TELKOM
NIKA
, Vol.14, No
.2, June 20
16
, pp. 548~5
5
4
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v14i1.3148
548
Re
cei
v
ed
No
vem
ber 2
5
, 2015; Re
vi
sed
April 20, 201
6; Acce
pted
May 3, 201
6
Efficient D2D Discover
y Scheme for Channel
Measur
e
ment of Interference Alignment
Zhanjun Liu, Yunpeng Li*
,
Zhonghua
Yu, Yue Shen
Cho
ngq
in
g Ke
y L
abor
ator
y
of
Mobil
e
Comm
unic
a
tion T
e
ch
nol
og
y,
Cho
ngq
in
g Uni
v
ersit
y
of Posts
and T
e
lecom
m
unic
a
tions (C
QUPT
), Chong
qin
g
, 400
065,
Chin
a
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: li
yun
pen
g2
29
@16
3
.com
A
b
st
r
a
ct
Co-ch
ann
el i
n
terferenc
e a
m
o
ng Dev
i
ce-to-
d
e
vice (D
2D) us
ers is major is
sue to be s
o
lv
ed w
h
e
n
utili
z
i
n
g
th
e sa
me fre
que
ncy
ban
ds. Cons
eq
uently, int
e
rf
er
ence a
l
i
g
n
m
e
n
t
(IA) as
an effective interfer
e
n
ce
ma
na
ge
me
nt
appr
oach
has
bee
n i
m
ple
m
ented
to D
2
D
co
mmun
icatio
ns for the fr
e
que
ncy sh
arin
g.
How
e
ver, the me
asur
e
m
ent
of chann
el state infor
m
ati
on (
C
SI) betw
een
trans
mitter an
d receiv
er of a D2D
pair a
nd cr
oss-chan
nel state
i
n
formatio
n
(C-
C
SI) amon
g D
2
D pa
irs are t
he
maj
o
r issu
e
s
that nee
d to
be
resolv
ed for
th
e i
m
p
l
e
m
e
n
tati
on
of IA from
theory to
pr
act
i
cal. T
her
efore
,
in this
pa
per,
w
e
prop
ose
a
n
effective D2
D
discov
e
ry sche
m
e t
o
overc
o
me this pr
ob
le
m,
w
h
ich can
me
asure th
e CSI
and
C-CSI b
a
s
ed
on th
e
discov
e
ry
mess
ages.
Si
mul
a
tio
n
re
sults sh
ow
that, und
er
perf
e
ct con
d
itio
ns,
eve
n
th
oug
h
the
prop
osed
D2D
discovery sch
eme incr
eases
the nee
de
d ti
me sl
ots to es
tablis
h D2D c
o
mmu
n
icati
ons
by
6.2% c
o
mpar
e
d
w
i
th the
co
nv
entio
nal
D
2
D
d
i
scovery
w
i
tho
u
t cons
ider
in
g I
A
, it incr
eases
the thro
ug
hput
up
to 50% tha
n
the conve
n
tio
nal
D2D w
i
thout co
nsid
erin
g IA, i.e., it improv
es the spectru
m
ef
ficiency.
Ke
y
w
ords
: D2D; Interference alignm
ent; D2D
discov
e
ry; Chan
nel state i
n
formatio
n
Copy
right
©
2016 Un
ive
r
sita
s Ah
mad
Dah
l
an
. All rig
h
t
s r
ese
rved
.
