TELKOM
NIKA
, Vol.13, No
.2, June 20
15
, pp. 478 ~ 4
8
6
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v13i2.995
478
Re
cei
v
ed
No
vem
ber 1
3
, 2014; Re
vi
sed
F
ebruary 21,
2015; Accept
ed March 1
2
, 2015
Integration of Signal and Artific
i
al Noise in MIMO
Wiretap Channel
Zhiliang Yang*
1,2
, Aihua Wang
1
, Xiqiang Qu
2
1
School of Info
rmation a
nd El
ectronics, Bei
j
i
ng
Institute of T
e
chnolog
y, B
e
iji
ng 1
0
0
081,
Chin
a
2
School of Info
rmation a
nd C
o
mmunic
a
tio
n
Engi
neer
in
g,
North Univ
ersit
y
of Chin
a T
a
i
y
u
an 03
00
51, Chi
n
a
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: sxt
yyz
l@si
na
.com
A
b
st
r
a
ct
In this pap
er, the inte
grate
d
sign
al-to-artific
i
al
no
ise (ISAN
) desig
n is ap
plie
d in MIMO w
i
retap
chan
nel
to
ens
ure w
i
re
less c
o
mmunic
a
tio
n
security.
W
h
e
n
the
infor
m
atio
n of
eav
esdro
pper
is
unk
no
w
n
,
the total
p
o
w
e
r is
divi
de
d i
n
to
tw
o parts: sig
n
a
l
and
artific
i
al
nois
e
. T
h
e
sig
nal
can
sec
u
re
certai
n q
u
a
lity
at
the leg
i
timate receiv
er. T
he artificial no
ise w
h
ich is
in the
null sp
ace of the receiv
er ch
ann
el matrix ca
n
deteri
o
rate
eav
esdro
pper c
h
a
nne
l by
th
e
me
thod of
bea
m f
o
rming. T
h
e ar
tificial
nois
e
p
o
w
er is distrib
u
te
d
even
ly i
n
ot
he
r spac
e, so th
at the
eav
esdr
opp
er ch
an
nel
is d
e
teri
orate
d
i
n
a
ll
directi
ons. T
h
e
sig
n
al
t
o
interface
an
d n
o
ise r
a
tio (SIN
R) is re
gar
ded
as the
effi
cie
n
t para
m
eter o
n
me
asuri
ng r
e
li
abil
i
ty an
d sec
u
rit
y
of infor
m
atio
n
at th
e l
egiti
mate
rec
e
iver.
T
he s
i
mul
a
ti
ons r
e
vea
l
th
at ISAN ca
n
deter
iorat
e
t
h
e
eaves
drop
per chan
nel an
d
sa
feguar
d
the inf
o
rmatio
n
tr
ans
miss
ion
on th
e
pre
m
ise
of the
give
n SINR of t
h
e
legitimate receiver.
Ke
y
w
ords
:
MIMO W
i
retap Chan
nel, ISAN, Averag
e SINR
1. Introduc
tion
With the ra
pid develo
p
m
ent of wi
reless
comm
unication bu
sine
ss, the
se
curity of
informatio
n transmi
ssion
h
a
s
bee
n pai
d
widely
att
ention. The
tra
d
itional
encrypti
on meth
od
s a
r
e
mostly ba
sed
on crypto
gra
phy [1]-[3], which b
u
ilt
se
curity mech
ani
sm ab
ove the netwo
rk layer.
In 1975,
Wy
ner
pro
p
o
s
e
d
the
wireta
p ch
ann
el m
odel
(WT
C), whi
c
h
provi
ded the
theo
retic
foundatio
n fo
r inform
ation
transmissio
n
over phy
sica
l
layer from th
e angle
of informatio
n the
o
ry.
The
WT
C p
r
oved th
at t
here
is a
secu
rity ca
pa
city
s
C
(
0
s
C
) when
the
cha
n
n
e
ls
of
eavesdro
ppe
r is inferio
r
to the sou
r
ce
. The in
form
ation ca
n be
transmitted
safely from t
he
sou
r
ce to the
destin
a
tion if
the data rate
is le
ss th
an
s
C
, in this
situati
on the e
a
vesdrop
p
e
r
can
receive the
d
a
ta, but he
cannot o
b
tain
any u
s
eful
inf
o
rmatio
n of
source [4]. In
this p
ape
r, th
e
sou
r
ce i
s
Alice, the legitim
a
te re
ceive
r
i
s
Bob,
a
nd th
e eave
s
d
r
opp
er i
s
Eve. Eve ca
n get
part of
the so
urce in
formation
by wireta
ppin
g
. The c
han
nel
betwe
en Alice and Bo
b is main chan
n
e
l,
and the chan
nel between
Alice and Eve
is
eavesdro
p
per chan
nel (see Fig
u
re 1).
