Communications and Network, 2013, 5, 338-343
http://dx.doi.org/10.4236/cn.2013.53B2062 Published Online September 2013 (http://www.scirp.org/journal/cn)
Resource Allocation Method of Device-to-Device
Communication
Wenji Feng, Yafeng Wang, Lei Yang
Wireless Theory & Technology lab (WT&T), Beijing University of Posts and Telecommunications, Beijing, China
Email: fwj10171017@sina.com, wangyf@bupt.edu.cn, yanglei5658@gmail.com
Received July, 2013
ABSTRACT
In this paper, we study D2D (Device-to-Device) communication underlying LTE-Advanced uplink system. Since D2D
communication reuses uplink resources with cellular communication in this scenario, it’s hard to avoid the inference
between D2D users and cellular users. If there is no restriction for D2D communication on using the whole uplink fre-
quency band, it will have a strong negative impact on cellular communication. In order to overcome this shortage, we
propose a resource allocation method that D2D users and cellular users use orthogonal frequency resources. This
method will effectively reduce the inference between both kinds of communication. However, an obvious disadvantage
of this method is no effective use of uplink resources. Based on this, we propose an optimized resource allocation
method that a specific cellular user will be chosen to reuse the RBs (Resource Block) of D2D users. These ideas will be
taken into system-level simulation, and from the results of simulation we can see that the optimized method has the
ability to improve overall syste m performance and limit inference for cell-edge users.
Keywords: Device-to-device Communication; LTE-Advanced System; Resource Allocation Method; up Link;
Cell-edge Users
1. Introduction
3GPP Long Term Evolution (LTE) technology has been
proved to have outstanding performance, especially in
the measures of spectral efficiency and the average
throughput, cell-edge and peak values in a cellular, fre-
quency reuse one network [1,2]. Hence, major efforts
have been spent on the development of LTE. Currently
the further evolution of such systems has been started
under the scope of LTE-Advanced [3-5].
The device-to-device communication (D2D) technol-
ogy, also known as proximity-based services (ProSe), is
introduced into LTE-Advanced system [6-7]. However,
D2D communication is sharing authorized frequency
band with cellular communication by the way of or-
thogonal method or multiplexing method. The introduc-
tion of D2D communication is to improve the throughput
of overall cellular system. It contributes to higher fre-
quency resource utilization as well. When D2D users
reuse cellular frequency resources, it is hard to avoid the
interference of other cellular users in cell. It may affect
the D2D user's communication quality to some extent. If
there are no restrictions for D2D communication on us-
ing the whole frequency band, it will have a strong n ega-
tive impact on cellular communication.
As mentioned above, appropriative resource allocation
methods of D2D communication may improve this situa-
tion. Because of frequency reusing with cellular commu-
nication, D2D communication should choose a better
resource allocation method to avoid inference against
each other. Hence, we hope to find out simulation sce-
narios of different resource allocation methods so that we
can choose an appropriative allocation method for D2D
communication.
This paper is organized as follows: Section 2 describes
the simulation platform. Section 3 shows different kinds
of D2D communication resource allocation methods, and
the performance will be analyzed in this section. Ac-
cording to section 3, we propose some optimization
method of D2D resource allocation method in section 4.
Finally concluding remarks are made in Section 5.
2. Simulation Platform
Assume that the D2D communication underlying LTE-
Advanced network only reuse uplink frequency resource.
The simulation platform consists of 7 eNB (21 sectors).
We distribute one pair of D2D UEs into one cell amount
to 21 pairs of D2D UEs. In our simulation platform, we
consider that there is one transmitter and one receiver in
a pair of D2D UEs, which are working only during up-
link slot. The distribution of eNB (evolved node B) and
D2D UEs is shown in Figure 1.
C
opyright © 2013 SciRes. CN
W. J. FENG ET AL. 339
-1000 -500 0500 1000
-800
-600
-400
-200
0
200
400
600
800
X(m)
Y(m)
D2DTx
D2DRx
BS
Figure 1. Distribution of eNB and D2D.
