorie
s,
ran
g
e
-
ba
sed
an
d
range
-fre
e [2-3]. Ran
ge-ba
sed
algorith
m
whi
c
h u
s
ing the
absol
ute distance b
e
tw
e
e
n
point and
point or p
o
int
of informatio
n
betwe
en the
sen
s
o
r
nod
e positio
ning.
Many algori
t
hms ba
se
d on dista
n
ce, all need a
c
t
ual
distan
ce
me
asu
r
em
ent.
Such
a
s
: T
O
A [4]
tec
h
nology, TDOA [5] tec
hnology, AOA
[6]
techn
o
logy, a
s
well a
s
the
RSSI [7] technology. At prese
n
t many node
s have
RF tran
smi
ssion
function. Fo
r
the hardware
over
he
ad a
n
d
ene
rgy con
s
umptio
n p
r
o
b
lems,
Used
to estimate t
h
e
distan
ce ba
sed on the RSSI.
But because the RF
signal is ea
sily affected by environm
ental
factors, u
s
in
g RSSI
ran
g
i
ng te
chn
o
log
y
to obtai
n t
he di
stan
ce
error is big
g
er, the
alg
o
ri
thm
accuracy i
s
n
o
t satisfa
c
to
ry, other ra
ngi
ng alg
o
rithm
based o
n
di
stance, althou
g
h
com
p
a
r
ed t
o
the RSSI algorithm imp
r
ov
ed on the po
sitioning a
ccu
racy, but it is limited
by power co
nsumpti
o
n
and th
e
co
st
of ha
rdware.
Ran
ge-f
r
ee
te
chn
o
logy
doe
s n
o
t ne
ed to
predi
ct th
e di
stan
ce
betwe
en
the se
nsor n
ode
s Angle i
n
formatio
n. The typi
cal
range
-fre
e al
gorithm i
n
cl
u
des
DV-Ho
p
[8]
algorith
m
, ce
ntroid
algo
rithm [9], MDS-MAP [10]
alg
o
rithm a
nd A
P
IT [11] algo
rithm, etc. M
D
S-
MAP algo
rith
m nee
ds glo
b
a
l net
work
structure info
rm
ation, do
es n
o
t apply to
thi
s
e
nergy limit
ed
wirel
e
ss
sen
s
or net
works. Centroid alg
o
r
ithm po
sition
ing usi
ng the
con
n
e
c
tivity o
f
beaco
n
no
d
e
s
and blind n
o
de easy to operate, with
out coo
r
din
a
tion, a small amount of calcul
ation, high
scalability, but the positioni
ng erro
r i
s
bigger.
DV-Hop [12-13] algo
rit
h
m is easy to implem
ent, on
gene
ral term
s of the hard
w
are env
iron
ment is a go
od ch
oice. Compa
r
ed
with
the range
-b
a
s
ed,
in the aspe
ct of cost, hard
w
ar
e,
power con
s
um
ption rang
e-free
al
gorithm requi
res le
ss, and less
affected by e
n
vironm
ental
factors. Thi
s
positio
ni
ng
of coa
r
se a
c
curacy to m
e
et the accu
ra
cy
requi
rem
ents
of the positio
n is
not stri
ctl
y
Application
s
.
DV-Hop al
go
rithm usi
ng t
he produ
ct of t
he averag
e
hop di
stan
ce
and nu
mbe
r
of hops
betwe
en
nod
es i
s
th
e di
stan
ce, the
n
use trilate
ra
l po
sitioning
algo
rithm to
obtain
location
informatio
n o
f
unkn
o
wn n
ode
s. But be
cau
s
e th
e al
gorithm
usi
n
g jump di
sta
n
ce
expre
s
s
line
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
The Re
se
arch of DV-HOP Positionin
g
Algor
ithm
Base
d on RSSI Ca
libration
(Ji
e
Cao
)
7453
distan
ce, the
r
efore, no mat
t
er ho
w far the actual
di
sta
n
ce b
e
twe
en
two adja
c
e
n
t node
s, its ho
ps
are
1, e
s
tima
te of the di
st
ance of a
d
ja
cent
node
s
a
r
e to j
u
mp from the
avera
ge. It will le
a
d
to
large
r
p
o
sitio
n
ing e
r
ror, while the exi
s
ting alg
o
rithm
in view of the
error of th
e
probl
em
s of the
first jump, le
ss
re
se
arch.