Figure 1. The
wiretap
cha
n
nel
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Integration of
Signal and Artificial Noi
s
e i
n
MIMO Wire
tap Cha
nnel (Zhiliang Ya
ng
)
479
The aim of integrate
d
sig
nal-to
-
artifici
a
l
noise (ISA
N) de
sig
n
is to improve
s
C
. There
were ma
ny rese
arche
s
on
the metho
d
s by whi
c
h
s
C
ca
n be im
prove
d
. The
se
cu
ri
ty capa
city of
WT
C is p
r
op
ose
d
ba
sed
o
n
the bro
a
d
c
asting
cha
n
n
e
l. The main
cha
nnel of th
e wiretap cha
nnel
is
n
o
isele
ss
chann
el, and the eaves
dro
pper
cha
nnel
is rece
ssion
cha
nnel of the main cha
n
nel
(Binary Sym
m
etric Chan
n
e
l). S. Le
ung
cal
c
ulate
d
un
der th
ese
con
d
itions that m
a
in
cha
nnel
a
nd
eavesdro
ppe
r chann
el are
Gaussian
chann
els
[5], and proved that when th
e eavesd
r
o
p
per
cha
nnel is th
e degraded
cha
nnel of main ch
anne
l,
s
C
is the differen
ce b
e
tween the mai
n
cha
nnel
an
d
the e
a
vesd
ropp
er chan
nel. To
in
crease
s
C
signifi
cantly, He
ro i
n
trodu
ce
d th
e
Multiple Inp
u
t Multiple Output (MI
M
O) an
d space diversi
t
y technolog
y into secu
rity
comm
uni
cati
on [6], an
d g
a
ve the fun
d
a
mental
co
m
puting m
e
tho
d
of
s
C
. A. Khis
ti [7]-[9] s
t
udied
the com
puti
ng metho
d
of
s
C
in the MIMOME scenari
o
s. In
most cases,
becau
se th
e
eavesdro
ppe
r is pa
ssive, Alice
can
not o
b
tain any
thin
g abo
ut the
chann
el
state i
n
formatio
n (CSI)
of Eve, Ne
gi [
10]-[13]
pro
p
o
se
d th
e
sch
e
me
of artifici
al noi
se
to
maximum the sec
u
rity rate when
the eavesd
r
o
pper i
s
pa
ssiv
e
. The basi
c
idea of t
he scheme is: Alice splits the transmitting
sig
nal
vector into two parts: the signal vecto
r
and the arti
fici
al noise ve
ctor. The artificial noise lie
s in
the null
spa
c
e of the
sign
a
l
vector, d
e
te
riorates
th
e
e
a
vesd
rop
p
e
r
cha
nnel and has no
effe
ct on
the main
ch
a
nnel. N.
Rom
e
ro
studi
ed t
he sch
e
me
of artificial
noi
se to secure
d
a
ta tran
smi
s
sion
in the MISO system [14]
-[15]. The
re
ce
iver sig
nal to
interface an
d
noise
ratio
(SINR) i
s
ap
pl
ied
as th
e me
asurem
ent of
s
C
b
y
A. Mukhe
r
j
ee, then
he
p
r
opo
se
d a
n
o
p
timized
met
hod of
po
wer
dis
t
ribution [16].
Based
on
former
re
sea
r
ch
es, ISAN
design is p
r
opo
sed in th
e MI
MO wi
reta
p
cha
nnel
whi
c
h can d
e
terio
r
ate the
eavesd
r
op
p
e
r ch
ann
el
a
nd safe
gua
rd
the informati
on tran
smi
ssi
on
over the wi
re
less ch
ann
el, meanwhile, se
curity rate
is always po
sitive and Bo
b can g
e
t the
information successfully.