The radius of each cell is 500 m. The distance be-
tween transmitter and receiver is 10 m to 20 m. D2D
communication reuses uplink resources of LTE-Ad-
vanced, which contains 46 RB. The transmission power
[8,9] of D2D transmitter is 20 dBm. The system band-
width is 10 MHz and the carrier frequency is 2 GHz. We
distribute 210 cellular UEs into overall system so that
there are 8 to 12 cellular UEs in each cell. The per-ma-
nent MCS of D2D communication is 28. The path loss
model and corresponding shadow fading model of D2D
communication are referred as model of Urban Macro
(UMa) in [10], where it can be further categorized as
Line-of-sigh t (LOS) and Non-line-of-sight (NLOS).
LOS:
10 10
16.9log ([])46.820log ([]/5.0)
c
PLd mfGHz
(1)
NLOS:
10 10
40log ([])30log ([])49
c
PLd kmfMHz  (2)
The probability of LOS is as follow:
14
=exp((4) /3),460
0, 60
LOS
d
Pdd
d
 
(3)
Where PL denotes the path loss, d is the distance be-
tween D2D users, fc is carrier frequency, and PLOS is the
probability of LOS.
The parameters of simulation are listed in Table 1.
3. Resource Allocation Method
3.1. Resource Allocation Method of D2D
In order to analysis the performance of different methods
of D2D resource allocation, first of all we assume Case
Basic, which is a LTE-Advanced uplink network without
D2D communication. In this simulation scenario, we set
D2D transmission power to fixed 20 dBm, and the MCS
will be the highest one. On this basis, we propose another
3 cases of D2D resource allocation method, which are
described below:
a) Case Basic:
This case is LTE-Advanced uplink system without
D2D communication.
b) Case All_RB_Reuse:
As Figure 2 shows, D2D communication reuses all
uplink frequency resource with cellular system.
c) Case 10_RB_Reuse:
As Figure 3 shows, D2D communication reuses 10
RB of cellular uplink frequency resource. Cellular UEs
still use all uplink frequency resource.
d) Case 10_RB_Sep:
As Figure 4 shows, D2D communication uses 10 RB
of uplink frequency resources. Cellular UEs only use
another 36 RB of uplink frequency resource. It means
that the RB of D2D and cellular is orthogonal.
Table 1. List of simulation parameters.
Parameters Value
Number of eNB 7 eNB (21 sectors)
Radius of cell 500 m
Distance of D2D
communication 10 m to 20m
System resources Uplink
Transmission power of D2D 20 dBm
System bandwidth 10 MHz
Carrier freque nc y 2 GHz
Number of cellular U E 210
Number of D2D pairs 21
Distribution 8 to 12 cellular UEs in each cell;
one pair of D2D-U E into one cell
MCS of D2D 28
Thermal noise density 174 dBm/Hz
Scheduling algorithm Proportional Fai r
RB 1 to 46
Cellular and D2D
Figure 2. Caes All_RB_Reuse.
RB 1 to 36 RB 37 to 46
Cellular
D2D
Figure 3. Case 10_RB_Reuse.
RB 1 to 36 RB 37 to 46
Cellular
D2D
Figure 4. Case 10_RB_Sep.
Copyright © 2013 SciRes. CN
W. J. FENG ET AL.
Copyright © 2013 SciRes. CN
340
3.2. Performance Analysis
We compare the simulation results of previous 4 cases
including average throughput of eNB, throughput of
cell-edge users, throughput of D2D communication,
throughput of overall system and BLER (Block Error
Rate) of D2D communication. Figure 5 illustrates the
simulation results of 4 cases.
From Figure 5, we recognize that if D2D communica-
tion reuses all resource with cellular communication, the
throughput of overall system gets a very large increase.
