Communications and Network, 2013, 5, 187-193
http://dx.doi.org/10.4236/cn.2013.53B2036 Published Online September 2013 (http://www.scirp.org/journal/cn)
A Resource Reuse Scheme of D2D Communication
Underlaying LTE Network with Intercell Interference*
Jia Liu1, Bingbing Li1, Bing Lan1, Junren Chang2
1National key lab of ISN. Xi’an, China
2Huawei Technol. Co., Ltd., Beijing, China
Received June, 2013
ABSTRACT
With the growing concern on data rates and resource utilization, Device-to-Device (D2D) communication has been
raised in 3GPP Long-Term Evolution (LTE) networks. In order to limit severe interference, previous studies mainly
focus on intra-cell interference that between cellular link s and local D2D links. In this paper, we consider both intra-cell
interference and inter-cell interference between D2D and cellular links. We propose a new resource reuse algorithm that
D2D users reuse the minimum interference uplink (UL) Semi-Persistent Scheduling (SPS) resources to reach the high-
est throughput. Th e simulation results show that this scheme reduces interference as well as improves throughput.
Keywords: Device-to-device; Inter-cell Interference; Interference Reduction; Semi-Persistent Scheduling; Resource
Sharing
1. Introduction
Device-to-device (D2D) communication as an underlay
coexisting with cellular networks has been proposed in
order to improve the utilization of the spectrum [1-4].
There are four resource allocation methods in D2D
communication underlaying cellular networks: cellular
mode, dedicated resource mode, reusing the resource of
only one cellular user, and reusing the resources of more
than one cellular user [4]. The first two modes bring no
interference to system because they use orthogonal re-
source. In the last two modes which belong to reuse
mode, D2D users share the same resource with cellular
users, so they interfere each other. The interference in-
cludes intra-cell interference (the interference between
attached cellular users and D2D users) and inter-cell in-
terference (the interference between neighbor cellular
users and attached cellular users; the interference be-
tween neighbor cellular users and attached D2D users;
the interference between neighbor D2D users and at-
tached cellular users; the interference between neighbor
D2D users and attached D2D users). However, the net-
work may achieve maximum resource utilization in reu se
mode.
When D2D users share the downlink (DL) resource
with cellular user, cellular users and neighbor cellular
users may suffer from interference by D2D users, and
D2D receiver may suffer from interference by eNodeB
(eNB). On the other hand, when D2D users share the
uplink (UL) resource with cellular, the eNodeB is the
victim of interference by D2D users, and D2D receiver
may suffer from interference by cellular users and
neighbor cellular users.
In order to reduce these interferences, a lot of work have
been done [5-11]. In [5], Xiao xiao proposes a power
optimization scheme with joint resource allocation and
mode selection in an OFDM system with integ rated D2D
communications, aiming at optimizing downlink power
consumption. In [6], Yu considers rate splitting and in-
terference cancelation in D2D communication underlay-
ing a cellular network. They assume that a transmitted
message is split into a private and a public part, using
fractions i

and 1i


of the total transmit power,
respectively. They also derive the optimal rate splitting
factors for most of the categorized channel conditions in
a two-link scenario. In [7], Janis, P. Koivunen proposes a
practical and efficient scheme where the D2D terminals
make power measurements during the uplink (UL) phase
of the cellular network in normal operation. In [8], H.
Wang and X. Chu propose a distance-constrained re-
source-sharing criterion for the base station to select a
cellular user for a D2D link, with the cellular-to-D2D
interference controlled by keeping a minimum distance
between them. In [9], the joint mode selection and power
allocation scheme is proposed to maximize the utility
function. In [10], Zulhasnine, M. formulates the problem
*This work was supported by the Huawei Innovation Research Pro-
gram, the Fundamental Research Funds for the Central Universities
(K5051301034) , and the 111 Project (B08038).
C
opyright © 2013 SciRes. CN