2. MIMOME s
y
stem mod
e
l
The num
ber of transmitter anten
na
s of Alice is
(2
)
aa
NN
, the numb
e
r
of receive
r
antenn
as of
Bob is
(1
)
bb
NN
, the n
u
mbe
r
of
eav
esd
r
op
pe
r an
tenna
s of Ev
e is
(1
)
ee
NN
(s
ee
Figure
2
)
. We su
ppo
se
that
the ch
annel betwe
en Alice an
d Bob i
s
Rayleigh flat f
ading
cha
nnel
(
b
H
)
w
h
ic
h
is
k
n
ow
n
b
y
Alic
e
an
d its
co
va
r
i
a
n
ce
ma
tr
ix is
2
b
h
I
;
the chan
nel b
e
twee
n
Alice a
nd Ev
e is
Raylei
gh
flat fading
chann
el(
e
H
) which is
un
kno
w
n by Alice its cova
rian
ce
matrix is
2
e
h
I
. The cha
nnel b
e
t
ween Alice and Bob an
d
noise ve
ctor (
b
n
) are su
peri
m
posed;
the
chan
nel b
e
twee
n
Alice and
Eve and
noise
ve
ctor (
e
n
)
a
r
e su
peri
m
posed. We suppo
se
that
the values of
b
n
at different times a
r
e inde
pend
ent
of each othe
r, the mean of
b
n
is ze
ro, and the
varian
ce of
b
n
(
2
b
)is c
o
mplex
Gau
ssi
an
mat
r
ix
.
V
e
ct
or
e
n
also ha
s such a p
r
ope
rty.
Res
p
ec
tively the varianc
e
matrix of
b
n
and
e
n
is:
2
2
{}
{}
H
H
bb
b
ee
e
E
E
nn
I
nn
I
(1)
We suppo
se t
hat the signal
s of Alic
e based on beam f
o
rmin
g techn
o
logy is
s
, the sign
al
received by Bob and Eve is as followi
ng:
e
bb
b
ee
y
y
Hs
n
Hs
n
(2)
The cova
ria
n
c
e matrix of
s
is
{}
H
s
E
Qs
s
, the total power of
s
is
0
{}
PT
r
s
Q
。
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 478 – 48
6
480
A
lic
e
B
o
b
Ev
e
En
c
o
d
e
r
D
ec
ode
r
D
ec
od
e
r
H
b
H
e
N
a
M
M
N
e
N
b
Figure 2. The
MIMOME wiretap ch
ann
el
3.
Integra
t
ed si
gnal-to-a
rtifi
c
ial noise de
sign
3.1. The model of ISAN
The si
gnal
s
can be
split in
to two pa
rts:
sign
al vecto
r
and a
r
tificial
noise ve
cto
r
. Th
e
total power o
f
s
is defined:
0
P
ab
, where
a
is th
e po
wer th
at
Alice wants t
o
se
nd a
nd
b
is
the artificial n
o
ise p
o
wer. T
hen the si
gna
l
s
can be defi
ned a
s
ad
b
st
η
(3)
Whe
r
e
t
is the
(1
)
a
N
normali
ze
d beam form
ing vector a
n
d
1
t
‖‖
.
d
is the scal
ar
compl
e
x info
rmation
symb
ol which Ali
c
e wants to
se
nd, an
d th
e
mean
of
d
is
{|
|}
1
Ed
.
η
is
the
(1
)
a
N
artificial noise
ve
ctor,
its cova
rian
ce matrix
i
s
{}
H
E
Q
ηη
, and
its t
r
ace i
s
{}
Tr
Q
(
{}
1
Tr
Q
).
The sig
nal
s that Bob and
Eve receive a
r
e
bb
b
ee
e
e
b
d
a
b
a
b
d
y
y
Ht
H
η
n
Ht
H
η
n
(4)
Becau
s
e Eve
is a passive receiver, Alice can
not get the CSI about
Eve.
In this case, we
sho
u
ld not regard
s
C
a
s
th
e security m
easure
m
ent
of the
syste
m
. The
re
cei
v
er
sign
al to
interferen
ce
and n
o
ise rati
o (SINR)
refl
ects th
e re
cei
v
er quality, so the SINR i
s
use
d
to indi
cate
the quality of receiver
sign
als.
A
ssu
me
b
w
denotes the
1
b
N
bea
m forming ve
ctor of Bob.