1. Introduc
tion
As device
-
to-device (D2
D
) commu
nicati
ons
have the
advantag
es o
f
suppo
rting a
highe
r
data rate, lo
weri
ng com
m
unication d
e
lay, and re
duci
ng ene
rg
y consumpti
on, it has b
een
con
s
id
ere
d
a
s
a key tech
n
o
logy for future 5G mobil
e
communi
cati
on system
s [1-2]. Hen
c
e,
co-
cha
nnel inte
rferen
ce (CCI) amon
g D2
D u
s
ers a
r
e
major i
s
sue
s
that mu
st be re
solve
d
for
su
ccessful t
r
ansmi
ssion
when th
ey a
r
e
utilizin
g th
e
same
fre
que
ncy b
and
s. T
o
redu
ce
the
co
-
cha
nnel
interf
eren
ce
am
on
g D2D u
s
ers,
seve
ra
l
tech
nique
s
have
been
propo
sed in
the
existing
literature. In [3], a novel lo
cation
-ba
s
e
d
resou
r
ce
allo
cation
schem
e is p
r
op
ose
d
. In this pa
p
e
r,
the D2
D use
r
s whi
c
h utilize the same f
r
equ
en
cy
ba
nds u
s
ually h
a
ve long dist
ance, therefo
r
e,
the interfe
r
en
ce
s fro
m
oth
e
rs’
tran
smitt
e
r a
r
e
re
duced be
ca
use
of the pat
hl
o
s
s of inte
rference
s
i
gnal. In [4],
an effec
i
tve time-domain interfer
e
n
ce co
ordin
a
tion (I
CIC) ap
proach
wa
s propo
se
d
by using alm
o
st blank subframes (ABS). Ev
en though the above proposed
schemes can
reduce
the interferen
ce
s amon
g D2D u
s
ers, th
e
spe
c
trum eff
e
icie
ny beco
m
es lo
w.
Re
cently, a
new inte
rfere
n
ce
ma
nage
ment
sc
hem
e called
inte
rferen
ce
align
m
ent (IA
)
wa
s p
r
e
s
ente
d
in [5] to
ali
gn the
interfe
r
en
ce f
r
om
the different trans
m
itte
rs
in
a s
p
e
c
tific s
i
gna
l
demen
sio
n
. T
hen, the rem
a
inde
r of the
dimen
s
ion
be
come
s inte
rfe
r
en
ce
-fr
ee sp
ace.
T
herefore,
IA can achie
v
e high multiplexing gain
and deg
re
es freedom of the chan
nel.
Howeve
r, the
majority of th
e pa
pers [6
-1
1] whi
c
h
co
n
s
ide
r
th
e
IA tech
nolo
g
for
conve
n
tional
cellul
a
r
network
or D2
D com
m
unication
s, they all assu
med that
the
chan
nel stat
e informatio
n
(CCI
) betwe
en
transmitter a
nd re
ceiver
of
a D2
D pair
a
nd cros
s-ch
a
nnel state
inf
o
rmat
io
n (C-CCI)
am
ong
D2
D
pairs a
r
e
kno
w
n to
ea
ch ot
her. Th
erefore, an effe
ctive metho
d
whi
c
h
can
mea
s
ure th
e CCI a
n
d
C-CCI is in
evitable for the a
pplcati
on of IA from theory
to practial.
Fortunately, D2D discov
ery in D2D comm
uni
cat
i
ons provides possibility for the
measurement
of CSI and C-CSI amo
ng D2
D pai
rs. Becau
s
e
an accu
rate
descriptio
n
of
sign
aling m
e
ssag
e exchan
ge bet
wee
n
t
he entitie
s of
the network
with recpe
c
t to the n
e
cessaary
informatio
n f
o
r id
entifying
a ne
w
D2
D p
a
irs.
T
he
exchang
e of di
scovery me
ssa
ges will
provide
the netwo
rk
with inform
ation abo
ut the
CSI
and C-CSI. Therefo
r
e, ba
sed
on
the descripti
on
provide
d
by
above, in thi
s
p
ape
r, we
pro
p
o
s
e a
n
effective D2
D di
scovery
for the
ch
an
nel
measurement
of IA, which
not only can
measur
e the CSI, but also
c
an me
asure the C-CSI.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Efficient D2D
Disco
v
e
r
y S
c
hem
e for Cha
nnel Mea
s
u
r
e
m
ent of Interferen
ce
… (Zh
anjun Li
u)
549
2. Sy
stem Model and De
scription o
f
IA
We con
s
ide
r
a single
-
cell netwo
rk
with N D2
D users,
and we a
s
su
me that each
of D2D
use
r
h
a
s two
anten
na
s (
M=
2
).