This
pape
r p
r
opo
se
s a
DV-Ho
p
imp
r
o
v
ed algo
rith
m ba
sed
on
RSSI
calib
ration, can improve t
he
positio
ning accuracy.
2. DV-HOP Algorithm
2.1. DV-HOP Algorithm
Nicule
scu an
d Nath propo
sed a
kind of distri
b
u
ted co
mputing lo
cali
zation al
gorit
hm. The
core id
ea of t
he alg
o
rithm
i
s
u
s
e th
e ave
r
age
di
stan
ce
per hop
an
d
the hop
s
bet
wee
n
un
kn
o
w
n
node
s an
d a
n
ch
or no
de
s instea
d of the distan
ce b
e
twee
n un
kn
own n
ode
s a
nd an
cho
r
no
des.
Whe
n
an un
known nod
e receive
s
thre
e
or more
an
chor no
de
s inf
o
rmatio
n, the unkn
o
wn no
de
throug
h the
t
r
ilateral me
a
s
ureme
n
t me
thod to
cal
c
ulate thei
r o
w
n l
o
cation i
n
formatio
n. T
he
algorithm mechanism follows.
1) Node b
r
oa
dca
s
t to their neighb
or no
de with
thei
r locatio
n
information, inclu
d
ing th
e
hop field, Initi
a
lize
d
to
zero
, record the
receivin
g no
d
e
with th
e mi
nimum n
u
mb
er to
ea
ch a
n
cho
r
node, ign
o
rin
g
the large
r
the numb
e
r of
hops a pa
cket from the same an
cho
r
node
s then reco
rd
the hop i
s
on
e and fo
rwarded to a n
e
ig
hbor
nod
e, in
this way, all the nod
es in
the network can
record do
wn to the minimu
m hop co
unt each an
cho
r
node.
2) T
h
ro
ugh
the received i
n
formatio
n, a
n
ch
or
nod
es
cal
c
ulatin
g a
v
erage
len
g
th of e
a
ch
jump, And b
r
oad
ca
st ave
r
age h
o
p
s
d
i
stan
ce to the entire n
e
t
work. The
unkno
wn no
d
e
cal
c
ulate
d
the distan
ce b
e
twee
n the anchor
n
ode
by node ho
ps from the
anchor a
nd t
h
e
averag
e pe
r-hop di
stan
ce
informatio
n.
3) Wh
en the
unkn
o
wn node calculat
es a di
stan
ce of three o
r
more a
n
ch
or nod
e
distan
ce info
rmation, thro
ugh the trilat
e
ral me
a
s
u
r
e
m
ent method
or the maximum likelih
o
od
estimation m
e
thod calcula
t
e. Using the
formula (1)
to
estimate the
actual di
stan
ce of every hop.
i
j
j
i
j
j
i
j
i
i
h
y
y
x
x
C
2
2
)
(
)
(
(1)
In the fo
rmul
a (1),
i
i
y
x
,
,
j
j
y
x
,
i
s
the
co
ordinate
s
of a
n
chor n
ode
s I
and
j, h
j
is
the
numbe
r of ho
ps bet
wee
n
a
n
ch
or no
de
s I and j(i
j).
2.2. DV-Hop Algorithm In
adequ
a
te
In Figure 1, a
n
ch
or no
de
s are L1,L
2
,L3.
A is the unknown node n
eed
s to locat
e
. Three
anchor nod
es kno
w
th
e di
stance bet
wee
n
ea
ch
ot
her,
as
sho
w
n
in
Figure a
r
e 3
0
,
30 and
40. T
h
e
distan
ce b
e
twee
n A and L1 is 15, ho
p
count is 1.T
h
e numbe
r of hop
s A to L2 and L3 a
r
e th
ree.
Assu
ming the
length of each side i
s
10.