However, on the other hand, D2D communication brings
a lot of interference against cellular communication so
that cell-edge users almost unable to transmit data cor-
rectly. At the same time, because of interference from
D2D communication, the throughput of cell centre users
drop sharply. Therefore, if there are no restrictions for
D2D communication on using the whole frequency band,
it will have a strong negative impact on cellular commu-
nication [11,12], which is not appropriate in the practical
communication system.
In the view of this, D2D communication only use part
of uplink frequency resource. In Case 10_RB_Reuse,
D2D communication only reuses 10 RB with cellular
system, and the all 46 RB is still used by cellular. In this
simulation case, comparing with Case All_RB_Reuse,
cellular system throughput gets a large increase, but the
throughput of cell-edge UEs is still very low so that cell-
edge users cannot communicate normally.
At the same time, we consider Case 10_RB_Sep,
which D2D commun icatio n uses 10 RB of cellular up link
frequency resource, and cellular UEs only use the other
36 RB of uplink frequency resource. From the results of
simulation, the performance of Case 10_RB_Sep is the
most equilibrium. What’s more, comparing with Case
Basic the throughpu t gets a lot of gain, and the BLER of
D2D is the lowest.
AverageThroughputofeNB
5887.64
1281.07
4680.12 4607.12
0
1000
2000
3000
4000
5000
6000
7000
BasicAll_RB_Reuse10_RB_Reuse10_RB_Sep
Throughput(kbps)
ThroughputofCelledgeUEs
259.01
0.09 4.42
206.67
0
50
100
150
200
250
300
BasicAll_RB_Reuse 10_RB_Reuse10_RB_Sep
Throughput(kbps)
ThroughputofD2D
8863.94
2350.36 2460.53
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
BasicAll_RB_Reuse10_RB_Reuse10_RB_Sep
Throughput(kbps)
Thro ughputofoveral lsystem
5887.64
10145.01
7030.48 7067.65
0
2000
4000
6000
8000
10000
12000
BasicAll_RB_Reuse 10_RB_Reuse10_RB_Sep
Throughput(kbps)
BLE RofD2DCommunication
34.92%
21.54% 17.87%
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
BasicAll_RB_Reuse 10_RB_Reuse10_RB_Sep
BLER
Figure 5. Results of four simulation cases.
W. J. FENG ET AL. 341
Figure 6 shows the cellular users SINR of 4 cases.
From Figure 6, we can still have the conclusions as
above. In Case All_RB_Reuse, the SINR of cellular UEs
is sharply decreased. It causes the exacerbation of overall
system. In Case 10_RB_Reuse, the effects of SINR are
more embodied at cell-edge users. The throughput of
cell-edge UEs is still very low so that cell-edge users
cannot communicate normally in this case. In Case
10_RB_Sep, the performance of SINR is almost the same
as Case Basic.
4. Optimization Scheme
4.1. Optimization Simulation Scenarios
We describe Case 10_RB_Sep in section 3, which has
the best performance of all 4 cases. However, in Case
10_RB_Sep, cellular communication gives up 10 RB to
assure the excellent performance of both cell-edge users
and D2D communication. The 10 RB, which are given
up by cellular communication and only used by D2D
communication, still negatively affect the performance of
overall system. Hence, we hope to find a possible way to
overcome the shortage. On the basis of Case 10_RB_Sep,
we consider choosing one cellular UE to reuse 10 RB
with D2D UEs. The simulation scenarios are described as
follow:
a) Case 10_RB_Sep:
This case is described in section 3.
b) Case Random_UE:
We choose a random cellular UE from current cell to
reuse 10 RB with D2D communication. The rest of cel-
lular users still use the other 36 RB.
c) Case Worst_UE:
We choose a cellular UE from current cell, which has
the slowest transmission rate currently, to reuse 10 RB
with D2D communication. The rest of cellular users still
use the other 36 RB.
d) Case Best_UE:
We choose a cellular UE from current cell, which has
the highest transmission rate currently, to reuse 10 RB
with D2D communication. The rest of cellular users still
use the other 36 RB.