The
max
i
mum n
u
mbe
r
of
D2
D p
a
irs
whi
c
h
ca
n f
o
rm
a g
r
ou
p
to
impleme
n
t IA is three [12]. For the conv
enien
ce of
no
tation, we rep
r
esent the three D2
D pai
rs in
the same
gro
up a
s
D2
D
1
, D2D
2
an
d D2
D
3
. In the di
scovery
proce
ss,
ea
ch
D2
D u
s
e
r
tra
n
smit
s
discovery m
e
ssage
s
with t
he ide
n
tity of itself and
targ
et user
on th
e selecte
d
di
scovery
re
sou
r
ce
block
(DRB),
and the
DRB
is b
r
oa
dca
s
te
d by t
he e
N
B
.
Con
s
eq
uent
ly, all the oth
e
rs’
D2D
use
r
s
can receive the discove
r
y messag
e if they
can de
co
d
e
t
h
is me
ssa
ge su
c
c
e
ssf
ul
ly
.
The inte
rfere
n
ce
ch
annel
model fo
r three D2D
p
a
irs is sh
own in
Figure 1. We
define th
e
three p
r
ecodi
ng matri
cs a
s
12
3
,,
VV
V
. To decode
the
M/2 =1
n
on-inte
rferen
ce stream
s al
ong the
colum
n
v
e
ct
ors
of
i
V
from the
M=
2
co
mpone
nts of
the
re
ceive
d
vecto
r
, the
dime
nsio
n
of
interferen
ce has
to be
le
ss
than or equ
al
to
M/2=1.
The follo
wing
three inte
rference alig
nme
n
t
equatio
ns e
n
s
ure that the dimen
s
ion of
the interferen
ce is e
qual to
M/2=1
at all the re
ceive
r
s.
12
2
1
3
3
()
(
)
s
pan
H
V
span
H
V
(1)
21
1
2
3
3
H
VH
V
(2)
31
1
3
2
2
H
VH
V
(3)
Whe
r
e
()
s
pa
n
A
repre
s
ent
s the vector spa
c
e
spa
nned by the column vecto
r
s of matrix
A
.
11
H
21
H
22
H
12
H
2
2
x
V
1
1
x
V
1
1
11
x
V
H
2
2
12
x
V
H
2
2
22
x
V
H
1
1
21
x
V
H
31
H
32
H
13
H
23
H
33
H
3
3
x
V
3
3
13
x
V
H
3
3
23
x
V
H
2
2
32
x
V
H
3
1
31
x
V
H
3
3
3
x
V
H
Figure 1. Example of interf
eren
ce
cha
n
n
e
ls for three
D2
D pairs
The sele
ction
of
i
V
for
satisf
ying the
abo
ve the e
quati
ons is the
ke
y point in
th
e IA.
Since the
ch
annel h
a
s
a full ran
k
of
M=
2
, the ab
ove equatio
ns
can be
e
quival
ently represe
n
ted
as:
11
()
(
)
s
pan
V
s
pan
EV
(4)
21
VF
V
(5)
31
VG
V
(6)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 14, No. 2, June 20
16 : 548 – 55
4
550
Whe
r
e:
11
1
31
32
12
13
23
21
()
()
(
)
,
EH
H
H
H
H
H
1
32
31
()
,
F
HH
1
23
21
()
.
GH
H
The re
ceive
d
sign
al at
i
receiver for the IA communi
ca
tions is give
n by
：
()
i
i
iii
i
j
j
j
i
Int
e
rf
e
r
ence
YH
V
x
H
V
x
N
i
j
(7)
Whe
r
e
i
N
is the white Ga
ussian n
o
ise.