Figure 1. The
Schemati
c
of DV-Hop Algo
rithm
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 10, Octobe
r 2014: 745
2
– 7458
7454
DV-h
op alg
o
ri
thm sho
w
s L
1
, L2, L3 cal
c
ulated a
s
follows:
L1
：（
30
+30
）
/
（
4+4
）
=7.
5
L2
：（
30
+40
）
/
（
4+6
）
=7
L3
：（
30
+40
）
/
（
4+6
）
=7
L1
，，
L2
L3
wi
ll be b
r
oa
dca
s
t their calcul
ations. So
A to L1
dista
n
ce
is 7.5,
be
cau
s
e
only
jump is p
e
r j
u
mp di
stan
ce
. the distan
ce A to
L2 an
d L3 are 7.5*
3=2
2
.5. Ho
wever, the act
ual
distan
ce
of A and L
1
is 15, with
DV
- Hop al
gor
ithm to estim
a
te the di
sta
n
ce
of 7.5, this
positio
ning e
r
ror i
s
rea
c
h
e
d
200%.
2.3. An Improv
ed Algorithm Based o
n
RSSI Auxiliar
y
Whe
n
the
ho
ps b
e
twe
en
n
ode
s A an
d a
n
ch
or
nod
e L
1
is
1 can
be
easily
obtain
ed. but
the DV-Hop
algorith
m
will
increa
se the
erro
r.
The
r
e
f
ore, the introdu
ction of
RSSI technol
ogy
assiste
d
p
o
si
tioning. T
he
RSSI ra
nging
metho
d
can
reflect
the
differen
c
e
which cau
s
ed
by
the
uneven di
stri
bution of nod
es from the a
c
tual ju
mp. But the transm
i
ssi
on po
we
r loss is used to
cal
c
ulate the
distan
ce i
s
often affected very gre
a
t by environ
menta
l
factor
s su
ch
as topog
rap
h
y,
geomo
r
p
holo
g
y. To avoid
this
pro
b
lem, t
h
is
pap
er
doe
s n
o
t directly
use
RSSI al
g
o
rithm, b
u
t wi
th
the ratio
bet
wee
n
e
a
ch
h
op b
e
twe
en t
w
o
nod
es
RS
SI path len
g
th an
d the
averag
e
RSSI path
length a
s
RSSI path len
g
th co
rrectio
n
facto
r
.
The
n
with
co
rre
c
tion
coeffici
ent and
DV
-Hop
cou
n
ts pe
r ju
mping from m
u
ltiplication to
fix each jump distan
ce.
Then u
s
e t
he co
rrectio
n
factor
wit
h
av
era
ge
distan
ce of
every jum
p
ing from
multiplicatio
n
to fix each j
u
mp di
stan
ce
. The jum
p
d
i
stan
ce
can
be u
s
ed
a
s
a jump f
r
om
the
distan
ce b
e
tween no
de
s is
1. And it is
not affected by environ
menta
l
factors.
Algorithm spe
c
ific process i
s
de
scribe
d a
s
follows:
First step,
li
ke with
the DV-Hop algo
rith
m, Ancho
r
nod
es floo
ding b
r
oa
dca
s
t their
information which include its locati
o
n
informatio
n, their own ID, the
initial value of 0 jump ho
ps
with
Fixed po
wer, re
ceive
s
t
he data
nod
e Ho
p+1 an
d
recorded i
n
th
e pa
cket. Unli
ke the
DV
-Ho
p
algorith
m
. the packet is ad
ded by the theoreti
c
al
mod
e
l of the measured RSSI p
a
th length value
calle
d
RSSID(n),
nod
e the
n
forwa
r
d
s
t
he p
a
cket
to its
neighbors
at
a fixed
power, after
the
neigh
bor
nod
e re
ceive
s
th
e forwarded
messag
e,
Th
e hop
s
numb
e
r pl
us
one,
RSSI re
ceive
d
by
the o
w
n
path
length
me
asured
value
a
nd the
value
of the
path l
e
ngth in
the
p
a
cket i
s
add
e
d
to
the previou
s
hop, and a
d
d
i
ng a pa
cket. When it ar
ri
ve at a node
by n jumps,
the data in the
p
a
c
k
e
t is
c
o
or
d
i
n
a
t
es
, ho
ps
R
SSID(
1)
+R
SSID
(
2)+
…
+
R
SSID(
N)=R
SSID
N
ID
,
H
o
ps
. T
h
e
sa
me
as the DV-Hop algo
rithm, when re
ceived the packe
ts of the sam
e
ID in the hop co
unt is less
than the exi
s
ting hop, to b
e
repl
aced, o
t
herwi
se
discard. when
th
e
data
of
a
n
anchor nod
e is
received to a
nother
an
cho
r
nod
e thro
u
gh N ju
mps.