Similarly,
e
w
denotes the
1
e
N
beam formi
n
g vector of Eve, the receiv
er vecto
r
of Bob and Eve can be
denoted
s
e
parately as:
ˆ
()
ˆ
()
bb
b
ee
HH
b
ee
e
e
bb
b
HH
e
ad
ad
zb
zb
wy
w
H
t
H
η
n
wy
w
H
t
H
η
n
(5)
And the re
cei
v
er SINR of Bob and Ev
e can be de
note
d
sep
a
rately
as:
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Integration of
Signal and Artificial Noi
s
e i
n
MIMO Wire
tap Cha
nnel (Zhiliang Ya
ng
)
481
2
2
2
2
)
)
(
(
H
bb
b
H
b
ee
e
ee
bb
b
H
e
H
ee
SINR
S
a
P
INR
H
b
H
wH
t
wH
Q
H
I
w
wH
t
wH
Q
H
I
w
(6)
If
be
SINR
SINR
, there i
s
th
e sch
e
me
of modulatio
n a
nd
codin
g
to
ensure
that B
ob
can
decode the m
e
ssag
e
d
su
ccessfully, at th
e same ti
me,
Eve cannot g
e
t anything a
bout
d
.
The MIMO wi
retap chan
nel
based o
n
integrate
d
sig
n
a
l
-to-a
r
tificial n
o
ise de
sig
n
can be
descri
bed a
s
:
0
mi
n
..
00
,
b
e
b
SI
N
R
s
tP
a
b
SIN
R
ab
t
(7)
Whe
r
e
b
is the given target value of Bob’s SINR.
3.2. Integrated signal-to-artificial noi
se design
The aim of in
tegrated
sign
al-to-artificial
noise
de
sign
is to cau
s
e Al
ice to se
nd th
e noise
in all di
re
ctio
ns, to influe
nce th
e eav
esd
r
op
pe
r a
s
much a
s
p
o
ssible, a
nd
to influen
ce
the
legitimate re
ceiver as little as po
ssible.
Alice’s b
eam
forming vector is
t
which is the eig
envecto
rs
co
rre
sp
ondi
ng
to the
maximum ei
genvalu
e
of
b
H
. Artific
i
al nois
e
vec
t
or (
η
) is the linea
r com
b
inatio
n of other
1
a
N
cha
r
a
c
t
e
ri
st
ic
v
e
ct
ors of
b
H
. As a re
sult, the signal vecto
r
s and a
r
tificia
l
noise vecto
r
s are
orthog
onal. Other
1
a
N
ch
ara
c
t
e
rist
ic
v
e
ct
o
r
s of
b
H
are
dist
ri
buted e
qually
, in this
prin
ci
ple,
η
is define
d
as:
2
1
1
a
N
i
a
i
i
N
η
t
(8)
Whe
r
e,
i
t
is the ith eigenve
c
tor of
b
H
, and
i
is a ran
dom
compl
e
x scal
ar whi
c
h h
a
s
unit amplitud
e and ra
ndo
m phase.
i
is defined a
s
：
i
j
i
e
.
The pha
se
of
i
is
i
which follows uniform
distributio
n, and
[0
,
2
]
i
.
Q
is defined
as:
2
1
1
a
N
H
ii
i
a
N
Qt
t
(9)
b
w
is Bo
b’s re
ceive be
am f
o
rmin
g ve
cto
r
(
bb
wH
t
). Eve’s SINR will
be
maximize
d
whe
n
his b
e
a
m
forming ve
ctor i
s
21
(+
)
H
ee
e
e
e
wH
Q
H
I
H
t
.
At this mome
nt, the SINRs of Bob and Eve are
b
SIN
R
and
e
SINR
res
p
ec
tively.
2
11
1
22
21
11
()
b
b
HH
b
HH
H
ee
e
e
ee
b
SINR
SIN
a
a
Ra
b
tHH
t
tH
H
Q
H
I
H
t
(10
)
whe
r
e,
1
is the
maximum eigenvalu
e
of
b
H
.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 478 – 48
6
482
The informati
on of eave
s
drop
pe
r ca
n
not
be sure,
and we ca
nnot en
su
re
that the
eavesdro
ppe
r is influe
nced
by manu
al n
o
ise,
so we can
dete
r
io
rat
e
eave
s
d
r
op
per ch
annel
as
far a
s
po
ssib
le statisti
cally
. SINR can
descri
be the
extent how t
he eave
s
d
r
o
pper
ch
ann
el
is
influen
ced by
manual noi
se. Unde
r the con
d
ition
s
that the total tr
ansmit po
we
r and Bob’s SINR
are con
s
tant
, all these probl
em
s co
me down to find Eve’s
minimum SINR with
b
SIN
R
con
s
trai
nts:
21
11
0
2
1
2
mi
n
m
i
n
[
)
(]
..
