An Interference Mitigation Scheme of Device-to-Dev
An Interference Mitigation Scheme of Device-to-Dev
An Interference Mitigation Scheme of Device-to-Dev
Article
An Interference Mitigation Scheme of
Device-to-Device Communications for Sensor
Networks Underlying LTE-A
Jeehyeong Kim, Nzabanita Abdoul Karim and Sunghyun Cho *
Department of Computer Science and Engineering, Hanyang University, 55 Hanyangdaehak-ro, Sangnok-gu,
Ansan, Gyeonggi-do 426-791, Korea; manje111@hanyang.ac.kr (J.K.); anzabanita@gmail.com (N.A.K.)
* Correspondence: chopro@hanyang.ac.kr; Tel.: +82-31-400-5670
Abstract: Device-to-Device (D2D) communication technology has become a key factor in wireless
sensor networks to form autonomous communication links among sensor nodes. Many research
results for D2D have been presented to resolve different technical issues of D2D. Nevertheless,
the previous works have not resolved the shortage of data rate and limited coverage of wireless
sensor networks. Due to bandwidth shortages and limited communication coverage, 3rd Generation
Partnership Project (3GPP) has introduced a new Device-to-Device (D2D) communication technique
underlying cellular networks, which can improve spectral efficiencies by enabling the direct
communication of devices in proximity without passing through enhanced-NodeB (eNB). However,
to enable D2D communication in a cellular network presents a challenge with regard to radio
resource management since D2D links reuse the uplink radio resources of cellular users and it can
cause interference to the receiving channels of D2D user equipment (DUE). In this paper, a hybrid
mechanism is proposed that uses Fractional Frequency Reuse (FFR) and Almost Blank Sub-frame
(ABS) schemes to handle inter-cell interference caused by cellular user equipments (CUEs) to D2D
receivers (DUE-Rxs), reusing the same resources at the cell edge area. In our case, DUE-Rxs are
considered as victim nodes and CUEs as aggressor nodes, since our primary target is to minimize
inter-cell interference in order to increase the signal to interference and noise ratio (SINR) of the target
DUE-Rx at the cell edge area. The numerical results show that the interference level of the target
D2D receiver (DUE-Rx) decreases significantly compared to the conventional FFR at the cell edge.
In addition, the system throughput of the proposed scheme can be increased up to 60% compared to
the conventional FFR.
Keywords: wireless sensor network (WSN); device-to-device (D2D); fractional frequency reuse (FFR);
almost blank sub-frame (ABS); long-term evolution (LTE); signal to interference plus noise ratio (SINR)
1. Introduction
Wireless sensor networks (WSNs) have been used in various fields including environment
monitoring, home automation, healthcare, agriculture, military, smart grids, and smart cars [1,2].
In these applications, sensors are equipped with wireless radio interfaces in order to form a wireless
network and communicate with other sensors or a data aggregator. WSN will also play an important
role to open the early Internet of Things (IoT) market [1,3].
Some technical barriers, which should be overcome to use WSN as a network infrastructure
of the early IoT market, still remain open issues. A representative open issue is the autonomous
communications among sensors. It is inefficient for a central node to control all sensors in WSN.
Thus device-to-device (D2D) communication is considered as a rising technology in WSNs to solve
this problem [4,5]. In D2D communications, sensor nodes or devices are able to communicate with
each other through autonomous manner. D2D communication enables sensor nodes in close proximity
to establish a direct link with each other as opposed to being routed by a controller or central node.
Recently, D2D communication technologies have been actively studied in oneM2M standard for
IoT [6,7] and 3rd Generation Partnership Project (3GPP) standard for Long Term Evolution Advanced
(LTE-A) systems [8,9]. Another open issue of WSN to boost IoT is the network coverage extension and
interworking with heterogeneous networks. The cellular networks can be an excellent candidate to
overcome this problem. The cellular networks usually cover the majority of countries and interoperate
with other networks. Thus, there is a growing trend to interwork WSNs with cellular networks.