Then, the interferen
ce ca
ncel
ation mat
r
ix
i
U
is used fo
r can
c
eli
ng th
e interferen
ce from
the received
sign
al of (7) f
o
r the
i
reciver is given by:
()
0
H
ii
j
j
UH
V
(8)
()
H
ii
i
i
i
i
i
i
YU
H
V
x
U
N
(9)
From
the abo
ve
analy
s
is, we ca
n see
that
the mea
s
urem
ent
of CCI
an
d C-CCI
for
th
e
acquisition of
i
V
and
i
U
are in
evitable, and
becau
se
the
detailed
cal
c
ulation of
i
V
and
i
U
beyond the
scop
e of this p
aper, the
r
efor
e, we do not
make fu
rther
analysi
s
.
3. Proposed
D2
D Discov
e
r
y
Scheme
The envole
d
node ba
se
station (e
NB) i
s
very
import
ant beca
u
se it can coo
r
di
nate the
messag
e in
some
step
s of
the sch
e
me,
and th
e
who
l
e D2
D
disco
v
ery sche
me
is di
spl
a
yed
in
Figure 2.
Figure 2. The
propo
se
d D2
D discove
r
y schem
e
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Efficient D2D
Disco
v
e
r
y S
c
hem
e for Cha
nnel Mea
s
u
r
e
m
ent of Interferen
ce
… (Zh
anjun Li
u)
551
Step 1:
1_
2
Tx
DD
ran
domly sele
cts a
discove
r
y res
our
ce
t
o
br
oad
ca
st
it
s di
scov
e
ry
messag
es which
inl
c
ude
s the id
entity of itself
an
d targ
et u
s
e
r
1_
R
2
x
DD
.The
2_
2
Tx
DD
and
3_
2
Tx
DD
follow the s
a
me proc
ess
.
Step 2:
1_
R
2
x
DD
sen
d
s to
e
N
B
with the
re
ceive
d
si
gnal
to i
n
terferen
ce
plu
s
n
o
ise
rati
o
(SINR)
a
nd measured CCI.
Furthe
rm
ore,
2_
2
R
x
DD
and
3_
R
2
x
DD
se
nd the m
e
a
s
ue
d C-CCI
betwe
en
1_
T
2
x
DD
an
d
2_
R
2
x
DD
&
3_
R
2
x
DD
resp
ectively to the eNB. In this step, if the
messag
es a
r
e not
re
ceive
d
by
1_
2
R
x
DD
&
2_
R
2
x
DD
&
3_
R
2
x
DD
, whi
c
h
mean
s that the
rece
ived
SINR i
n
1_
2
R
x
DD
&
2_
R
2
x
DD
&
3_
R
2
x
DD
is not
satisfied,
retra
n
smissi
on
s
occurs.
The
2_
2
R
x
DD
and
3_
2
R
x
DD
also sen
d
the simila
r messag
es to
eNB and foll
o
w
the sa
me p
r
ocess. In this pap
er,
we a
s
sume that the D2
D
Rx
can de
co
de the discove
r
y messag
es
su
ccessfully.
Step 3:
Th
e eNB cal
c
ul
ates the pre
c
o
der vecto
r
i
V
and interferen
ce ca
ncelatio
n matrix
i
U
ba
sed
on
th
e me
asued
CCI an
d
C-CCI of all
th
e D2D users.
T
h
en,
the
eNB
sen
d
s
sy
ste
m
informatio
ns
to all the D2D
Tx
and D2D
Rx
with the pre
c
od
er
vector
i
V
and interfere
n
ce
can
c
el
ation matrix
i
U
, res
pec
tively.
Step 4:
All
the
D2
D u
s
ers
no
mat
t
er tan
s
mitte
rs
and
rceiv
e
rs intiate t
he
D2
D
comm
uni
cati
on by usin
g the IA technol
ogy after
re
ce
iving the syst
em informatio
n from eNB.