J. LIU ET AL.
188
of RB allocation to the D2D communication as a mixed
integer nonlinear programming (MINLP), and proposes
an alternative greedy heuristic algorithm to solve it. In
[11], Interference Alignment (IA) is used in a D2D un-
derlay network to enhance spectral efficiency. But in
these papers, authors do not consider inter-cell interfer-
ence to D2D users.
In this paper, we consider both intra-cell interference
and inter-cell interference between D2D and cellular
links in D2D reuse mode. A new resource reuse algo-
rithm that D2D user reuse the UL SPS resource is pro-
posed. We choose the minimum interference resource to
reach high throughput. It’s shown that our algorithm can
bring less interference and higher throughput.
The rest of this paper is organized as follows. In Sec-
tion 2 we give the system model of D2D communication
underlay LTE network. In Section 3, we formulate our
resource algorithm and mechanism. In Section 4, some
numerical results are given. Finally, some concluding
remarks are drawn in Section 5.
2. System Model
In this paper, we assume D2D users reuse LTE UL SPS
(Semi-Persistent Scheduling) resource. When D2D users
reuse UL resource, D2D receiver may suffer from inter-
ference by attached cellular users, neighbor cellular users,
and neighbor D2D users. An example of interference
scenario in the uplink is given in Figure 1. CUE denotes
cellular users and DUE denotes D2D users. D2D users
are in anchored eN B (A- eNB ) an d its ne ighbor eN B is N-
eNB. Communication links are indicated by the solid line
while interference links are indicated by the dotted line.
DUE2 may suffer from interference by , ,
, , , and . And when
trans mit to A-eNB, A-eNB may suffer from inte r-
ference by , and DU . As in Figure 1
the interference of in N-eNB to D2D link is
strong, so we should avoid choosing the same resource of
for D2D link.
1
1
CUE
E3
2
1
CUE
3
1
CUE1
1
CUE
1
2
CUE
1
2
CUE
DU
2
2
CUE
E1 CU
CU
3
2
CUE
1
2
E1
2
E
DU
E3
Some useful symbols are defined as:
,
dc
PP
The transmission power of D2D user and cel-
lular user
dd
G A channel gain of D2D link
cB
G A channel gain of cellular link
j
i
y A binary variable which satisfies 1
j
i
y

if use
r
i and user j use the same resource
o
N Noise power at the receiver
i User device index in attached cell
User device index in neighbor cell
j
Considering intra-cell interference and inter-cell inter-
ference, the UL SINR of D2D link and cellular link can
be expressed by:
0
UL ddd
ddd D
ii idjjjddddd
ijd
PG
NyPGyPG yPG
  
 
(1)
0
cccB
eNB cd c
iiiBi dddjjjB
id j
PG
NyPG yPGyPG
  
 
(2)
Because SPS resource has its own regularity, we
choose it to reuse. SPS is a feature that significantly re-
duces control channel overhead for applications that re-
quire persistent radio resource allocations such as VoIP.
It means that the size of packets and the arriving time
intervals are constant over a period of time, and users are
allocated with resource periodically. The period length is
the time that a traffic occupies resource periodically.
Usually, the algorithm of resource selection might take
some time. The most likely scenario is that we compute
RB (i) of user A is reasonable to reuse, but cannot reuse
RB (i) because this RB is not assigned to user A when
starting the step of sharing resource. Using SPS resource
can solve this problem due to its fixed resource assign-
ment in a period of time.
1
1
CUE
2
1
CUE
3
1
CUE
1
2
CUE
2
2
CUE
3
2
CUE
Figure 1. An example of interferences in the LTE uplink.

J. LIU ET AL. 189
Taking typically VoIP as an example, the packets ar-
riving time interval is 20 ms. eNB gives SPS scheduling
indication to users through PDCCH, then users can
transmit or receive data in this schedule and transmit or
receive new VoIP data on the same resource after every
20 ms. The SPS resource schedule is given in Figure 2.
In this paper, we assume eNBs can exchange SPS re-
source allocation information through X2 interface.
A-eNB controls D2D pairs reuse SPS resource, consults
with N-eNBs, and measures correlates, then selects the
most appropriate SPS resource to reuse.
3. Resource Reuse Algorithm Considering
Inter-cell Interference
In this section, we will describe the SPS resource reuse
algorithm in details. and define D2D
users. The proposed scheme is presented as follows.
DUE1 DUE2
Step 1: All cellular users and D2D users register to its
anchored eNB (A-eNB) and A-eNB reports its SPS re-
source using information to its neighbor eNBs (N-eNBs)
by interface X2.
Step 2: According to the information obtained from
interface X2, N-eNBs decide which resource can be used
as SPS resource. In this way, A-eNB and N-eNBs would
use the same SPS resource.
Step 3: D2D users ( and ) report their
position information to A-eNB.
DUE1 DUE2
The position information can be obtained from GPS
(Global Position System) or A-GPS (Assisted GPS). For
example, the coordinates of DU and are E1 DUE2 11 (,)
x
y,
22
(, )
x
y
dd
. If the distance between and
1,20 (0
d is the threshold), and the channel condi-
tion is good (we can use a detecting signal, and judge if
the block error rate (BLER) is less than a threshold), they
can form a D2D pair, and then go to next Step. As what
is said above,
DUE1 DUE2
2
1,22121
()(dxxyy
2
)
pairs is
(3)
Step 4: A-eNB computes the distance between cellular
users and D2D users, choose n (e.g. three) maximum
distance, and takes their corresponding SPS resource as
candidates. The distance between cellular users and D2D
22 12 12
,()()
xx yy
dx y