0,
0
HH
H
ee
e
e
e
e
bb
b
AV
E
a
st
a
b
SIN
R
a
P
b
a
tH
H
Q
H
I
H
t
(11
)
4. Results a
nd Analy
s
is
Becau
s
e th
e
eavesdro
ppe
r cha
nnel i
s
u
n
kn
ow
n, the
Monte Carl
o
method i
s
ap
plied to
examine h
o
w the integ
r
ate
d
sig
nal-to
-
a
r
tificial
noi
se
desi
gn affe
cts the e
a
vesd
ropp
er
ch
an
nel.
All the simul
a
tions in th
e pape
r set the times
of Mo
nte Ca
rlo ex
perim
ent to b
e
5000, a
nd
the
sign
al-to
-
interferen
ce
-an
d
-noise-ratio (S
I
NR)
of the legitimate re
ceiver (Bo
b
) i
s
5
b
dB
. Onc
e
Alice
kno
w
E
v
e’s ave
r
ag
e
SINR in
CSI, and
the
si
tu
ation of B
ob’
s SINR i
s
proper,
Alice
could
put artificial
noise in Eve’s ch
ann
el space ea
sily in ord
e
r to d
e
terio
r
ate th
e eavesdrop
per
cha
nnel.
The method
applied wh
en Alice kn
ows the exact Eve’s locatio
n
is u
s
ed for
referen
c
e.Th
e ISAN an
d
Ref a
r
e a
p
p
lied
sep
a
rat
e
ly whe
n
th
e num
ber
of the ante
nna
s is
8,
4
,
4
b
ae
NN
N
. The Figu
re
3 sh
ows the
averag
e SINR
of eave
s
d
r
oppe
r in
all situations.
The re
sults
reveal
that th
e SINR of legitimate re
cei
v
er must be prop
er, in this situation, Eve’s
SINR
gets smaller an
d al
locate
d p
o
we
r of
artifici
al
noise b
e
com
e
s
high
er wit
h
the
gradu
a
lly
increa
sing t
r
a
n
smit p
o
wer,
and the
se
cu
rity perfo
rm
a
n
ce i
s
better.
If Alice knows that the Ev
e’s
CSI, more ta
rgeted
actio
n
s have
bee
n ta
ken
by Ali
c
e
t
o
inte
rfere
wit
h
Eve. Th
e
re
sult
sho
w
s th
at
Eve’s SINR
woul
d be
kep
t
belo
w
13
dB
if the
tran
smit p
o
wer of Ali
c
e
is
big e
noug
h.
The a
bove
method i
s
su
ppo
sed
as
a
Ref. For IS
AN, the artifi
cial n
o
ise is
distrib
u
ted ev
enly from ev
ery
dire
ction,
an
d the
sim
u
lat
i
on
re
sults re
veal that
su
ch a
metho
d
can
dete
r
io
ra
te eave
s
d
r
op
per
cha
nnel effe
ctively thoug
h its SINR i
s
wo
rse
tha
n
the Ref. For exam
ple
,
if the transmit
power
0
15
P
dB
, rece
ption SINR wou
l
d belo
w
13
dB
(whi
ch is a
pproximates to zero
).
2
4
6
8
10
12
14
16
18
-2
5
-2
0
-1
5
-1
0
-5
0
5
P
0
(d
B
)
AV
e
(d
B
)
I
SAN
(
b
2
=
e
2
)
I
SAN
(
e
2
=0
)
Ref
(
b
2
=
e
2
)
Ref
(
e
2
=0
)
Figure 3. The
average
e
SIN
R
with respec
t to diffe
rent tran
smit powe
r
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Integration of
Signal and Artificial Noi
s
e i
n
MIMO Wire
tap Cha
nnel (Zhiliang Ya
ng
)
483
4.1. Ho
w
doe
s the numbe
r of antenna
influence
s
SINR of e
a
v
esdropper
The num
be
rs of transmitti
ng anten
na
s
and re
ceiving
antenna
s
will affect the SINR of
the re
ceive
r
. The p
e
rfo
r
mance of e
a
v
esdropp
er‘
s
SINR i
s
an
alyzed
wh
en
the num
bers of
transmitting
antenn
as a
n
d
re
ceiving
antenn
as
ch
ange o
n
the
premi
s
e th
at the SINR of
legitimate
re
ceiver is con
s
tant(
5
b
dB
).In the
method
of I
SAN, the p
e
r
forma
n
ce of
Eve’s
averag
e SINR co
rrespon
ding to
the n
u
mbe
r
of the
antenna
s (
4,
4
be
NN
)
is differe
nt (see
Figure 4
)
. Th
e sim
u
lation
s reveal th
at the in
crea
se o
f
the numb
e
rs of tra
n
smitting ante
nna
s
and
transmit power will reduce
e
SIN
R
dramati
c
all
y
, especi
a
lly
ae
NN
.For example,
e
SIN
R
will
rea
c
h the lev
e
l of
13
10
dB
when
0
8,
1
0
a
NP
d
B
. It is found that
the gain of legitimate re
ceiver
cha
nnel a
n
d
the dimen
s
ion of artificial noi
se vector
will increase wh
en
the numbe
r o
f
transmitting a
n
tenna
s in
cre
a
se
s. As a re
sult,
the eave
s
dropp
er
will be interfe
r
ed
seri
ou
sly.
