Internati
o
nal
Journal of Ele
c
trical
and Computer
Engineering
(IJE
CE)
V
o
l.
6, N
o
. 4
,
A
ugu
st
2016
, pp
. 15
70
~
1
576
I
S
SN
: 208
8-8
7
0
8
,
D
O
I
:
10.115
91
/ij
ece.v6
i
4.1
001
7
1
570
Jo
urn
a
l
h
o
me
pa
ge
: h
ttp
://iaesjo
u
r
na
l.com/
o
n
lin
e/ind
e
x.ph
p
/
IJECE
Implementation of Fuzzy Based
Simulation
for Clon
e Det
e
cti
o
n
in Wirel
ess S
e
ns
or Networks
M
a
n
j
un
at
ha
R
.
C
.
1
,
Re
kh
a K. R.
2
, Na
t
r
aj K
.
R
.
2
1
Department of Electronics
and
Communication Engineering,
Resear
ch s
c
holar
, J
a
in Univ
ers
i
t
y
2
Department of Electronics
an
d
Communication Engineering,
SJB Instutute of
Technolog
y
Bangalor
e
,
Karn
atak
a,
India
Article Info
A
B
STRAC
T
Article histo
r
y:
Received
Ja
n 29, 2016
Rev
i
sed
Ap
r 2, 20
16
Accepted Apr 14, 2016
Wireless sensor networks ar
e
usuall
y
left un
attended
and s
e
rve hostile
environment, th
erefore can easily be
compromise
d. With compromised nodes
an attack
er can
conduct sev
e
ral
inside
and outside attacks.
Node replicatio
n
att
ack is
one
of
them
which c
a
n
caus
e
sev
e
re
damage to
wirele
ss se
nsor
network if lef
t
undetected
. This pa
per presents fuzzy
b
a
sed simulation
framework for detection
and revocati
on of compromised nodes in wireless
sensor network. Our proposed scheme
uses PDR statistics
an
d neighbor
reports to d
e
t
e
r
m
ine the prob
ab
ilit
y
of
a
clust
e
r
being
com
p
romised. Nodes
in compromised cluster ar
e then re
voked and software attestation is
performed.Simulation
is carr
i
ed o
u
t on MATLAB 2010a and p
e
rf
ormance o
f
proposed scheme is compared
with convention
a
l algor
ithms on the basis of
communication
and storage ov
er
head. Si
mulation
results show that proposed
scheme requir
e
less comm
unication and storag
e
overhead
than
convention
a
l
algorithms.
Keyword:
Clu
s
ter
Fuzzy logic
Rep
lica no
d
e
detectio
n
Trust aggre
g
at
or
W
i
rel
e
ss se
ns
o
r
net
w
or
ks
Copyright ©
201
6 Institut
e
o
f
Ad
vanced
Engin
eer
ing and S
c
i
e
nce.
All rights re
se
rve
d
.
Co
rresp
ond
i
ng
Autho
r
:
Natara
j K
R,
Proffesor a
n
d
Head,
Dep
a
rtm
e
n
t
o
f
ECE, SJB
In
stitu
te of Techn
o
l
o
g
y
,
Bangal
o
re, Ka
rnataka, India.
Em
a
il: Nataraj
.
sjb
it@g
m
ail.co
m
1.
INTRODUCTION
Security is o
n
e
o
f
prim
e
o
b
j
ective wh
ile d
e
sign
ing any wireless sensor net
w
ork architecture,
esp
ecially wh
en
sen
s
o
r
n
e
two
r
k
is ex
po
sed
to
h
o
stile
e
n
v
i
ron
m
en
t. In
m
a
n
y
o
f
wireless sen
s
or network
applications such as military
ope
rations an
adve
rsa
r
y can capture an
y node a
nd
gain
access to encryption
keys.
Once e
n
cryption
keys
are e
x
tracted a
dve
rsa
r
y can c
r
eate as m
a
ny as re
plica node
s
and
de
ploy t
h
em
a
t
d
e
sir
e
d
lo
cation
s
in
th
e
n
e
two
r
k
.
Th
is typ
e
o
f
attack
is kn
own
as no
d
e
r
e
p
lica attack
an
d
f
a
lls under
th
e
category of ins
i
de attacks.
Node re
p
lica attack ca
n cause s
e
vere
dam
a
ge to
the system
if left undetected.
As
t
h
ese re
pl
i
ca n
ode
s gai
n
t
h
e t
r
ust
of
nei
g
h
b
o
u
r
h
oo
d
no
des t
h
ey
can l
a
unc
h
a veri
t
y
o
f
at
t
acks i
n
cl
udi
ng
bl
ac
k
hole attack,
worm
hole attack, false
da
ta inj
ectio
n, can
d
i
v
e
rt n
e
t
w
ork
tr
affic towards t
h
e attacker, ca
n leak
secret inform
ation t
o
the
attacker etc.
Th
e m
a
in
proble
m
in
th
e
d
e
t
ectio
n
o
f
rep
licatio
n
a
ttack resides in the
resource
scarcit
y
of se
nsor
n
e
two
r
k
.
To
effectiv
ely d
e
tect th
e rep
e
titive u
s
e
o
f
sam
e
secret k
e
y
n
e
twork-wi
d
e
com
p
ariso
n
of locatio
n
depe
n
d
ent
aut
h
ent
i
cat
i
o
n i
n
f
o
rm
at
i
on i
s
re
qui
red
.