  (4)
e n SPS resource groups
an
terference to neighbor D2D users on the
SP
chosen users in step6 and gives feedback to
A
its corresponding resource as the SPS resource to
re -eNB allocates the SPS resource group to
D: D2Dling and data can be transmitted
be of our proposed algorithm can
be l
ne as ana
22
cd cc
Step 5: A-eNB reports thes
d D2D positions to N-eNB.
Step 6: N-eNB searches its cell for users who use the
same resource, and then orders the corresponding users
compute the in
S resource.
Step 7: N-eNB receives the interference information
from the
-eNB.
Step 8: A-eNB searches the smallest interference and
choose
use.
Step 9: A
2D users.
Step 10 signa
tween DUE1 and DUE2 .
The complete procedure
illustrated in Figure 3.
It should be pointed out that, A-eNB may have severa
ighbors, and we should consider several N-eNBs reality.
Taking Figure 1 exmple. If the distance be-
tween two users ,0ij dd

(0 ddisthre,
an a D2D pair (DUE1 , DUE2 ). 1
1
CUE ,
2
1
CUE , 3
1
CUE are three cellular users in A-eNB, which
are farthest to the D2D pairs. The RB groups (RBG) as-
signed to them are RBG (1), RBG (3
spectively. In the neighbor cell, (1
2
CUE , 2
2
CUE , 3
2
CUE )
use this three RB groups. A-eNB and N-eNB exchange
their SPS resource allocation information through X2
erface. Assuming the transmitting power of user i is
i
P, and pathloss between c
user is d
i
PL . Then compute 2
22
min{/,1, 2, 3}
d tii
PPLi
and find i. So RBG (i) can be assigned to D2D p
is a tance shold)
they cform
G
Then D2D users can communication on this RBG.
te
the performance of a hybrid system by using our idea.
(2),
ellular
and RB
user i
), re-
and D2D
airs.
intthe
4. Simulation and Performance Analysis
In this section, simulation results are shown to evalua
Figure 2. The SPS resource schedule.

J. LIU ET AL.
190
Figure 3. The proposed SPS resource reuse scheme.
.1. Simulation Parameters
p-around-cell layout.
(5
where,
 is the configured maximum UE transmitted
powe

4
We consider a 7 hexagonal wra
Cellular users are randomly located in cells and D2D
users are located in cells’ edge. When computing the
inter-cell interference to D2D link, the 6 cells around
should be considered. The distance between D2D users is
less than 50m and the distance between two eNBs is
500m. There are 30 cellular users and 6 D2D pairs in
each cell. In this simulation, the 3-sector antenna is used
for each eNB. LTE power control scheme [13] is utilized
by controling the power of cellular users and a constant
is used to express the power of D2D users. The setting of
the UE transmit power for the physical uplink shared
channel (PUSCH) transmission in subframe i is defined
by
PUSCHCMAX10 PUSCH
O_PUSCH TF
( )min{,10log(())
( )()()()}
PiPM i
PjjPLifi 


)
CMAX
P
r. ()
PUSCH
M
i
as is the bandwidth of the PUSCH re- signment expressed in number of resource
jis a parameter composed of the sum
source
blocks valid fo r sub frame i.
 _OP P()
ell specific
USCH
of a c nominal component.
 For j =0 or 1, {0,0.4,0.5,0.6,0.7,0.8,0.9,1}


is
a 3-bit cell specific parameter provided by higher layers.
 e downlink pathloss estimate calculated in
the UE in dB.
PL is th
 ( )10log((21))
S
MPR K
P
USCH


TF offset
i1.25
S
K for

an0d for 0
S
K

where S
K
is given by the UE spe-
cific parameter deltaMCS-Enabled provided by higher
layers.