For these backgrounds, this paper considers D2D communication technologies for wireless sensor
networks. In particular, D2D communication technologies underlying LTE-A systems are investigated
to apply wireless sensor networks to IoT [8,10]. D2D communication underlying LTE-A systems
can be defined as a technology enabling direct communication between proximal sensor nodes or
user equipment (UE). In D2D communication, sensor nodes do not need to pass through cellular
infrastructural nodes such as an enhanced node-B (eNB) or mobility management entity (MME) to
setup a D2D communication link. In the beginning of D2D study, integrating D2D in an LTE-A
network was strongly supported by Qualcomm, which previously developed FlashLinQ. FlashLinQ
is a proprietary technology which allows cellular devices to discover each other automatically and
communicate with thousands of other FlashLinQ enabled devices within 1 km range [4,11]. To avoid the
loss of network throughput in D2D communication underlying LTE-A systems, the licensed spectrum
of cellular networks should be shared with D2D communications. Unfortunately, the interference
among cellular user equipment (CUE) and D2D user equipment (DUE) is inevitable. Because of this, a
considerable amount of research on the interference management has been conducted with regard to
D2D communications being laid into LTE-A cellular networks where CUEs and DUEs share the same
resources. In [12], the authors proposed a radio resource allocation scheme for D2D communication
underlying cellular networks using strict fractional frequency reuse (FFR) to reduce the inter-cell
interference of a DUE on receiving mode (DUE-Rx) at the cell edge area. In [13], the authors proposed
the modified FFR to improve the coverage of CUEs and DUEs in cell edge areas. In this scheme,
different resources were allocated to DUEs based on their location. Wenbin et al. proposed a resource
allocation scheme in which dedicated frequency resource blocks (FRBs) are assigned to D2D links
to avoid interference [14]. In [15], an intelligent power control mechanism has been proposed. It is
based on the soft FFR that allocates radio resources to CUEs and DUEs with variable ratios. The
drawback of prior systems, which are based on strict FFR, is the inefficient use of radio resources and
the decrease of throughput. In the strict FFR, the entire frequency band can not be fully used in an
entire cell. Some portion of resources are used only for CUEs or cell center users and the remaining
portion is allocated to DUEs or cell edge users. Thus, the waste of resources and the loss of throughput
are inevitable in the strict FFR. On the other hand, the entire frequency can be used for the entire users
in the soft FFR, but the interference occurs from CUEs to DUE-Rx. The interference can prevent a
DUE-Rx from receiving data from a CUE on transmitting mode (CUE-Tx) in cell edge areas.
In this paper, a hybrid mechanism of fractional frequency reuse and almost blank sub-frame
(FFR-ABS) schemes are proposed to mitigate inter-cell interference caused by CUEs of neighboring
cells to D2D pairs in cell edge areas. The proposed scheme can guarantee the Quality-of-Service
(QoS) of both CUE and D2D pairs in terms of throughput and signal to interference and noise ratio
(SINR), and increase the throughput of UEs in the cell edge area. The rest of this paper is organized as
follows: Section 2 describes proximity services and resource allocation for D2D in LTE-A standards.
The system model and proposed scheme are discussed in Section 3. Mathematical modeling, analysis,
and discussion of the performance evaluations are provided in Sections 4 and 5, respectively. Finally,
concluding remarks are summarized in Section 6.
Sensors 2017, 17, 1088 3 of 18
accordingly, the overall cell throughput increases in soft FFR. The drawback of soft FFR is that DUEs
or CUEs can be suffered from inter-cell interference. Therefore, it is necessary to study an efficient
resource allocation scheme for D2D communications that can efficiently mitigate interference without
throughput loss.
Figure 3. System model and resource allocation for soft fractional frequency reuse (FFR).
a D2D connection request from DUE-Tx i to DUE-Rx j, it lists CUEs whose distance to the DUE-Rx
ABS . The list is a set ABS CUE of CUEs that should be silent on ABS sub-frames for
j is less than Dmax j
the DUE-pair i, j. The eNB then sends control signals including the ABS pattern for the DUE-pair i, j.
The final phase is data transmission by DUEs and CUEs. In this phase, the eNB checks the DUE-pair i, j
and the set ABS CUE
j for the DUE-pair i, j. The CUEs in the set should be silent during ABS sub-frames.
In the proposed scheme, the CUEs that cause severe interference are selectively chosen to be silent in
ABS to minimize the throughput loss.