B
u
t
l
i
m
i
t
e
d
m
e
m
o
ry
an
d
p
o
we
r s
u
ppl
y
put
rest
ri
ct
i
ons
o
n
t
h
e am
ount
o
f
aut
h
e
n
t
i
cat
i
on
i
n
f
o
rm
at
i
on st
ore
d
a
nd e
x
c
h
a
nge
d
wi
t
h
i
n
t
h
e net
w
or
k.
He
nce ene
r
gy
effi
ci
ency
,
less sto
r
ag
e and
co
mm
u
n
i
cati
o
n
o
v
e
rh
ead
will b
e
th
e k
e
y issu
es in
d
e
cid
i
n
g
u
tility o
f
th
e alg
o
r
ith
m
.
Nod
e
R
e
pl
i
cat
i
on
at
t
ack has dra
w
n
i
n
t
e
rest
of
several
re
sea
r
chers
si
nce
l
a
st
deca
de,
p
r
o
t
ocol
s
f
o
r
det
ect
i
ng
replication attack a
r
e categorized as ce
ntral
i
zed and
di
st
ri
but
e
d
det
ect
i
o
n pr
ot
oc
ol
s.
C
e
nt
ral
i
zed det
e
ct
i
on
protoc
ols suc
h
as Random
ized key pr
e
-
distribution [1] and SET [2] use ba
se station as centralise cont
rolling
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
Im
pl
eme
n
t
a
t
i
o
n
of
F
u
zzy B
a
s
e
d
Si
m
u
l
a
t
i
o
n f
o
r C
l
o
n
e
Det
e
c
t
i
on i
n
Wi
rel
e
s
s
Se
ns
or .
...
(
M
anj
un
at
h
a
R. C
.
)
1
571
au
tho
r
ity wh
ile d
i
stribu
ted
detectio
n
techn
i
q
u
e
s su
ch as
Determ
in
istic
m
u
l
ticast [3
],
Ran
d
o
m
ized
an
d Lin
e
selected m
u
lticast [4], RE
D [5] a
nd
l
o
cal
ized m
u
lticast [6]
uses
witness base
d a
p
proac
h
for clone node
d
e
tectio
n.
Existing
detection techniques
sh
ow a tra
d
e-off bet
w
een
detecti
on accura
cy and c
o
mm
u
n
ication
or
sto
r
ag
e
o
v
e
r
h
ead
.
Th
er
efo
r
e
th
is p
a
p
e
r
pr
esen
ts a fu
zzy
based
ar
ch
itectur
e fo
r d
e
tection
o
f
cl
o
n
e
nodes in
wi
rel
e
ss sen
s
o
r
net
w
or
k.
Ou
r
pr
op
ose
d
sche
m
e
i
s
ext
e
nsi
on o
f
w
o
r
k
p
r
es
ent
e
d
by
Geet
ha et
al
[7]
and us
e
s
p
ack
et deliv
ery ratio
[PDR], tru
s
t v
a
lu
es calcu
l
ated
b
y
rep
o
rting
and
n
e
ig
hb
ouring
cluster to
d
e
tect rep
lica
no
des i
n
a cl
us
t
e
r base
d scen
ari
o
. R
e
st
o
f
t
h
i
s
paper i
s
ar
ra
nge
d as f
o
l
l
o
w
s
:
Sect
i
on –
II
di
scuss
net
w
or
k an
d
t
h
reat
m
odel
,
i
n
sect
i
o
n-
II
I
pr
op
ose
d
fuzzy
base
d re
pl
i
ca n
ode
det
ect
i
o
n
(
F
R
N
D
)
pr
ot
oc
ol
i
s
gi
ve
n.
Se
ct
i
on-
IV
d
i
scu
ss sim
u
latio
n
resu
lts
an
d fi
n
a
lly sectio
n-V con
c
ludes th
is
p
a
p
e
r.
2.
NETWO
R
K
AN
D TH
RE
A
T
MO
DEL
2.
1.
Netw
o
r
k M
o
d
e
l
C
onsi
d
er
a
wi
r
e
l
e
ss sens
o
r
ne
t
w
o
r
k
wi
t
h
n
o
d
e
s
un
if
or
m
l
y
d
i
str
i
bu
ted in
an
ar
ea
o
f
100
x1
00
m
e
ter
sq
uares i
n
a
h
o
s
tile env
i
ron
m
en
t. Network
fo
llows a
clu
s
ter b
a
sed
arch
itecture as sh
own
in
Fi
g
u
re
1
,
furthe
rm
ore the network use
s
locali
zation protoc
ol and each
node
knows its location. Nodes are st
ationary
aft
e
r
de
pl
oy
m
e
nt
an
d
t
i
e
d
wi
t
h
R
S
A
base
d
pu
bl
i
c
key
cry
p
t
o
sy
st
em
.B
ase st
at
i
on i
s
ce
nt
ral
a
n
d
c
ont
r
o
l
l
i
n
g
authority whic
h is responsi
b
l
e
for all routing related ta
sk, furt
herm
ore at a fixed tim
e
interval each c
l
uster
sen
d
t
r
ust
re
p
o
r
t
t
o
base
st
at
i
on
whe
r
e
FR
N
D
p
r
ot
ocol
i
s
e
v
a
l
uat
e
d.
2.
2.