As multipl
e
D2
D p
a
irs
m
a
y se
nd
di
scovery
me
ssa
ges con
c
u
rre
ntly, which fo
llows th
e
rand
om a
c
ce
ss
process,
thus the
D2
D
Tx
sh
ould t
r
an
smit the
discovery m
e
ssag
es i
n
the
begin
n
ing of each time slo
t
by using a certain tr
a
n
smi
ssi
on proba
bi
lity to avoid
the colli
sion.
We
assume
that
su
ccessful
di
scovery me
ssge rece
ption
can ta
ke
pla
c
e if there i
s
n
o
colli
sio
n
. Le
t
N
denote
s
the
numbe
r of D2D pai
rs
whi
c
h will p
o
tent
ially particip
a
t
e in the discovery pro
c
e
s
s at
the same time.
N
al
so
stan
ds fo
r th
e
nu
mber of
disco
v
ery me
ssag
es
sim
u
ltane
ously t
r
an
smi
tted,
and it shoul
d be estimat
ed by the D2D
Rx
. Als
o
, l
e
t
t
p
denotes the optimal transmi
ssi
o
n
probability that can be utilized by D2D
Tx
. The probability of a successful transmi
ssi
on of
a
discovery me
ssage in o
n
e
time slot for one D2
D
Tx
i
s
the probab
ility that only one D2
D
Tx
is
transmitting while all the o
t
her D2
D
Tx
are idle. This p
r
obability is gi
ven by:
1
(1
)
N
succe
s
s
t
t
pp
p
(10)
The o
p
timal tran
smi
ssi
on
prob
ability is
the one th
at
maximize
s th
e su
cce
ss
probability
in (10), i.e. [1
3]:
1
t
p
N
(11)
No
w, let
X
denotes the n
u
m
ber of successful tran
sm
issi
on of one
D2
D
Tx
in
n
t
i
me slot
s.
In ord
e
r to i
m
pleme
n
t IA, the p
r
ob
abili
ty of hav
ing
more
than
on
e succe
s
sful
transmissio
n
in
n
time s
l
ots
is
:
Pr
[
1
]
1
Pr
[
0
]
1(
1
)
n
su
c
ces
s
XX
p
(12)
If we nee
d to sati
sfy (12
)
with the
su
cce
ss
pro
babil
i
ty higher tha
n
in
n
time
s
l
ots,
then:
Pr
[
1
]
X
(13)
Thus, we
hav
e:
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 14, No. 2, June 20
16 : 548 – 55
4
552
ln
(
1
)
ln
(
1
)
s
u
c
cess
n
p
(14)
The re
sult of
(14) give
s the numb
e
r o
f
needed tim
e
slots
whi
c
h
one D2
D u
s
er
can
transmit successfully. Furt
herm
o
re,
a
s
t
he n
u
mb
er
of su
cce
ssful
transmi
ssion
f
o
r
our p
r
op
osed
scheme i
s
at least three, therefo
r
e, we h
a
ve:
31
0
Pr
[
3
]
1
(
)
(
1
)
Xn
X
su
c
c
es
s
s
u
c
ce
s
s
X
n
Xp
p
X
(15)
Whi
c
h lea
d
s:
2
0
2
2
2
2
Pr[
3
]
1(
)
(
1
)
1
(1
)
)
(
1
)
(1
)
(1
)
(
1
)
(1
)
1
0(
1
)
1
l
n
(
1
)
0
,
(
ln
Xn
X
su
cc
es
s
s
u
c
ce
s
s
X
n
s
u
c
cess
s
u
c
cess
s
u
c
cess
n
su
c
c
e
ss
su
cc
ess
s
u
c
cess
s
u
c
cess
su
cc
ess
s
u
c
c
e
s
s
X
n
pp
X
pp
p
p
p
p
pp
pp
n
2
2
1)
(
1
)
/l
n
(
1
)
.