2
4
6
8
10
12
14
16
18
20
-2
0
-1
5
-1
0
-5
0
5
N
a
AV
e
(d
B
)
P
0
=1
d
B
P
0
=5
d
B
P
0
=1
0
d
B
P
0
=1
5
d
B
Figure 4. The
average
e
SINR
with respe
c
t to different tran
smitting anten
nas of Alice
2
4
6
8
10
12
14
16
18
20
-1
0
-8
-6
-4
-2
0
2
N
b
AV
e
(d
B
)
P
0
=
1dB
P
0
=
5dB
P
0
=
10dB
P
0
=
15dB
Figure 5. The
average
e
SIN
R
with respe
c
t to different RX an
tenna
s of Bob
In the m
e
th
od of
ISAN, Eve’s
e
SINR
varie
s
a
c
co
rdin
g
to the
ch
an
ging
b
N
when
a
N
and
e
N
are con
s
t
ant
(
4,
4
ae
NN
) (see Fi
gure
5
). Th
e simulation
s
re
veal that the increa
sing
of
b
N
will decrease
e
SINR
. As
a
result, the interferenc
e
is
gr
eater
within this
c
ontext . It c
a
n not
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 478 – 48
6
484
be ign
o
red th
at
e
SINR
is affe
cted
by
a
N
as well
a
s
b
N
,and the
ch
angin
g
b
N
has l
e
ss influe
nce
than ch
angi
n
g
a
N
becau
se th
e increa
sing
b
N
only raise the
diversity gain
of legitimate receiver.
4.2. Ho
w
doe
s the ch
anne
l
v
a
riance influences SINR of eav
esdr
opper
In the MIMO
Wiretap
cha
nnel, the im
portant
co
ncl
u
sio
n
of info
rmation t
r
an
smissi
on
se
curity i
s
th
e quality of l
egitimate receiver
cha
nnel
is
sup
e
rio
r
t
o
eave
s
d
r
op
per
ch
ann
el. The
signifi
cant a
d
v
antage of I
SAN is that t
he inform
atio
n ca
n be tra
n
sferre
d security even if the
quality of legitimate receiver chan
nel is inferior
to e
a
vesd
rop
p
e
r
cha
nnel. In the se
ction,
we
analyze the
i
m
pact
s
of
th
e qu
ality of le
gitimate
recei
v
er chan
nel
and
eave
s
dro
pper chan
nel
on
the
e
SIN
R
of eavesd
r
opp
er (se
e
F
i
gure 6
)
.
0
5
10
15
20
-2
-1
0
1
2
-2
0
-1
5
-1
0
-5
0
P
0
(d
B
)
h
e
2
/
h
b
2
(d
B
)
AV
e
(d
B
)
(a).
0
5
10
15
20
-1
6
-1
4
-1
2
-1
0
-8
-6
-4
-2
P
0
(d
B
)
AV
e
(d
B
)
h
e
2
/
h
b
2
=-
1
d
B
h
e
2
/
h
b
2
=0
d
B
h
e
2
/
h
b
2
=1
d
B
h
e
2
/
h
b
2
=2
d
B
-2
-1
0
1
2
-1
4
-1
2
-1
0
-8
-6
-4
-2
h
e
2
/
h
b
2
(dB
)
AV
e
(d
B
)
P
0
=1
d
B
P
0
=5
d
B
P
0
=1
0
d
B
P
0
=1
5
d
B
(b).