P
USCH

is a UE specific correction value, ()
f
i

(1) ()
P
USCH PUSCH
fii K


.

J. LIU ET AL. 191
There ar kind of th loss in a D2D scenario
[12]. The path loss model between cellular users and
eNB is C
e threepa
OST 231 Hata model, giving
()36.735lg( )pathloss dBd (6)
The path loss between cellular users and D2D users is
Xia model, giving
66.540lg( ),50
{100.720lg( ),
dd
pathloss dd 50


 (7)
And the path loss between two D2D users is free space
model, giving
38.420lg( )pathloss d
d is link distance in meter. The left parameters are pre-
se
4.2. Simulation Results and Discussion
es the sum throughput o f system only
users and
e, and the
he system throughput
(8)
nted in Table 1.
s
We will show some simulation results to confirm our
method’s advantages in this part.
Figure 4 compar
with cellular users, the system with cellular
D2D users that not handle inter-cell interferenc
system with cellular users and D2D users using the
proposed method. We observe that t
Table 1. System parameters.
Parameter Value
Noise Power Density -174 dBm/Hz
RB bandwidth
Carrier freque nc y 2 GHz
RB number
ain 14 dbi
d deviation
Antenna pattern (horizontal)
fixed antenna p
180 kHz
100
Max UE Power 200 mW
Min UE Power 3.2 mW
D2D UE Power 2 mW
Max BS antenna g
UE antenna gain 0
Shadowing standar8 dB
Between cells 0.5
Shadowing
correlation Inner cell 1
(For 3-sector cell sites with
atterns)
2
3
(
dB
()min[12),]
m
A
A

 where
degr dB.
ure5 dB
QPSK, 16QAM,64QA
7 (Data)


370 dB
 ees, 20
m
A
BS noise fig
Modulation and coding
scheme (MCS) M
1 Pilot +6 Number of symbols per slot
can be significantly improved by usinD communi-
cation. When using our resourse reuse algorithm con-
ell interfeut is
itional algorithm.
dy is that to D2D
g D2
sidering nei
higher than trad
ghbor crence, the throughp
The focus of our stuthe interference
users can be reduced and the throughput of cellular do
not be significantly reduced. This can be validated in
Figure 5 and Figure 6. Because D2D users are distrib-
uted in each cell edge and they share the same reuse with
cellular users which are far from them, the interference to
from D2D users to cellular users is small. Figure 7
shows the throughput per cellular user in the three cases
above. It can be seen that no matter handle neighbor cell
interference or not, cellular users’ throughput changes
little. Figure 8 shows the interference received by cellu-
lar users in the three cases above. The cellular users’ in-
terference stays the same. So we can conclude that our
algorithm has little effect on cellular users.
10 15 20 25 30 35 40 45
0
0. 1
0. 2
0. 3
0. 4
0. 5
0. 6
0. 7
0. 8
0. 9
1
W i thout D2D
not handle inter-cell interference
wit h consi derat ion of int er-cel l i nterference
t hroughput :Mbps
CDF
Figure 4. The sum throughput of system only with cellular
users, the system with cellular users and D2D users that not
handle inter-cell interference, and the system with cellular
users and D2D users using the proposed me thod.
00.5 11.5 22.5
0
0.1
0.2
0.3
0.4
0.5
0.6
1
0.7
0.8
0.9
c ell ul ar users throughput:M b ps
CDF
W i thout D2D
not handl e inter-cell interference
with consideration of inte r-cell interference
Figure 5. The throughput per cellular user.