Threa
t Mo
del
Let
us
ass
u
m
e
t
h
at
t
h
e
at
t
ack
has a
part
i
a
l
co
nt
r
o
l
o
v
e
r
t
h
e
depl
oy
m
e
nt re
gi
o
n
a
n
d
m
a
y
capt
u
re
a
sub
s
et
of avai
l
a
bl
e no
des.
Af
t
e
r gai
n
i
n
g acc
ess ove
r secret
key
s
, at
t
acker
m
a
y l
a
unch v
a
ri
o
u
s i
n
si
de a
t
t
acks
through c
o
m
p
rom
i
sed nodes
.
Furt
herm
ore it
is also as
su
m
e
d that
each c
o
m
p
romised node is
surrounde
d
by at
least one
legitim
ate node
[4].
Fi
gu
re
1.
Net
w
or
k a
rra
n
g
em
ent
3.
FUZ
Z
Y
BASED REPLICA NODE
DETECTION SCHE
ME (FRNDS)
Fuzzy
base
d r
e
pl
i
ca no
de de
t
ect
i
on (FR
N
D
)
pr
ot
oc
ol
ad
o
p
t
s
a regi
o
n
ba
sed ap
pr
oac
h
f
o
r det
ect
i
o
n
of c
o
m
p
rom
i
sed n
ode
s o
p
erat
i
ng i
n
t
h
e e
nvi
ro
nm
ent
.
The al
go
ri
t
h
m
di
vides t
h
e net
w
o
r
k area i
n
t
o
a n
u
m
b
er
of re
gions; wit
h
each re
gion
has
a clusterhead node
with som
e
co
mm
o
n
node
s sha
r
ing betwee
n the
other
regi
ons
.T
he algorithm
relies
on t
r
ust
value
for each cluste
r and
detects th
e cluste
r trustworthi
n
ess
bas
e
d
on
th
e clu
s
ter tru
s
t v
a
lu
e. On
ce a clu
s
ter is
flagg
e
d to
b
e
unt
ru
st
wo
rt
hy
, s
o
ft
ware m
o
d
u
l
e
s
of al
l
t
h
e
se
nso
r
n
o
d
e
s
bel
o
ngi
ng
t
o
t
h
at
cl
ust
e
r i
s
t
e
st
ed by
t
h
e n
e
t
w
o
r
k
o
p
erat
or
fol
l
owe
d
by
t
h
e
det
ect
i
on a
n
d re
v
o
ca
t
i
on
of
com
p
rom
i
sed no
des i
n
t
h
at
c
l
ust
e
r. A si
m
p
l
e
app
r
oac
h
f
o
r
unt
rust
wo
rt
hy
cl
ust
e
r det
ect
i
o
n m
i
ght
be bas
e
d o
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l. 6
,
N
o
. 4
,
Au
gu
st 2
016
:
15
70
–
1
576
1
572
com
p
ari
n
g
a s
i
ngl
e t
r
ust
val
u
e
wi
t
h
a t
h
r
e
sh
ol
d;
ho
we
ver wit
h
th
is ap
pro
ach an error in clu
s
ter tru
s
t
calcu
latio
n
will d
i
rectly affect th
e
o
u
t
p
u
t
of th
e al
g
o
rith
m
.
To m
i
n
i
mize th
e im
p
act o
f
su
ch
erro
rs FRND
pr
ot
oc
ol
u
s
es
m
u
lt
i
p
l
e
t
r
ust
val
u
es a
n
d
pac
k
et
del
i
v
e
r
y
ra
t
i
o
t
o
deci
de
whet
her t
h
e cl
ust
e
r i
s
t
r
ust
w
ort
h
y
o
r
n
o
t
. Mu
ltip
le
tru
s
t
v
a
lu
es are co
llected
fro
m
tru
s
t a
ggreg
at
o
r
p
r
esent in
sam
e
clu
s
ter as
well as th
e
ove
rl
ap
pi
n
g
n
ode
s o
f
t
h
e
n
e
i
g
h
b
o
u
r
i
n
g cl
ust
e
r.
Fuzzy
base
d ap
pr
oac
h
i
s
ap
pl
i
e
d t
o
com
p
r
o
m
i
se no
de
detection and
revocation as
follows: eac
h
node in a cl
uste
r
i
s
act
as a t
r
ust
ag
gre
g
at
o
r
i
n
ro
u
n
d
r
obi
n m
a
nne
r
.
In each tim
e span, t
h
e t
r
ust
aggre
g
ator c
o
m
putes trust
va
lue
and pac
k
et delivery ratio for
i
t
s
cluster
and report
it to
th
e b
a
se statio
n
.
Th
e base statio
n
th
an
p
e
rfo
r
m
FRND
p
r
o
t
o
c
o
l
to
ev
al
u
a
te cluster’s tru
s
two
r
th
in
ess;
o
n
c
e a cl
u
s
ter
is d
ecid
e
d to
be un
tru
s
two
r
t
hy, th
e
n
e
twork
o
p
e
rator
p
e
rfo
rm
s so
ftware attestatio
n
s
ag
ain
s
t all
sens
or
n
o
d
e
s t
o
det
ect
an
d
re
vo
ke t
h
e c
o
m
p
rom
i
sed n
ode
s
i
n
t
h
at
cl
ust
e
r.