Figure 6. The
average
e
SINR
; a) with res
p
ec
t to different
0
P
and
22
/
eb
hh
holis
tically; b) with
respec
t to different
0
P
and
22
/
eb
hh
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Integration of
Signal and Artificial Noi
s
e i
n
MIMO Wire
tap Cha
nnel (Zhiliang Ya
ng
)
485
In MIMOME,
if the number of antenna
s are fixed (
8,
4
,
4
ab
e
NN
N
), Eve’s avera
g
e
e
SINR
cha
nge
s with
the ratio of eavesd
r
op
p
e
r ch
ann
el varian
ce a
nd
the receivers chan
nel
var
i
anc
e
(
22
/
eb
hh
). Fig 6a
sho
w
s ho
w the
e
SIN
R
chang
es with the
ch
ang
e of
22
/
eb
hh
and
0
P
holisti
cally. Fig 6b
sho
w
s h
o
w the
e
SIN
R
ch
ang
es with
the chang
e
of
22
/
eb
hh
and
0
P
resp
ectiv
e
ly
.
T
h
e
s
i
mu
la
tion
s
r
e
ve
a
l
th
at ISAN
c
a
n so
lve
th
e
in
flue
n
c
e
ca
us
e
d
b
y
th
e
d
i
ffe
re
n
c
e
b
e
t
w
een
legitimate re
ceiver chan
nel
and eave
s
d
r
oppe
r chan
n
e
l efficiently. For exam
ple,
whe
n
0
10
P
dB
,
22
/2
b
e
hh
dB
,the varien
ce
of eave
s
d
r
o
pper chan
nel
is
100
time
s the
varie
n
ce of le
gitimate
receiver chan
nel, an
d the
method
of IS
AN
can
redu
ce
e
SIN
R
to the level of
9
10
dB
. The
high
er
the transmit power is, the
more effecti
v
e the I
SAN
deal with the
difference betwee
n
legitimate
receiver
cha
nnel and e
a
vesd
rop
p
e
r
chan
nel. The sim
u
lati
ons reveal
that when t
h
e
eavesdro
ppe
r ch
ann
el i
s
superi
o
r to th
e
re
ceive
r
ch
a
nnel, ISAN can e
n
sure th
e qu
ality of B
ob’s
rec
e
iv
er
b
SINR
an
d re
du
ce E
v
e’s ave
r
ag
e
e
SIN
R
.
As a
re
su
lt, the cha
n
nel capa
city of
eavesdro
ppe
r will be red
u
ced marke
d
ly and informat
i
on's
se
cu
rity can b
e
en
sured effectively.
5. Conclusio
n
In this pa
per,
the method
of integrate
d
sign
al
-to
-
artifi
cial n
o
ise is
applie
d in MI
MO to
improve th
e
se
curity of
si
gnal
s tra
n
smi
ssi
on. We
d
e
s
ign th
e
spe
c
ific
step
s to
impleme
n
t the
algorith
m
. The method in whi
c
h the tra
n
smitter
kno
w
s the info
rm
ation of eavesdropp
er is u
s
e
d
as a refe
ren
c
e metho
d
, and we
com
pare the inte
rru
ption
s
of eavesdro
ppe
r from differe
nt
method
s. Th
e influen
ce
s
with eave
s
d
r
oppe
r a
r
e a
n
a
lyzed from t
he pe
rspe
ct
ives of the
nu
mber
of the a
n
ten
nas an
d cha
nnel va
rian
ce. The
sim
u
lations reve
al
that ISAN
can
re
du
ce t
he
averag
e SINR of eave
s
d
r
oppe
r effectiv
ely and the
se
cu
rity of informatio
n tran
smissio
n
will
be
improve
d
on the co
ndition
of fixed SI
NR at the legitimate receiver.
Ackn
o
w
l
e
dg
ments
This
wo
rk
wa
s
sup
porte
d i
n
pa
rt by
NS
F of
China
wi
th gra
n
ts
612
7125
8, the
Rese
arch
Fund for the
Do
ctoral Pro
g
ram of Hig
h
e
r Edu
c
ation
with gra
n
ts 2
0131
1011
100
27
,Natio
nal 863
Program with
grants
2014
AA01A707.
Referen
ces
[1] JA
T
homas,T
M
Cover.
Ele
m
e
n
ts of Informati
on T
heory
. W
i
l
e
y
-
Int
e
rscie
n
ce
. 2006.
[2]
Sur
y
ad
i MT
, Sukirman
E, Ag
us MM. T
he i
m
pleme
n
tatio
n
of h
eno
n ma
p
alg
o
rithm f
o
r
digit
a
l
imag
e
encr
y
pti
on.
T
e
l
k
omnik
a
(T
el
e
c
ommunic
a
tio
n
Co
mp
utin
g El
ectronics
an
d
Contro
l)
. 20
14;
12(3):
65
1-
656.