J. LIU ET AL.
192
1
00.5 11.5 2
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
2.5
x 10
-9
0
0.1
CDF
not handl e inter-c ell interference
with c o nsidera tion of inter-cell interferenc e
without D2D
c el l ul ar us ers i nterference
Figure 6. The interference of cellular users.
1 23 4 56 7 8910
0
5
10
15
20
25
D2D pa ir s num be r
D2D users throu ghput (M bps)
wit hout consi derati on of inter-c el l int erferenc e
wit h consi derati on of inter-c el l i nterference
igure 7. The D2D throughput wi th D2D numbe r c hange d.
F
12345678910
0
0. 5
1
1. 5
2
2.5 x 10
-13
D2D pairs num ber
D2D user s interference
not handle i nter -c el l i nterferenc e
with considerati on of i nter-cell i nterference
Figure 8. The interference of D2D users with D2D number
changed.
To investigate the interference D2D users received, we
change D2D pairs’ number from 1 to 10, and plot the
D2D users’ interference (Figure 8). When D2D pairs’
number is small, the interference to D2D users is small,
too. But when the number increases, D2D users may
suffer from high interference by cellular users and
neighbor cellular users. It’s shown that our algorithm
greatly reduced the interference to D2D users.
5. Conclusions
In this paper, we proposed a new resource reuse algorithm
that D2D user reuse the UL SPS resource. We consider
both intra-cell interference and inter-cell interference
between D2D and cellular links, and it’s more closely to
practical conditions. By choosing the reasonable resource to
onal Pros-
Advanced Networks,” ICC Workshops,
.
for Device-to-Device
Communication Underlaying Cellular Networks,” Ve-
hicular Technology Conference), 15-18 May
2011.
s, V. Koivunen, Ribeiro, etc. “Interference-aware
reuse, the systerm throughput grows higher. Simulation
results show that the proposed algorithm significantly
improves systerm throughput but causes little effect to
the cellular users’ throug hput. Wh en we u se the p rop osed
algorithm, the D2D throughput can be increased and the
interference to D2D links is reduced.
REFERENCES
[1] K. Doppler, M. P. Rinne, P. Janis, C. Ribeiro and K. Hugl,
“Device-to-Device Communications; Functi
pects for LTE-
14-18 June 2009
[2] S. Hakola, T. Chen, J. Lehtomaki and T. Koskela, “De-
vice-to-Device (D2D) Communication in Cellular Net-
work-Performance Analysis of Optimum and Practical
Communication Mode Selection,” Wireless Communica-
tions and Networking Conference (WCNC), 18-21 April
2010.
[3] G. Fodor and N. Reider, “A Distributed Power Control
Scheme for Cellular Network Assisted D2D Communica-
tions,” Global Telecommunications Conference
(GLOBECOM 2011), 5-9 Dec. 2011.
[4] B. Wang, L. Chen, X. H. Chen, X. Zhang and D. C. Yang,
“Resource Allocation Optimization
(VTC Spring
[5] X. Xiao, X. M. Tao and J. H. Lu, “A QoS-Aware Power
Optimization Scheme in OFDMA Systems with Inte-
grated Device-to-Device (D2D) Communications,” Ve-
hicular Technology Conference (VTC Fall), 5-8 Sept.
2011.
[6] C.-H. Yu and O. Tirkkonen, “Device-to-Device Underlay
Cellular Network Based on Rate Splitting,” Wireless
Communications and Networking Conference (WCNC),
1-4 April 2012.
[7] P. Jani
Resource Allocation for Device-to-device Radio under-
laying Cellular Networks,” Vehicular Technology Con-
ference, 26-29 April 2009.

J. LIU ET AL.
193
e-constrained Re-
wer Allocation Scheme for Power-Efficient De-
nderlaying LTE Network,” Wireless and Mobile
elecommunications (ICT),
[8] H. Wang and X. Chu, “Distanc
source-sharing Criteria for Device-to-device Communica-
tions Underlaying Cellular Networks,” Electronics Let-
ters , Vol. 48, No. 9, 2012, pp. 528-530.
[9] M. Jung, K. Hwang and S. Choi, “Joint Mode Selection
and Po
vice-to-Device (D2D) Communication,” Vehicular Tech-
nology Conference (VTC Spring), 6-9 May, 2012.
[10] M. Zulhasnine, C. C. Huang and A. Srinivasan, “Efficient
Resource Allocation for Device-to-Device Communica-
tion U
Computing, Networking and Communications (WiMob),
11-13 Oct. 2010, pp. 368-375.
[11] Elkotby, H. E. Elsayed, Ismail and M.H., “Exploiting
Interference Alignment for Sum Rate Enhancement in
D2D-Enabled Cellular Networks,” Wireless Communica-
tions and Networking Conference (WCNC), 1-4 April
2012, pp. 1624 -1629.
[12] X. Y. Zhu, S. Wen, etc., “A Cross-layer Study: Informa-
tion Correlation Based Scheduling Scheme for De-
vice-to-Device Radio Underlaying Cellular Networks,”
International Conference on T
23-25 April 2012.
[13] “Evolved Universal Terrestrial Radio Access (E-UTRA);
Physical Layer Procedures (Release 9),” 3GPP TS 36.213
V9.2.0, 2010.