The
det
a
i
l
e
d d
e
scri
pt
i
o
n
of
f
u
zzy
base
d r
e
p
l
i
ca nod
e det
e
c
t
i
on p
r
ot
oc
ol
i
s
gi
ve
n as
fol
l
ows:
P
r
i
o
r t
o
the de
ploym
e
nt, each node in the
networks
is allotted a
unique ID a
nd
network is
di
vided int
o
number
of
o
v
e
rlapp
i
ng
cl
u
s
ters. Co
mm
u
n
i
catio
n
co
st
of th
e system
will d
e
p
e
nd
en
t on
clu
s
ter size,
alth
o
ugh
th
ere
is n
o
restrictio
n ov
er th
e size and th
e sh
ap
e of th
e clu
s
ter
bu
t an
i
n
crease in
clu
s
te
r size will in
crease in
tra
co
mm
u
n
i
catio
n
clu
s
ter co
st
as th
e lo
cal trust repo
rt will
req
u
i
re m
o
re ho
p
e
s to
reach
at th
e tru
s
t aggreg
at
o
r
.
Wh
ile
k
eep
i
n
g clu
s
ter size small i
t
will b
e
d
i
fficu
lt to
d
e
t
ect co
m
p
ro
m
i
s
e
n
o
d
e
s. Fu
rt
herm
o
r
e, secret
k
e
yin
g
m
a
t
e
ri
al
i
s
prel
oade
d i
n
t
o
eac
h se
ns
or
n
ode
f
o
r
pai
r
wi
se
ke
y
est
a
bl
i
s
hm
ent
by
base
st
at
i
o
n [
8
]
,
[
9
]
.
T
h
e e
n
t
i
r
e
process
can be
descri
bed in t
h
ree steps.
3.
1.
Cl
uster
f
o
rm
a
t
i
o
n
an
d T
r
us
t
ag
gre
g
a
t
or
s
e
l
ecti
o
n
After d
e
p
l
o
y
men
t
, each
no
d
e
d
e
term
in
es its lo
cation
an
d fin
d
s
ou
t th
e cl
u
s
ter t
o
wh
ich it b
e
long
s,
th
is clu
s
ter is
referred to
as
ho
m
e
clu
s
ter to
th
e nod
e
wh
ile o
t
h
e
r cl
u
s
ters
will b
e
foreign clu
s
ters. Th
e
sen
s
o
r
no
de t
h
e
n
di
sc
ove
rs t
h
e
I
D
o
f
al
l
t
h
e nei
g
h
b
o
u
r
i
n
g n
o
d
es
i
n
i
t
s
hom
e cl
ust
e
r an
d est
a
bl
i
s
hes
pai
r
wi
se
secret
keys
with the
m
. Selection of trust a
g
gre
g
at
or is done
t
h
en in a
round
robin m
a
nner
a
s
follows: each cluster
i
s
associated wit
h
a
series
of time slots;
i
n
a
p
s
eu
do
ra
nd
om
or
der
eac
h
no
d
e
deci
des i
t
s
d
u
t
y
t
i
m
e
sl
ot
an
d act
as
a trust aggregator. T
h
ese trust aggre
g
ator nodes a
r
e re
spo
n
si
bl
e f
o
r
t
h
e sen
d
i
n
g t
r
ust
re
po
rt
s an
d P
D
R
ch
aracteristics to
th
e
b
a
se statio
n
.
3.
2.
Trust c
a
lculation
and
F
o
rwardin
g
In each tim
e
interval
, neighbourhood trust is co
m
puted by each cluster
in
ev
er
y n
ode.
Neigh
bou
rho
o
d
trust is d
e
fi
n
e
d
as t
h
e d
i
fferen
ce b
e
t
w
een
th
e pro
b
a
b
i
lity d
i
strib
u
tio
n
s
of th
e informatio
n
g
e
n
e
r
a
ted
an
d in
f
o
r
m
at
io
n
sen
t
to
th
e n
ode in
co
n
s
i
d
er
atio
n
b
y
its n
e
ig
hbo
ur
ing
node in
cu
r
r
e
n
t
clu
s
ter
.
Neigh
bou
ri
n
g
tru
s
t is related to
th
e au
th
enticit
y o
f
no
d
e
an
d
it in
creases with
dat
a
t
r
a
n
sm
i
ssi
on bet
w
ee
n
n
e
igh
bou
ri
n
g
n
o
d
e
s. Th
e t
r
u
s
t info
rm
atio
n
can
also
b
e
tran
sm
itted
t
o
th
e
b
a
se statio
n
b
y
t
h
e
nod
es
o
f
nei
g
hb
o
u
ri
ng
c
l
ust
e
r
whi
c
h a
r
e o
n
e
ho
p a
w
a
y
fr
om
t
h
e cur
r
e
nt
cl
ust
e
r
.
T
h
e arra
n
g
em
ent
i
s
gi
ve
n i
n
Fi
g
u
re
2
.
Fi
gu
re
2.
C
l
ust
e
red
net
w
o
r
k
wi
t
h
no
des
o
v
e
r
l
a
p
p
i
n
g C
l
ust
e
rs [
1
0]
3.
3.