[3]
Nurp
eti E, Surya
d
i MT
, Widya D.
Performa
nce of cha
o
s-b
a
sed e
n
cr
ypti
o
n
alg
o
rithm for
digit
a
l ima
ge.
T
e
lko
m
nik
a
(T
elec
o
m
mun
i
cat
i
on C
o
mputi
ng
Electron
ics an
d Control).
20
14; 12(3): 6
75-
682.
[4]
AD W
y
ner. T
he W
i
re-T
ap Chann
el .
The bell system
tech
nical journal.
197
5; 54(8): 13
55-
138
7.
[5]
S Le
ung
YC,
Hellm
an ME.
T
he Gaussian
w
i
re-tap
ch
ann
el.
Infor
m
ati
o
n
T
heory, IEEE
T
r
ansacti
on
s
on
. 197
8; 24(4)
: 451-45
6.
[6]
AO Hero. Secure sp
ace-tim
e
commu
nicati
on.
Infor
m
atio
n T
heory, IEE
E
T
r
ansactio
n
s
on
. 20
03;
49(1
2
): 323
5-3
249.
[7]
A Khisti, Wornell,
Gregory
,
W
i
ese
l
, Ami
,
Eldar,
Yon
i
na.
On the
Gaussian
MI
MO Wiretap
Cha
nne
l
.Infor
m
ation T
heor
y
,
2007. ISIT 2007. IEEE In
ternation
a
l S
y
mpo
s
ium. 200
7: 24
71-2
475.
[8]
A Khisti, Wornell, Gregory
W.
Secur
e
T
r
ansmission
W
i
th
Multipl
e
A
n
ten
nas I: T
he MISOME W
i
reta
p
Cha
nne
l.
Inform
ation Theory, IEEE Transactions on
. 201
0; 56(7): 30
88-
31
04.
[9]
A Khisti, Wornell, Gregory
W.
Secure T
r
ansmission With
Multip
le Antennas-Part II: T
h
e MIMOME
W
i
retap Ch
ann
el.
Information
Theory, IEEE Transactio
n
s o
n
. 2010; 5
6
(11)
: 5515-5
5
3
2
.
[10]
R Neg
i
, S Goel. Guar
ante
e
in
g
Secrec
y usin
g Artifici
al No
ise.
IEE
E
T
r
nsactions
on W
i
rel
e
s
s
Co
mmun
icati
o
ns.
2008; 7(
06)
: 2180-2
1
8
9
.
[11]
S Goel,
Ne
gi
R. Guar
ante
e
in
g
Secr
ec
y
usin
g Artifici
al
Nois
e.
W
i
re
l
e
ss C
o
mmun
i
c
ations, IEEE
T
r
ansactio
n
s o
n
. 2008; 7(
6): 2180-
218
9.
[12]
R Neg
i
, Goel S
.
Secret co
mmunic
a
tion
usi
n
g
artificial
no
ise
.
Vehic
u
lar T
e
chno
log
y
Co
nfer
ence, 2
0
0
5
.
VT
C-2005-Fall.
2005 IEEE 62
nd. 200
5: 190
6
-
191
0.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 478 – 48
6
486
[13]
S Goel, R N
egi
.
Secret
co
mmunic
a
tion in pre
s
enc
e of
coll
ud
ing eaves
dro
p
pers
. Proc. MIL
C
OM:1501-
150
6.
[14]
N Romero Z
,
Ghogh
o M, Mc
lern
on D. Outage Prob
abi
lit
y
Based Po
w
e
r Distributi
on Be
t
w
e
en Dat
a
and Artifici
al N
o
ise for Ph
ysic
al La
ye
r Securi
t
y
.
Signal Proc
essing Letters, IEEE.
2012; 19
(2): 71-74.
[15]
Li J, Petrop
ul
u AP. On Erg
odic S
e
crec
y
Rate for Gaus
sian MISO W
i
retap C
han
ne
l
s
.
Wireless
Comm
unications, IEEE Transactions on
. 201
1; 10(4): 11
76-
118
7.
[16]
A Mukherje
e, S
w
i
n
dle
hurst AL. Robust Bea
m
fo
rming for Securit
y
in MIMO W
i
retap Chan
nels W
i
t
h
Imperfect CSI.
Signal Process
i
ng, IEEE Transactions on.
20
11; 59(1): 3
51-
361.
Evaluation Warning : The document was created with Spire.PDF for Python.