1
s
u
cc
es
s
su
cce
s
s
su
cc
ess
s
u
c
c
e
ss
p
p
pp
(16)
The result of
(16
)
give
s th
e num
ber of
need
ed time
slots
whi
c
h
a
gro
up
can i
m
pleme
n
t
I
A
suc
c
e
ssf
ull
y
.
4. Performan
ce Ev
aluation
In this se
ctio
n, we de
scrib
e
the simul
a
ti
on enviro
n
me
nt and re
sult
s of the perfo
rman
ce
for the prop
o
s
ed
sch
eme.
Firstly, from the Figure
3
,
we can get
that the satisfactio
n
of three
D2D transmit
ters’
successful sending
will incr
ease the time delay co
m
pared with one
D2D
transmitter’s su
ccessful se
nding
by
6.2
%
, which
ca
n
also
see fro
m
the equ
atio
ns (14
)
and
(16),
i.e., obviously:
2
2
(1
)
(
1
)
ln
(
1
)
ln
/
l
n
(
1
)
.
ln
(
1
)
1
su
cc
es
s
su
cc
es
s
su
cce
s
s
su
cc
es
s
s
u
c
cess
p
p
p
pp
Figure 3. Nee
ded time slot
s with the in
creased
num
be
r of simultan
e
ous tra
n
smitted discove
r
y
messag
es
3
4
5
6
7
8
9
10
10
20
30
40
50
60
70
80
N
u
m
ber
of
S
i
m
u
l
t
aneous
T
r
an
s
m
i
t
t
ed
D
i
s
c
o
v
e
ry
M
e
s
s
ages
N
eeded T
i
m
e
S
l
ot
s
3
su
cce
ssf
u
l
t
r
a
n
s
m
i
sssi
o
n
1
su
cce
ssf
u
l
t
r
a
n
s
m
i
ssi
o
n
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Efficient D2D
Disco
v
e
r
y S
c
hem
e for Cha
nnel Mea
s
u
r
e
m
ent of Interferen
ce
… (Zh
anjun Li
u)
553
Then, we an
alyze the p
e
rforman
c
e
co
mpari
s
o
n
wit
h
the obtain
e
d
throu
ghp
ut after the
establi
s
hm
en
t of D2D com
m
unication
s.
The
detaile
d
simulatio
n
pa
ramete
rs a
r
e
listed i
n
T
abl
e 1.
In addition,
we a
s
sume th
a
t
the pe
rform
ance de
gr
ada
tion of am
plifying noi
se i
s
negligibl
e
d
u
e to
the
i
U
, i.e,
()
H
ii
UN
in equatio
n (9
), which is
be
cau
s
e th
e o
p
timal deg
re
es freedo
m
of the
cha
nnel can
be obtaine
d in the case of high sig
nal
-to
-
noi
se (S
NR) [14]. Furthermore, we use
the
loc
a
tion of the D2D
Rx
as the refere
nce
locatio
n
of th
e D2
D p
a
ir,
becau
se the
relative di
sta
n
ce
betwe
en the
D2
D
Tx
and
D2
D
Rx
of a D2
D pai
r i
s
signifi
cantly
smalle
r tha
n
the cell rad
i
us.
Therefore, we can rega
rd
the coo
r
din
a
tes of D2
D
Rx
as the po
sition
of a D2D pai
r.
Table 1. Simulation Para
meters
Parameters
Val
ues
Carrier Fre
quenc
y
2.62
GHz
Number of
RBs
1
Number of
T
x
an
tennas
2
Number
of Rx an
tennas
2
Cell Radius
500 (m)
Path Loss
PL(dB)=35.4+22
.