Com
p
ro
mise No
de Detec
t
io
n
a
nd Re
voc
at
ion
Once
a cl
uster-trust statem
en
t is receive
d
at the
base
station
by the
trust
aggre
g
ator
node of c
u
rrent
clu
s
ter; firstly its au
th
en
ticity an
d
th
e
freshness o
f
the re
p
o
rt is ch
eck
ed
at th
e b
a
se statio
n. For au
th
enticit
y
secret
key
sha
r
ed bet
w
een t
h
e
base st
at
i
on a
nd t
r
ust
ag
gre
g
ator is chec
ked whe
r
eas
f
o
r f
r
e
sh
ness o
f
the
rep
o
rt
ti
m
e
r asso
ciated
with
it is ch
eck
e
d. Un
au
th
en
tic o
r
ex
p
i
re
d reports are
dis
carde
d by the
base station. For the
detection of c
o
m
p
romised trust aggre
g
ator, t
h
e
base sta
tion m
a
in
tain
s th
e
record of each tru
s
t agg
r
eg
at
o
r
b
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
Im
pl
eme
n
t
a
t
i
o
n
of
F
u
zzy B
a
s
e
d
Si
m
u
l
a
t
i
o
n f
o
r C
l
o
n
e
Det
e
c
t
i
on i
n
Wi
rel
e
s
s
Se
ns
or .
...
(
M
anj
un
at
h
a
R. C
.
)
1
573
b
i
nd
ing
its ID
to
its ho
m
e
clu
s
ter.
Th
is will
p
r
ev
en
t a co
m
p
ro
m
i
sed
tru
s
t
ag
greg
ato
r
from
c
l
ai
min
g
mu
ltip
le
hom
e clustersa
n
d launchi
n
g replay attack
with fa
ke cl
uster-trust statem
ents.
To
h
a
nd
le
th
e n
on-lin
earity
asso
ciated
with
th
e
prob
l
e
m
a
fuzzy
based
ap
pr
oac
h
i
s
p
r
ese
n
t
e
d i
n
t
h
i
s
researc
h
.
FIS a
r
chi
t
ect
u
r
e i
s
p
r
o
p
o
sed
fo
r t
h
e det
ect
i
on
of
unt
rust
wo
rt
hy
cl
ust
e
r o
n
t
h
e
basi
s o
f
t
r
ust
r
e
po
rt
s
fro
m
sa
m
e
an
d
n
e
igh
bou
ri
ng
cluster and p
a
ck
et
d
e
livery ratio (PDR) statistics of th
e clu
s
ter
u
n
d
e
r
co
nsid
eration.
Clu
s
ter form
at
i
o
n is m
o
re efficient usi
n
g fuz
z
y logic
[11].
The a
r
chi
t
ect
u
r
e
of
p
r
o
p
o
se
d sy
st
em
i
s
gi
ven
i
n
Fi
g
u
re
3,
wi
t
h
t
r
ust
re
po
rt
f
r
om
cl
ust
e
r
un
de
r
co
nsid
eration an
d its immed
i
ate n
e
ighb
ou
r an
d p
a
ck
et
d
e
livery ratio b
e
i
n
g
th
e inpu
t to
t
h
e syste
m
. Th
e
ou
tpu
t
o
f
t
h
e system
i
s
th
e prob
ab
ility th
at th
e clu
s
ter is tru
s
t
w
ort
h
y or no
t. Th
e
p
r
op
o
s
ed
FIS stru
cture is b
a
sed
on
th
e set of ru
les g
i
v
e
n
i
n
Table 1
.
Based
o
n
th
e ru
le set th
e
p
r
o
b
a
b
ility o
f
a clu
s
ter
b
e
in
g
un
trustwo
r
th
y is
calcu
lated
,
if d
e
tected
un
tru
s
t
w
orth
y software atte
st
at
i
on i
s
pe
rf
or
m
e
d o
v
er t
h
e
n
ode
o
f
cl
u
s
t
e
r i
n
consideration.
Fi
gu
re 3.
F
u
zz
y
l
ogi
c base
d R
e
pl
i
ca
N
ode
Det
ect
i
on Sc
he
m
e
(F.R
.
N
.
D
.S
.)
Tabl
e
1. R
u
l
e
s
e
t
fo
r
pr
o
pose
d
sy
st
em
S. No.
TA Report fro
m
sa
m
e
clus
ter
TA Report fro
m
neighbouring clus
ter
PDR S
t
atisti
cs
Cluster
Trustw
o
r
thiness
1
Lo
w
Lo
w
Lo
w
Lo
w
2
Lo
w
Lo
w
Med
i
u
m
Lo
w
3 Low
Low
High
Low
4
Lo
w
Med
i
u
m
Lo
w
Lo
w
5 L
o
w
M
e
diu
m
M
e
diu
m
M
e
diu
m
6 L
o
w
M
e
diu
m
High
M
e
diu
m
7 L
o
w
High
L
o
w
M
e
diu
m
8 L
o
w
High
M
e
diu
m
M
e
diu
m
9 L
o
w
High
High
M
e
diu
m
1
0
Med
i
u
m
Lo
w
Lo
w
Lo
w
11
M
e
diu
m
L
o
w
M
e
diu
m
L
o
w
12
M
e
diu
m
L
o
w
High
M
e
diu
m
13
M
e
diu
m
M
e
diu
m
L
o
w
L
o
w
14
M
e
diu
m
M
e
diu
m
M
e
diu
m
M
e
diu
m
15
M
e
diu
m
M
e
diu
m
High
M
e
diu
m
16
M
e
diu
m
High
L
o
w
M
e
diu
m
17
M
e
diu
m
High
M
e
diu
m
M
e
diu
m
18
M
e
diu
m
High
High
High
19 High
Low
Low
Low
20
High
L
o
w
M
e
diu
m
L
o
w
21
High
L
o
w
High
M
e
diu
m
22
High
M
e
diu
m
L
o
w
L
o
w
23
High
M
e
diu
m
M
e
diu
m
M
e
diu
m
24
High
M
e
diu
m
High
M
e
diu
m
25
High
High
L
o
w
M
e
diu
m
26
High
High
M
e
diu
m
High
27
High
High
High
High
TA Report
fro
m
sa
m
e
cluster
TA Report
fro
m
neighbouri
ng cluster
PDR
Statistics
F.R.N.D.S
.