Thro
ugh the
method
DV-h
op algo
rithm
to
cal
c
ulate the
averag
e hop
distan
ce. At the sam
e
time
, calcul
ate the averag
e ju
mp RSSI average
path length, j
u
st like F
o
rm
ula (2
).
n
n
RSSID
RSSID
RSSID
RSSID
av
)
(
)
2
(
)
1
(
(2)
Then
broad
casts the
ave
r
age
ho
p di
stance
(H
OP
size) an
d the
avera
ge
RS
SI path le
ng
th
(RSSIDav
)
in the way of DV-Ho
p
algo
rithm.
Secon
d
step,
establi
s
h co
rrectio
n
facto
r
of
jump distance, with the any first jump RSSI
path length
RSSID(1
)
div
i
ded by the averag
e pat
h
length RSSIDav, like Formula (3
), as a
corre
c
tion
co
efficient of Ju
mp distan
ce.
av
RSSID
RSSID
)
1
(
(3)
Therefore,
when the h
op
cou
n
t betwe
e
n
node
s i
s
1,
and the di
st
ance betwee
n
cal
c
ulate t
w
o
points, such as Fo
rmula
(4).
size
Hop
d
(4)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
The Re
se
arch of DV-HOP Positionin
g
Algor
ithm
Base
d on RSSI Ca
libration
(Ji
e
Cao
)
7455
Whe
n
the
nu
mber of h
o
p
s
between
two
nod
es is
g
r
e
a
ter th
an
1, calcul
ate the
d
i
stan
ce
betwe
en
two poi
nts
according
to th
e traditio
nal
DV-HOP
al
go
rithm. Ho
p
co
unt is
1 of th
e un
kno
w
n
n
ode
can u
s
e th
e
informatio
n table which i
n
clu
de the v
a
lue of the
R
SSID and t
he average
path
length, with this can ea
sy to cal
c
ul
ate th
e distan
ce to
the anchor n
o
de.
The third ste
p
, use the t
r
ilateration i
n
formul
a(5
)
,
ca
lculate
the un
kno
w
n coo
r
di
nates of
unkno
wn
no
des,
after
un
kno
w
n
no
de
s use the
me
asu
r
ed
value
s
of j
u
mp
di
stan
ce
whi
c
h
is
unkno
wn no
de to ea
ch
anchor
nod
e. Then the
t
r
ilateral me
asurem
ent met
hod i
s
used
to
estimate the
unkno
wn
coo
r
din
a
tes
of it
s
own.
Known n
node
s co
ordin
a
tes were
（）
，
（
）
，
（
）
x1,y1
x
2,y2
x3,y
3
…
(
x
n,yn) in or
der
,
dis
t
ance fro
m
no
de
D, re
sp
ect
i
vely d1, d2,
d3,
..., dn, and it i
s
known that the co
ordinates of the node D is (x, y).