7
log10(R)
Shado
w
i
ng Stan
dard
Deviation
4 dB
Noise Figure
5 dB
Noise Spectrum
Densit
y
-175 dBm/Hz
Max
i
mum D2
D Tx
Po
w
e
r
23 dBm
Traffic Model
Full buffer
Carrier Fre
quenc
y
2.62
GHz
Number of
RBs
1
Figure 4. Through
put of D2D pai
rs:
With IA
and With
out IA
Acco
rdi
ng to Figure 4, we
can o
b
serve t
hat
the imple
m
entation of
IA improves t
he D2
D
pairs’ throug
h
put co
mpa
r
ed
with no
IA by 50%. Th
is is
becau
se by u
s
ing IA, on
e reso
urce bl
o
ck
(RB)
ca
n su
p
port thre
e D2
D pai
rs’
com
m
unication at
the same ti
me, and ea
ch D2
D pai
rs
can
get 1/2 of RB without i
n
te
rferen
ce
s among different D2D
pairs. Ho
we
ver, for the D2D
comm
uni
cati
ons witho
u
t
IA, each
D2
D p
a
irs
ca
n
only get
1/3 of RB
by
usin
g o
r
thog
onal
freque
ncy ba
nd.
5. Conclusio
n
In this pa
pe
r, an effective D2
D di
sco
v
er
y schem
e
is p
r
opo
se
d
for the p
u
rpose of
measuri
ng the CSI and C-CSI amon
g D2
D pairs, wh
ich is u
s
e
d
for the imple
m
entation of IA.
Firstly, the D2D u
s
e
r
s
me
asu
r
e the
CSI and C-CS
I by using th
e feedb
acks
of the discov
ery
messag
es f
r
o
m
each othe
r, and then
se
nd the
s
e fe
e
dba
cks to th
e ba
se
statio
n. Furthe
rmo
r
e,
the ba
se
station calculate
s
the tran
smitter an
d
receiver p
r
e
c
odin
g
s
of IA by usi
ng the
CSI and
C-CSI. Finall
y
, the base
station se
nd
s these calc
ulat
ed precodin
g
s to all the
D2D u
s
e
r
s
ba
sed
on the flexible s
o
lution of IA.
Simulation res
u
lt
s show that, un
der p
e
rfe
c
t con
d
ition
s
, the
prop
osed
D2D di
scovery sch
eme l
o
we
rs th
e
need
ed tim
e
slot
s to
establi
s
h
D2
D
comm
uni
cati
ons
co
mpa
r
e
d
with th
e convention
a
l
D2
D di
scove
r
y witho
u
t consi
deri
ng IA
by
6.2%, it incre
a
se
s the
thro
ughp
ut up to
50% t
han th
e
conve
n
tional
D2
D
comm
u
n
icatio
ns
with
out
con
s
id
erin
g IA, i.e., it improves the spe
c
trum effici
en
cy.
Ackn
o
w
l
e
dg
ements
This
wo
rk is
sup
porte
d by
the National
Natu
ral S
c
ie
nce
Fou
ndati
on of
China
(NSF
C)
6140
1053.
Referen
ces
[1]
F
eng D, Lu L, F
eng G. Device-to-Devic
e
co
mmunicati
ons i
n
cellu
lar n
e
t
w
orks.
IEEE Comm
unication
Maga
z
i
ne
. 2
0
1
4
; 52(4): 49-5
5
.
0
1
2
3
4
5
6
7
8
x 1
0
6
T
o
t
a
l
T
h
r
oug
hput
o
f
D2
D P
a
ir
s
(
bps
)
W
i
th
I
A
W
i
th
o
u
t I
A
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 14, No. 2, June 20
16 : 548 – 55
4
554
[2]
T
ehrani M, U
y
sal M, Yaniko
merogl
u H. De
vice
-to-Dev
i
ce
commmunic
a
ti
on in 5G ce
llu
l
a
r net
w
o
rks:
Chal
le
nges, So
lutio
n
s, and F
u
ture Directi
ons.
IEEE Comm
unication Maga
z
i
ne
. 20
14; 52(
5
)
: 86-92.