Com
p
ro
mise
Cluster
Detec
t
io
n
Fuzz
if
ier
Inferenc
e Engine
D
e
f
u
zzif
ier
F
u
zzy
R
u
le
Ba
s
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l. 6
,
N
o
. 4
,
Au
gu
st 2
016
:
15
70
–
1
576
1
574
4.
SIMULATION RESULTS
To eval
uat
e
t
h
e perf
orm
a
nce
of p
r
o
p
o
sed f
u
zzy
base
d re
pl
i
ca no
de det
ect
i
on p
r
ot
oco
l
, M
A
TLAB
base
d fram
e
wor
k
ha
s bee
n
prese
n
t
e
d
.
Th
e
per
f
o
r
m
a
nce cri
t
e
ri
on i
s
set
t
o
com
m
uni
cat
i
on o
v
er
hea
d
and
st
ora
g
e
ov
er
he
ad. C
o
m
m
uni
cat
i
on
o
v
er
head
bei
n
g
t
h
e
n
u
m
b
er of m
e
ssage
s transm
i
tted and storage
ove
rhea
d
being the m
e
mory re
quired
by each node
, L
e
t
bei
ng t
h
e
n
u
m
b
er of
no
de
s prese
n
t
e
d i
n
t
h
e net
w
o
r
k
,
bei
n
g
av
erag
e d
e
g
r
ee o
f
n
e
i
g
hbo
urhoo
d, p
b
e
i
n
g
th
e prob
ab
ility o
f
clu
s
terhead
electio
n
an
d
and
b
e
in
g th
e
num
ber o
f
wi
t
n
ess n
o
d
es an
d
num
ber of cl
u
s
t
e
rhea
d re
po
rt
i
ng t
o
bas
e
st
at
i
on t
h
e com
m
uni
cat
i
on an
d st
ora
g
e
ove
rhead can be com
puted
as
follows:
Tabl
e 2. Si
m
u
lat
i
on
Pa
ram
e
t
e
r
B
a
sed o
n
t
h
e
com
put
at
i
onal
form
ul
a gi
ve
n i
n
Tabl
e
3,
com
m
uni
cat
i
on an
d st
o
r
age
ove
rhea
d o
f
di
ffe
re
nt
al
g
o
ri
t
h
m
s
have
bee
n
cal
cul
a
t
e
d a
n
d
com
p
are
d
wi
t
h
pr
o
pose
d
fuz
z
y
base
d
r
e
pl
i
ca n
ode
de
t
ect
i
on
schem
e
. Fi
gure
4 i
n
di
cat
es t
h
e pr
op
ose
d
m
o
d
e
l
of f
u
zzy
bas
e
d re
pl
i
ca n
o
d
e
det
ect
i
on sy
s
t
em
consi
s
t
i
ng
t
h
re
e
in
pu
t p
a
ram
e
te
rs lik
e Tru
s
t agreeg
at
fro
m
same clu
s
ter,
T
r
ust agree
g
ate from
neighb
ouri
ng cluster and
Packet
d
e
liv
ery ratio
.
Tabl
e
3. C
o
m
p
ari
s
o
n
of
com
m
uni
cat
i
on an
d c
o
m
m
uni
cat
ion
o
v
e
r
hea
d
Para
m
e
ter
M
e
thod
Br
oadcast
Centralise
Detection
Deter
m
inistic
Multicast
Rando
m
i
sed
Multicast
Line selected
Multicast
Pr
oposed
sche
m
e
Co
m
m
unication Over
head
Storage Overhead
Fi
gu
re 4.
Pr
o
p
o
se
d
f
u
zzy
bas
e
d repl
i
ca no
de
det
ect
i
o
n
sy
st
em
Sim
u
l
a
t
i
on res
u
l
t
s
gi
ve
n i
n
F
i
gu
re5 a
nd
6 s
h
o
w
t
h
at
p
r
op
ose
d
schem
e
requi
res l
e
ss co
m
m
uni
cat
i
o
n
and st
o
r
a
g
e ov
erhea
d
. Let
n=
10
0,
d=4
0
, p=
0.0
5
, g=
2 an
d
s=1, t
h
e com
m
uni
cat
i
on ov
erhea
d
f
o
r b
r
o
a
dcas
t
m
e
thod
will be
10, ce
ntralise
detecti
on
will be
100, determ
inistic
m
u
lticas
t will be
10, ra
ndom
ised m
u
lticast
will b
e
1
000
,
lin
e selected
m
u
l
ticast will
b
e
1
0
0
and
for pro
p
o
s
ed
sche
m
e
will b
e
5. Furth
e
rm
o
r
e sto
r
ag
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
Im
pl
eme
n
t
a
t
i
o
n
of
F
u
zzy B
a
s
e
d
Si
m
u
l
a
t
i
o
n f
o
r C
l
o
n
e
Det
e
c
t
i
on i
n
Wi
rel
e
s
s
Se
ns
or .