There are the
following formula:
2
2
2
2
1
2
1
2
1
)
(
)
(
)
(
n
n
n
d
y
y
x
x
d
y
y
x
x
(5)
Subtra
cting the last eq
uati
on from
the first sta
r
t equat
ions
were:
22
2
2
2
2
1
1
111
22
2
2
2
2
22
2
2
2
2
2
22
22
11
1
1
1
2(
)
2
(
)
2(
)
2
(
)
2(
)
2
(
)
nn
n
n
n
nn
n
n
n
n
n
n
n
nn
n
n
nn
xx
x
x
x
y
y
y
y
y
d
d
xx
x
x
x
y
y
y
y
y
d
d
x
x
x
x
x
yy
yy
y
d
d
(6)
Rep
r
e
s
ente
d
as a line
a
r eq
uation AX = b
,
Among them:
A=
)
(
2
)
(
2
)
(
2
)
(
2
1
1
1
1
n
n
n
n
n
y
y
x
x
y
y
x
x
n
(7)
b=
2
1
2
2
2
1
2
2
1
2
1
2
2
n
2
1
2
1
2
n
n
n
n
n
n
n
n
d
d
y
y
x
x
d
d
y
y
x
x
(8)
X=
y
x
(9)
Finally, use t
he mini
mum
mean
sq
ua
re
erro
r meth
o
d
to o
b
tain t
he
coo
r
din
a
tes
of the
node
s.
X=
b
A
A
A
T
T
1
(10)
3. The Simulation Results
In the matlab
environm
ent
for the propo
sed m
odified
DV-Hop al
go
rithm simulati
on, and
comp
arative analysi
s
o
n
t
he o
r
iginal
DV-Ho
p
alg
o
rit
h
m. Assumin
g
the a
c
tual
coordi
nate
s
of
the
unkno
wn
nod
e (xi,yi), e
s
ti
mates for the
nod
e
coo
r
di
n
a
tes fo
r
(x
r,y
r
), c
o
mmu
nic
a
tion r
adiu
s
i
s
the
R.
Absolute p
o
si
tioning erro
r shown in Form
ula (11
)
.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 10, Octobe
r 2014: 745
2
– 7458
7456
2
2
)
(
)
(
i
r
i
r
y
y
x
x
err
(11)
Relative po
sit
i
oning e
r
ror shown in Form
ula (12
)
.
R
err
error
n
i
(
1
2
)
3.1. Diffe
ren
t
Communica
tion Radiu
s
, Performan
c
e
Comparison
In orde
r to more
com
p
re
hen
sive anal
ysis an
d eva
l
uation of the
improved al
gorithm,
cha
nge th
e n
ode
com
m
un
ication
ra
diu
s
in turn. The
experi
m
ent
comp
are pe
rf
orma
nce un
d
e
r
different com
m
unication radiu
s
. As sh
own in
Figu
re 2, unde
r t
he environm
ent of 100 x
100
recta
ngul
ar
simulation ex
perim
ent, the numb
e
r
of node
s 2
00,
the an
cho
r
node
20 in
the
environ
ment.
Co
mmuni
cat
i
on radiu
s
f
r
o
m
10
sta
r
ts
sup
e
rp
ositio
n
in tu
rn, afte
r ea
ch
data
ru
n
indep
ende
ntly, and then
the data a
r
e
averag
ed.
Ran
g
ing
erro
r can b
e
cal
c
ulate
d
bet
ween
node
s.
It can be see
n
from the Figure 2, impro
v
ed
algorithm
when the co
mmuni
cation
radiu
s
is
small
relative
to the origi
n
al algo
rithm can great
ly improve the
po
sitioning a
c
curacy, after radi
us
is more than
20, the improved algo
rith
m has
n
o
t improve
d
alg
o
rithm pe
rformance. Beca
use
whe
n
R i
s
small, lo
cated
nod
es withi
n
the
scope
of
1-hop, l
o
cated n
ode
s
ca
n u
s
e little
or no
anchor n
ode
s. Introdu
cin
g
auxiliary RSSI ranging
at this time to ensure th
e
estimation e
rro
r
betwe
en un
known nod
e a
nd an
cho
r
no
de between i
s
sm
aller. Ho
wever,
with the increa
se
o
f
R
,
Average
hop
distan
ce i
n
n
e
twork
be
co
mes l
a
rg
er. L
o
cate
d no
de
s can
u
s
e m
o
re an
cho
r
n
o
d
e
s
within the sco
pe of 1-ho
p, Sensitive to
the ch
ang
e of the radiu
s
is
not so.