4.2. Performance Analysis
We propose 3 cases to compare with Case 10_RB_Sep,
the results of performance are shown in Figure 7.
-40 -30 -20 -10010 20 30 40
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
SINR(dB)
C D F
Basic
AllRBResue
10RBResue
10RBSep
Figure 6. Cellular UEs SINR.
AverageThroughputofeNB
4607.12
4528.54
4795.63
4571.52
4350
4400
4450
4500
4550
4600
4650
4700
4750
4800
4850
10_RB_Sep Random_UEWorst_UEBest_UE
Throughput(kbps)
ThroughputofCelledgeUEs
206.67 187.18
3.97
221.21
0
50
100
150
200
250
10_RB_Sep Random_UEWorst_UEBest_UE
Throughput(kbps)
ThroughputofD2D
2576.15
2537.96
2450.88
2532.05
2380
2400
2420
2440
2460
2480
2500
2520
2540
2560
2580
2600
10_RB_Sep Random_UEWorst_UEBest_UE
Throughput(kbps)
Throughputofoverallsystem
7066.51
7246.51
7103.58
7183.27
6950
7000
7050
7100
7150
7200
7250
7300
10_RB_Sep Random_UEWorst_UEBest_UE
Throughput(kbps)
Figure 7. Results of optimized simulation cases.
Copyright © 2013 SciRes. CN
W. J. FENG ET AL.
342
From Figure 7, we can draw the following conclu-
sions:
In Case Random_UE, the throughput of eNB, D2D
and overall system is all lower than Case 10_RB_Sep,
and the throughput of cell-edge users also lower than
Case 10_RB_Sep. So Case Random_UE is not an appro-
priate case to further optimize the performance.
In Case Worst_UE, th e throughput of eNB and over all
system increases a lot, but the throughput of D2D com-
munication fall sharply. And worst yet, the throughput of
cell-edge UEs is the lowest so that cell-edge users canno t
communicate normally. So Case Worst_UE is still not an
appropriate case to further optimize the performance.
In Case Best_UE, comparing with Case 10_RB_Sep,
the average throughput of eNB is reduced by 0.8%, and
the overall system throughput is reduced by 1.1%. At the
same time, the throughput o f cell-edge users is increased
by 7.1%. So in this case, we giv e up a little th roughp ut of
eNB and overall system, for the purpose of cell-edge
users’ throughput improvement. Case Best_UE is an ap-
propriate case to further optimize the performance.
Figure 8 shows the cellular UEs SINR of 4 optimized
cases.
From this figure, we can obviously see that the cellular
UEs of Case Worst_UE has the worst SINR. In Case
Best_UE, due to choosing the highest transmission rate
in current slot, the SINR is not as good as Case 10_
RB_Sep. But the improvement of cell-edge users’
throughput is what we want.
5. Conclusions
Based on the simulation and analysis above, we can rec-
ognize that after introducing D2D communication into
LTE-Advanced uplink system, the performance of over-
all system gains dramatically increase and it also im-
prove the system spectrum efficiency. At the same time,
when D2D communication reuses frequency resource
-40 -30 -20 -10010 2030
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
SINR(dB)
CDF
Last10Sep
OneBestMS
OneRanMS
OneWorstMS
Figure 8. Cellular UEs SINR of optimized cases.
with cellular users, cellular UEs will be interfered
strongly. It leads to the unable transmission of cell-edge
users. Therefore, proposing an appropriate resource allo-
cation method for D2D communication is a key factor to
improve the performance of the overall system. This pa-
per firstly consider different kinds of D2D communica-
tion resource allocation methods and choose the case that
D2D communication uses 10 RB of uplink frequency
resource and the cellular UEs only use another 36 RB of
uplink frequency resource, which has the best perform-
ance. On the basis of Case 10_RB_Sep, we consider
choosing one cellular UE, which has the highest trans-
mission rate in current slot, to reuse 10 RB with D2D
UEs. This optimized simulation case sacrifice a little
throughput of overall system in order to improve the
performance of cell-edge users. After verification and
analysis, Case Best_UE has the ability to improve overall
system performance and limit inference for cellular
communication, especially for cell-edge cellular users.