...
(
M
anj
un
at
h
a
R. C
.
)
1
575
o
v
e
rh
ead fo
r bro
a
d
cast m
e
th
o
d
will
b
e
4
0
,
cen
tralise
d
e
tectio
n
will
b
e
40
, d
e
term
in
isti
c
m
u
lticas
t
will
b
e
2,
rando
m
i
sed
mu
lticast will b
e
1
0
,
lin
e selected
m
u
lticas
t will b
e
1
0
an
d
for p
r
op
osed
sch
e
m
e
will b
e
1.C
o
m
p
arat
i
v
e g
r
ap
h
fo
r c
o
m
m
uni
cat
i
on an
d st
ora
g
e
ove
r
h
ead
are
gi
ven
i
n
Fi
g
u
r
e
5 a
n
d
6
res
p
ect
i
v
el
y
.
100
200
300
400
500
600
700
800
900
1000
0
1
2
3
4
5
6
7
8
9
10
x 1
0
5
N
u
m
ber
of
nodes
N
u
m
ber
of m
e
s
s
age
s
t
r
a
n
s
m
i
tted and r
e
c
e
i
v
ed
C
o
m
m
uni
c
a
t
i
on Ov
er
head
B
r
oadc
as
t
C
ent
ral
i
z
ed M
e
t
hod
det
er
m
i
ni
s
t
i
c
M
u
l
t
i
c
as
t
R
andom
i
z
ed M
u
l
t
i
c
as
t
Li
ne s
e
l
e
c
t
ed
M
u
l
t
i
c
as
t
P
r
opos
ed S
c
hem
e
Fi
gu
re
5.
C
o
m
p
ari
s
on
o
f
c
o
m
m
uni
cat
i
on o
v
e
rhea
d
Fi
gu
re 6.
C
o
m
p
ari
s
on
o
f
St
or
age ove
r
h
ead
5.
CO
NCL
USI
O
NS
Thi
s
pape
r
pr
esent
s
si
m
u
l
a
ti
on
fram
e
wor
k
f
o
r
f
u
zzy
b
a
sed
repl
i
ca n
ode
det
ect
i
o
n
schem
e
i
n
cl
ust
e
red
wi
rel
e
ss sens
or net
w
o
r
k
s
. Th
e pr
op
ose
d
schem
e
u
s
es p
a
ck
et deliv
ery ratio
(PDR) and
tru
s
t repo
rts
to
d
e
term
in
e t
h
e prob
ab
ility
o
f
a clu
s
ter
b
e
in
g
co
m
p
ro
m
i
s
e
d
.
Perfo
r
m
a
n
ce o
f
propo
sed sch
e
m
e
is co
m
p
ared
with broa
dcast, centralize detection,
rand
o
m
i
zed
m
u
lticas
t,
d
e
term
in
istic
m
u
l
ticast an
d
l
i
n
e
selected
mu
lticast
m
e
t
hods
o
n
t
h
e basi
s
o
f
c
o
m
m
uni
cat
i
on a
n
d
st
o
r
ag
e
ove
rhea
d
re
qui
re
d
by
t
h
e al
g
o
r
i
t
h
m
.
In c
o
nv
en
t
i
onal
al
go
ri
t
h
m
s
com
m
uni
cat
i
on a
nd
st
or
age
o
v
e
r
hea
d
i
s
f
u
nct
i
on
o
f
n
u
m
b
er
of
n
odes
p
r
ese
n
t
e
d i
n
t
h
e sy
st
em
,
avera
g
e de
g
r
ee
of n
e
i
g
hb
or
h
o
od a
n
d n
u
m
b
er of wi
t
n
ess n
o
d
es w
h
e
r
eas i
n
pr
op
ose
d
sc
he
m
e
bot
h are
fu
nct
i
o
n
of
num
ber o
f
cl
ust
e
rs p
r
ese
n
t
and n
u
m
b
er of re
p
o
rt
i
n
g cl
ust
e
rs i
n
nei
g
h
b
o
r
ho
o
d
. Si
m
u
l
a
t
i
on res
u
l
t
s
sho
w
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l. 6
,
N
o
. 4
,
Au
gu
st 2
016
:
15
70
–
1
576
1
576
t
h
at
wi
t
h
sam
e
param
e
t
e
rs t
a
ken;
pr
op
ose
d
schem
e
requi
r
e
s l
e
ss com
m
u
n
i
cat
i
on a
nd st
ora
g
e o
v
e
r
hea
d
t
h
a
n
con
v
e
n
t
i
onal
al
go
ri
t
h
m
s
.
REFERE
NC
ES
[1]
R. Brooks,
et al.
, “On the detection of clon
es in sensor
networks using
random
key
pr
edistribu
tion
,
”
Sy
ste
m
s,
Man,
and Cybernetics
,
Part C:
App
lica
tions
and
Reviews, IEEE Transactions on
,
vo
l/issue: 37(6)
, pp
. 12
46-1258, 2007
.
[2]
H. Choi,
et al.
, “SET: Detecting node clone
s
in sensor networks,”
S
ecurity and Privacy
in
Communicatio
ns
Networks and
th
e Workshops, 20
07, Secure Co
m
m
2007, Third In
terna
tional Conference on
. IEEE, 2007
.