Therefore
wh
en the ra
dius is lesse
r
, the im
proved al
g
o
rithm can b
e
large
r
pe
rfo
r
man
c
e
improvem
ent.
Figure 2. Ran
g
ing Error u
n
der Diffe
rent
Comm
uni
cati
on Ra
diu
s
3.2. Diffe
ren
t
Number of
Anch
or Nod
es, the Algo
r
i
thm Perfor
mance Com
p
arison
Simulation
s a
r
e in the 1
0
0
×
10
0 network
environme
n
t, the total number of n
ode
s 200,
the comm
uni
cation radiu
s
of 20, cha
nging the
n
u
mbe
r
of an
cho
r
nod
es i
n
the netwo
rk.
Simulation e
x
perime
n
ts perfo
rman
ce
und
er
di
ffe
rent
numb
e
r of a
n
chor
node
s. An
d
the
prop
osed alg
o
rithm is me
asu
r
ing the
positio
n
information of the node an
d the actual n
ode
locatio
n
und
e
r
different co
mmuni
cation
radiu
s
. at
last
obtain the rel
a
tive positioni
ng error.
The simul
a
tion re
sults a
r
e sho
w
n in Figur
e 3, as the number of anchor n
ode
s to
improve, ori
g
i
nal algo
rithm
and improve
d
algor
ithm
can be gre
a
tly improved. When the an
ch
or
node
rea
c
h
e
s
a
ce
rtain p
e
rcentag
e, then imp
r
ov
e t
he num
be
r of
anchor nod
e
s
, the alg
o
rit
h
m
performance incr
eased
slowly, however,
the improved al
gorithm i
s
still better than the
DV-HOP
algorith
m
.
0
10
20
30
40
50
60
0
1
2
3
4
5
6
7
8
9
10
C
o
m
m
uni
c
a
t
i
o
n
r
a
di
us
/
m
R
ang
i
ng er
ror
/
m
DV
-
H
o
p
H
y
br
i
d
al
go
r
i
t
h
m
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
The Re
se
arch of DV-HOP Positionin
g
Algor
ithm
Base
d on RSSI Ca
libration
(Ji
e
Cao
)
7457
Figure 3. Different An
cho
r
Nod
e
s Po
sitioning Erro
r
3.3. Net
w
o
r
k
Location
Co
v
e
rage under the Differ
e
n
t Numb
er o
f
Anc
hor No
des
In the expe
ri
ment of p
o
siti
oning
cove
ra
ge, Se
n
s
or n
ode
s a
r
e
ran
domly di
strib
u
ted in th
e
environ
ment of 100×1
00, node nu
mbe
r
chang
es
b
e
twee
n 100
-20
0
.The numb
e
r
of ancho
r n
ode
s
is
20
.N
od
e
co
mmu
n
i
c
a
tion
ra
d
i
us
is
10
. T
h
e n
o
de
distrib
u
tion
is sh
own in
Fi
gure
4, i
m
pro
v
ed
algorith
m
co
mpared with
DV-Hop al
gorithm pos
itioni
ng cove
rag
e
as sho
w
n in
Figure 5.
Figure 4. The
Rand
om No
de Di
stributio
n
Figure 5. Positioning Coverage un
der a
Different Nu
mber of Node
s
As ca
n be
se
en in Figu
re
5, with the in
cre
a
se
of nu
mber of n
ode
s, po
sitioning
coverage
also
will b
e
in
cre
a
si
ng. Be
cause covera
g
e
ca
us
ed by
node
den
sity. The g
r
eate
r
t
he de
nsity, the
greate
r
the p
o
sitioni
ng coverag
e
, be
cau
s
e this im
pro
v
ed algo
rithm
according to
the hop
cou
n
t is
1 node ad
opt auxiliary RSS
I
ranging. Ho
p cou
n
t
is 1 of the unkno
wn node
s whi
c
h need only o
ne
anchor no
de
informatio
n can be
po
sitio
ned. While
traditional
DV-Hop
algo
rith
m to get at l
east
three
an
cho
r
node
informa
t
ion to
solve
the un
kn
own
nod
e lo
catio
n
. It is
evide
n
t to
see
in t
he
numbe
r of n
ode
s is le
ss
than 90, the
den
sity
of node in net
work nod
e a
r
e rare F
r
om fig
u
re
(5).T
hen
aux
iliary p
o
sitio
n
ing to
the
unkno
wn
no
de
whi
c
h
ho
p count
is
1 by th
e
RSSI
techn
o
logy. It can re
du
ce
relian
c
e o
n
anchor
node
s. Positioni
n
g
cove
rag
e
wa
s sig
n
ifica
n
tly
better than DV-Ho
p
algo
rithm.