6. Acknowledgements
This paper is supported by National Key Technology
R&D Program of China under grant No.
2012ZX03003011.
REFERENCES
[1] H. Ekstrom, A. Furusk¨ar, J. Karlsson, M. Meyer, S.
Parkvall, J. Torsner and M. Wahlqvist, “Solutions for the
3G Long-term Evolution,” IEEE Communications Maga-
zine, Vol. 44, No. 3, 2006, pp. 2432-2455.
doi:10.1109/MCOM.2006.1607864
[2] ITU, “ITU-R; Recommendation M. 1645 Framework and
Overall Objectives of the Future Development of
IMT-2000 and Systems Beyond IMT-2000,” 2003.
[3] “Estimated Spectrum Bandwidth Requirements for the
Future Development of IMT-2000 and IMT-Advanced,”
ITU.
[4] D. C. Lee and Y. H. Kwon, “Performance Benefits of
Uplink Packet Relay Protocols for Cellular-Like Systems:
Quantitative Analysis,” IEEE Transactions Wireless
Communications, Vol. 5, No. 7, 2006, pp. 1569-1574.
doi:10.1109/TWC.2006.1673062
[5] S. Haykin, “Cognitive Radio: Brain-empowered Wireless
Communications,” IEEE Journal on Selected Areas in
Communications, Vol. 23, No. 2, 2005, pp. 201-220.
doi:10.1109/JSAC.2004.839380
[6] H.-Y. Hsieh and R. Sivakumar, “On Using Peer-to-peer
Communication in Cellular Wireless Data Networks,”
IEEE Transactions on Mobile Computing, Vol. 3, No. 1,
2004, pp. 57-72. doi:10.1109/TMC.2004.1261817
[7] P. J¨anis, Y. Chia-Hao, K. Doppler, C. Ribeiro, C. Wi-
jting, K. Hugl, O. Tirkkonen and V. Koivunen, “De-
vice-to-device Communication Underlaying Cellular
Communications Systems,” Submitted to: International
Copyright © 2013 SciRes. CN
W. J. FENG ET AL. 343
Journal of Communications, Network and System Sci-
ences (IJCNS).
[8] V. Chandrasekhar and Z. Shen, “Optimal Uplink Power
Control in Twocell Systems with Rise-over-thermal Con-
straints,” IEEE communications letters, Vol. 12, No. 3,
2008, pp. 173-175. doi:10.1109/LCOMM.2008.071697
[9] C.-H. Yu, K. Doppler, C. Ribeiro and O. Tirkkonen,
“Power Optimization of Device-to-device Communica-
tion Underlaying Cellular Communication Systems,” in
Proc. of IEEE International Conference on Communica-
tions, June 2009.
[10] Rep. ITU-R M.2135, “Guidelines for Evaluation of Radio
Interface Technologies for IMT-Advanced,” Revision 1
to Document 5D/TEMP/90-E, 1st July 2008.
[11] C.-H. Yu, K. Doppler, C. B. Ribeiro and O. Tirkkonen,
“Resource Sharing Optimization for Device-to-Device
Communication Underlaying Cellular Networks. Wireless
Communications,” IEEE Transactions on. 8. 2011, pp.
2752-2763.
[12] M. Zulhasnine, C. C. Huang and A. Srinivasan, “Efficient
Resource Allocation for Device-to-Device Communica-
tion Underlaying Lte Network,” in Wireless and Mobile
Computing, Networking and Communications (WiMob),
2010 IEEE 6th International Conference on, 2010), pp.
368-75.
Copyright © 2013 SciRes. CN