[3]
Hu
L.
Res
ourc
e
all
o
cati
on fo
r netw
o
rk-assi
sted devic
e-to-
devic
e disc
ove
r
y
. IEEE VITA
E. Aalbor
g.
201
4: 1-5.
[4]
Babu
n
L.
Intercell i
n
terferenc
e coord
i
n
a
tion
for D2D disc
overy in LT
E-
A HetNets
. IEEE WCNC.
Istanbul. 2
014:
2202-
22
07.
[5]
Viveck R Ca
damb
e
, A Jafar. Interferenc
e al
i
gnme
n
t and d
egre
e
o
f
freedom of the K-User
interfere
n
ce ch
ann
el.
IEEE Tr
ansaction on Inform
ation Theory
. 2008; 52(
8): 3425-
34
41.
[6]
Maso M, D
ebb
ah M, Va
ng
elis
ta L. A d
i
stribu
t
ed a
ppro
a
ch t
o
interf
erenc
e
alig
nme
n
t in
OF
DM base
d
t
w
o-ti
ere
d
net
w
o
r.
IEEE Transaction on Vehicular Technology.
2013; 6
2
(5)
:
1935-1
9
4
9
.
[7]
Shar S, Chatzi
S, Otter B.
Interference
alignment
for spectral coex
istence
of heterogeneous net
w
o
rk.
EURASIP J. Wirel. Co
mmu
n
. Netw
.
2013; 6:125.
[8]
An J, He G, Xu Y. Distri
buted c
o
o
pera
t
ive
multic
ell
preco
d
in
g b
a
s
ed o
n
l
o
cal c
han
nel st
at
e
information.
T
E
LKOMNIKA T
e
leco
mmunic
a
tion, Co
m
puti
ng, Electro
n
ics
and Co
ntrol
. 201
5;
13(3):
851-
858.
[9]
Hussai
n
E, Khale
d
F
,
Mahmoud H.
Explo
i
ti
ng interfer
ence
align-
ment for sum rate en
h
ance
m
ent in
D2D-e
n
a
b
le
d c
e
llu
lar n
e
tw
ork
. IEEE WCNC.
Shan
gh
ai. 201
2: 1624-
16
29.
[10]
Yunp
en
g L,
KH C.
Interferenc
e al
ig
n
m
e
n
t w
i
th multipl
e
a
n
ten
n
a
s for sin
g
le
grou
p D2
D
communic
a
tio
n
s
. KICS W
i
nter Confer. Inche
on. 201
5: 136-
138.
[11]
Li
X, Liu F
,
Va
n N. Enha
nce
d
chan
nel
estim
a
tion
alg
o
rithm
for dedic
a
ted
short-ran
ge co
mmunicati
on
s
y
stems.
T
e
lko
m
n
i
ka T
e
l
e
co
mmu
n
ic
ation, Co
mp
utin
g, Electronics a
nd C
ont
rol
. 201
5; 13(3)
: 870-87
9.
[12]
Che
n
S, Ro
ge
r S. Cluster
ing
for interfer
en
ce al
ig
nment
i
n
multi
u
ser
int
e
rferenc
e n
e
t-w
o
rk.
IEEE
T
r
ansactio
n
on
Vehicu
lar T
e
c
hno
logy
. 20
14
; 63(6): 261
3-2
324.
[13]
Ahli
n L, Z
a
n
d
e
r
J Slima
n
e
.
Pr
incip
l
es of
W
i
re
less Communications
.
Lu
nd,
S
w
e
d
en: Stud
entlitteratur.
200
8.
[14]
A
y
ac
h O, Heath W
.
T
he practical
ch
alle
ng
es of i
n
terferenc
e al
ignm
ent.
IEEE Wireless
Co
mmun
icati
o
ns
. 2013; 2
0
(1)
:
35-42.
Evaluation Warning : The document was created with Spire.PDF for Python.