4. Conclusio
n
In re
cent yea
r
s,
schola
r
s
prop
osed m
a
ny im
prove
d
algorith
m
To
the lack of
DV-HOP
algorith
m
. Th
e re
sea
r
ch of 1 hop dista
n
c
e e
rro
r in th
e DV-HOP al
gorithm i
s
no
t a lot. In
this
pape
r, the i
m
prove
d
alg
o
rithm i
s
p
r
o
posed in
ord
e
r to
red
u
ce
this effe
ct. By introdu
cin
g
the
auxiliary RSSI
ranging, reduce
the estim
a
tion
di
st
ance error
betwe
en nodes. At
last achi
eve t
he
0
10
20
30
40
50
60
0
0.
05
0.
1
0.
15
0.
2
0.
25
0.
3
0.
35
0.
4
0.
45
0.
5
个
A
n
c
h
o
r
no
de nu
m
b
e
r
/
R
e
l
a
t
i
v
e
po
s
i
t
i
on
i
n
g e
r
r
o
r
DV
-
H
o
p
H
y
br
i
d
al
go
r
i
t
h
m
0
10
20
30
40
50
60
70
80
90
100
0
10
20
30
40
50
60
70
80
90
100
* R
e
d
i
s
t
h
e
be
a
c
o
n
n
o
d
e
s
.
B
l
a
c
k
i
s
t
h
e
un
k
n
o
w
n
no
de
0
20
40
60
80
10
0
12
0
14
0
16
0
180
200
0
20
40
60
80
100
120
N
u
m
b
er o
f
no
des
P
o
s
i
t
i
on
i
ng c
o
v
e
ra
ge/
%
DV
-Ho
p
H
y
bri
d
a
l
gori
t
h
m
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 10, Octobe
r 2014: 745
2
– 7458
7458
goal of
eventually re
duce
the po
si
tionin
g
erro
r. The
simulatio
n
re
sults
sh
ow, t
he alg
o
rithm
did
not
in
crea
se the
am
ount o
f
cal
c
ulatio
n and ha
rd
ware eq
uipme
n
t. The
imp
r
ove
d
alg
o
rithm
Can
better improve the positio
n
i
ng perfo
rma
n
ce. Stay
alg
o
rithm
s
are d
i
scusse
d on the ba
sis of the
simulatio
n
experim
ents in
this pap
er. In
the future How to implem
ent in pra
c
tical system, an
d
test the
algo
rithm p
e
rfo
r
mance in
the
real
env
i
r
on
ment is al
so
an
urg
ent n
eed to
solve
the
probl
em.
Ackn
o
w
l
e
dg
ements
This work was supp
orte
d by the National Natu
ral Scien
c
e
Found
ation
of China
(612
630
31)
Referen
ces
[1]
KIM S, KO JG,
YOON J, et al.
Multipl
e
-o
bjec
tive
foe pl
acin
g multi
p
le
b
a
s
e
statio
n i
n
w
i
reless s
ens
o
r
netw
o
rk.
Proce
edi
ng of th
e 2r
d Internati
o
n
a
l
S
y
mp
osi
u
m o
n
W
i
reless P
e
r
v
asive C
o
mp
uting. Kor
ea,
200
7: 627-
631
[2]
Jenn
ifer Yick, Bis
w
a
nnath M
u
kher
jee
，
Dip
ak Ghosa
l
. W
i
reless se
nsor
n
e
t
w
o
r
k.
Co
mp
u
t
er Netw
orks
.