[3]
L.
Eschenau
er
and V. D. Gligor, “A key
-
manag
e
ment scheme for
distribut
ed sens
or networks,”
Pr
oceed
ings
of th
e
9th ACM
confer
ence on Comput
er and communications security
, ACM,
2002.
[4]
B. Parno,
et al.
,
“Distributed detection of node repli
cation attack
s in sensor networks,”
S
ecurity and Privacy, 2005
IEEE Symposiu
m on
. I
EEE, 200
5.
[5]
M. Conti,
et al.
, “A randomized, efficien
t,
and dist
ributed pro
t
ocol for th
e detecti
on of node r
e
plication attacks in
wireless sensor
networks,”
Proceedings of the 8th ACM
international symposium
on Mobile ad hoc networking an
d
computing,
ACM, 2007.
[6]
B. Zhu,
et al.
, “
L
oca
liz
ed
m
u
lticast:
effic
i
ent and distr
i
buted
repl
ic
a dete
ction in
l
a
rge-s
cal
e s
e
ns
or
networks,”
Mobile Computing,
I
EEE Transactio
ns on
, vol/issue:
9(7), pp
. 913-92
6, 2010
.
[7]
R. Geetha,
et a
l
.
,
“Fuzzy
logic based
compromise
d node d
e
tection and r
e
vocation
in cluster
e
d wireless
sensor
networks,”
Info
rmation Commu
nication and Embedded Systems
(
I
CICES)
, 2
014 Internation
a
l Conference
on
.
IEEE, 2014
.
[8]
T. Park
and
K. G. Shin
, “
S
oft tamper-pr
oofing vi
a pro
g
ram integr
ity verif
i
ca
tion in wire
le
ss se
nsor
networks,”
Mobile Computing,
I
EEE Transactio
ns on,
vol/issue:
4(3), pp
. 297-30
9, 2005
.
[9]
A.
Seshadri,
et al.
, “
S
watt: S
o
ftware-based
a
ttesta
tion fo
r e
m
bedded devi
c
e
s,”
S
ecurity an
d Privacy, 200
4,
Pr
oceed
ings
, 20
04 IEE
E
S
y
mpos
ium on
. IEEE, 2
004.
[10]
M. Beldj
e
hem
,
“
T
oward a Multi-Hop, Mult
i-
Path Fault-
To
le
rant and
Load
Balan
c
ing Hier
archi
cal Rou
tin
g
Protocol for
Wir
e
less Sensor Network,”
W
i
reless Sensor
Network
, 2013.
[11]
A.
K.
Kaushik,
“
A Hy
br
id Appr
oach of Fuzzy
C
-
means Clustering
and Neur
al n
e
twork to mak
e
Energ
y
-Eff
icient
heterog
e
neous
Wireless Sensor network,”
International
Journal of Electrical
an
d Computer Eng
i
neering (
I
JEC
E
)
,
vol/issue: 6(2),
2
016.
BIOGRAP
HI
ES OF
AUTH
ORS
M
a
njunatha R
C
obtained hi
s
B.E and M
.
T
ech Degre
e
from
Vis
v
es
hwaraya Univ
ers
i
t
y
,
Karnatak
a,
India, in
2006 and
2008 respectiv
ely
in
Telecommunication
En
g
i
neer
ing. He is
working as Assistant professor at Achar
y
a Ins
titute of Techno
lo
g
y
, Bangalor
e
, and Karnataka.
He is
current
l
y
purs
u
ing his
P
h
.D at J
a
in Unive
r
s
i
t
y
, Karna
t
ak
a.
His
current r
e
s
earch
includ
es
Clone d
e
te
ction
in wire
less Sensor Networks. He
is a
m
e
m
b
er of
I
S
TE and
IE
.
Dr K.
R.
Rekha
obtained her
ME degree from Bangalo
re University
, Ind
i
a in
2000. She i
s
working as a
Professor in th
e D
e
partm
e
nt
of
Ele
c
troni
cs and Co
m
m
unication in
SJB Institute of
Techno
log
y
, Bangalore. She has pursued her
Ph.
D. degree in
Dr MGR
University
, Chennai.
Her res
ear
ch
inter
e
s
t
s
inc
l
ud
e W
i
rel
e
s
s
co
m
m
unication,
F
P
GA im
plementa
tion,
and
Microcontro
ller
and Embedded s
y
stem d
e
sign.
S
h
e is
a m
e
m
b
er
of M
I
E,
M
I
S
TE
and IE
TE
Dr K.
R.
Nataraj
obtain
e
d his
ME degree from
Bangalor
e
Univ
ersity
, Ind
i
a in 2
000. He worked
as Professor and Postgraduate Coord
i
nato
r in the Dep
a
rtm
e
nt of E
l
ectron
i
cs
and
Com
m
unicationEngine
ering. C
u
rrentl
y
h
e
is Head
of the D
e
partm
e
nt
in SJB Institute of
Techno
log
y
, B
a
ngalor
e
. His research
inte
r
e
sts include Wir
e
less communication, FPGA
implementation, and Microcontr
o
lle
r and
Embed
d
ed s
y
stems design. He is a member of MIE,
MIST
E
, IE
TE
and IE
EE
Evaluation Warning : The document was created with Spire.PDF for Python.