200
8; 52(1
2
): 2292-
223
0
．
[3] Mizugaki
K
，
Fu
j
i
w
a
r
a
R
，
N
a
kaga
w
a
T
，
et al.
Accur
a
te wireless
loc
a
tion/
communic
a
tion system with
22-c
m
error us
ing UW
B- IR
. Radi
o an
d W
i
reless S
y
mp
os
ium
．
Washi
ngt
on. DC. IEEE. 2007: 4
55-
458
．
[4] F
A
NG
Z
heng
，
Z
H
AO Z
han
.Rang
ing
an
al
ysis
bas
ed
on
RSSI.
Chi
n
e
s
e Jour
na
l of
Sens
ors a
n
d
Actuators
. 200
7; 22(11): 2
526
-253
0
．
[5] WAN
X
i
nsheng
，
ZHAO Yanjing
，
LI H
a
it
ao R
e
searc
h
on impr
ove
m
ent of al
go
rithm DV-Ho
p
.
Co
mp
uter Scie
nce
. 201
1; 38(
2): 76- 78.
[6] LI
M
，
LIU YH.
Ren
dere
d
p
a
t
h
: Ran
ge- fre
e
loca
lizati
on
in
anis
o
tropic
se
nsor n
e
t
w
orks
w
i
t
h
h
o
l
e
s.
IEEE ACM Tra
n
sactions on Networking
. 20
1
0
; 18(1): 32
0- 332.
[7]
MA Xia
o
x
i
ao, Z
H
AO Anjun, MA Guagsi, R
e
sear
c
h
on l
o
calizati
on tech
nol
og
y of
w
i
re
less sens
o
r
net
w
o
rks.
Jour
nal of Co
mpute
r
Applic
ations
.
201
0; S1(6): 2
4
-26.
[8]
Guoqi
an
g Ma
oBaris, F
i
d
a
n
Loca
lizati
on
al
gorithm
s an
d strategies
for w
i
rel
e
ss
se
ns
or
net
w
o
rks.
Information Sci
ence R
e
fere
nc
e, Rele
ase Dat
e
. 2009.
[9]
Yassine F, Saf
ah.
A hybrid D
V
-Hop for loca
l
i
z
a
t
i
o
n
in larg
e scale w
i
reless
sensor n
e
tw
ork
. Proceedi
n
g
of the
6th Inter
natio
nal
C
onfe
r
ence
on
Mo
bil
e
T
e
c
hnol
og
y. Appl
icatio
n an
d
S
y
stems. N
e
w
Y
o
r
k: ACM
.
200
9; 1-6
．
[10]
Shan
g Y, R
u
m
l
W
,
Z
hang
Y, et al.
Locali
z
a
tion
from
mere connectivity.
Mobil
e
H
o
c, An
n
apo
lis, MD
,
USA. 2003.
[11]
He T
,
Huan
g
C, Blum
B M,
et al.
Ra
nge-fr
ee
loca
li
z
a
ti
on
sche
m
es for
l
a
rge
scal
e
s
e
n
s
or n
e
tw
orks.
MobiC
o
m, San
Di-eg
o, CA, USA. 2003.
[12]
YUAN Z
h
eng
wu, LIANG J
unj
un. Mu
lti-ho
ps
local
i
zati
on
al
g
o
rithm
base
d
on v
i
rtual
force
for
w
i
rel
e
s
s
sensor
net
w
o
rk
s. Journ
a
l
of C
hon
gqi
ng
Un
iv
ersit
y
of P
o
sts and
T
e
lecomm
unic
a
tions:
Nat
u
ral Sc
ie
nce
Editio
n. 201
0; 22(1): 12
2-1
2
6
.
[13]
Z
H
ANG Jia,
W
U
Yanh
ai, S
H
I F
eng. L
o
ca
lizatio
n
alg
o
rit
h
m bas
ed
on
DV-Hop
for
w
i
reless s
ens
or
net
w
o
rks. Jour
nal of comp
ute
r
A
pplic
ations.
201
0; 30(2): 32
3-32
6.
Evaluation Warning : The document was created with Spire.PDF for Python.