8: send control signal to DUE-Tx i including ABS pattern for the DUE-pair i, j
9: end while
Figure 4 depicts different interference sources and the sub-frame allocation for DUE-pair and
CUE in the proposed FFR-ABS scheme. In the proposed FFR-ABS, ABS is applied only to CUEs whose
distance is too close to D2D links. As shown in Figure 4, there are two different interference sources
ABS , the node is called an
such as outer and inner CUEs. If the distance to the D2D link is longer than Dmax
outer CUE. Since the signal from an outer CUE does not interfere with the D2D link, the ABS method
is not applied to an outer CUE. It is defined as a non-ABS (NABS) case in the proposed scheme. In an
NABS case, even outer CUEs transmit data on the non-orthogonal frequency with DUEs, and the
interference to DUEs can be tolerable. On the other hand, a CUE that is closer to the D2D link than
ABS is called an inner CUE. In the proposed scheme, ABS is applied to an inner CUE because it
Dmax
can severely interfere with the D2D link. An eNB broadcasts ABS patterns to the entire cell. In the
proposed FFR-ABS scheme, DUE-Tx transmits data on the ABS subframes while inner CUEs should
be silent on the ABS subframes to avoid interference with D2D links as shown in Figure 4.
Sensors 2017, 17, 1088 7 of 18
Table 2. Resource allocation for cellular user equipment (CUE) and device-to-device (D2D) pairs.
We propose a hybrid FFR-ABS scheme to mitigate the inter-sector interferences as well as inter-cell
interferences. CUEs in cell-edge areas should have much higher transmission power to communicate
with eNB located in the cell center. Thus, CUEs also cause interferences to DUE-Rxs in neighboring
sectors. The proposed FFR-ABS scheme can prevent this problem and improve overall throughput.
Table 3. Notations.
Notation Definition
Φb Homogeneous Poisson point processes (PPP) for eNB distribution
Φc Homogeneous Poisson point processes (PPP) for CUE distribution
Φd Homogeneous Poisson point processes (PPP) for DUE-Tx distribution
λb Intensity of homogeneous Poisson point processes (PPP) for eNB distribution Φb
λc Intensity of homogeneous Poisson point processes (PPP) for CUE distribution Φc
λd Intensity of homogeneous Poisson point processes (PPP) for DUE-Tx distribution Φd
CUE ,H D2D
Hx,y Complex channel gain between node x to y for cellular and D2D links, respectively
x,y
n ,K n
Ktx A set {1, ..., K } of K DUE-Tx/DUE-Rx terminals using D2D links for frequency Fn
rx
Cn A set {1, ..., C } of C UE using cellular links for frequency Fn
Dmin , Dmax The minimum/maximum distance between DUE-Tx and DUE-Rx, respectively
ABS
Dmax Maximum distance between CUE and DUE-Rx for adopting ABS scheme
ABS CUE j A set of CUEs whose distance from DUE-Rx j is less than Dmax ABS
ABS cDUE A set of DUEs whose distance from CUE j is less than Dmax ABS
To derive the distribution of network elements, we assume that eNBs are distributed with
homogeneous PPP Φb of intensity λb . CUEs are located by independent stationary point process Φc
with density λc . It is assumed that UE-Txs are distributed in a homogeneous PPP Φd with density λd .
DUE-Rxs, the receiver of D2D communication, are uniformly distributed following DUE-Tx during
the interval [Dmin , Dmax ]. ABS CUE
j
ABS .
is a set of CUEs whose distance to DUE-Rx j is less than Dmax
The CUEs in ABS CUE
j should be silent mode on ABS subframes. Similarly, ABS cDUE is also defined as
ABS . The reference
a set of DUE-Tx i in case the distance between CUE c and the DUE-Rx j is less than Dmax
distance is the distance between CUE c and DUE-Rx j. Therefore, the set of DUE-Tx i that interferes
with the CUE c is determined by the distance between DUE-Rx j and CUE c.
Sensors 2017, 17, 1088 9 of 18
CUE = | H CUE |2 . Similarly, the received signal power of DUE-Rx from DUE-Tx i to DUE-Rx j,
where Gc,j c,j
D2D , can be defined as follows:
Pi,j
D2D D2D D2D
Pi,j = Gi,j p , (2)
D2D = | H D2D |2 .
where Gi,j i,j
Pi,jD2D
SI NR D2Dn
FFR = D2D + CUE
. (3)
γ2 + ∑k∈Ktx
n ,k 6 =i P
k,j ∑k∈C n Pk,j
i,j∈ Fn
Similarly, the SINR of CUE with the conventional FFR can be defined as:
Pi,cCUE
SI NRCUEn
FFR = . (4)
γ2 + ∑k∈Ktx D2D CUE
k,c + ∑k ∈C n
n P P
k,c i ∈ Fn
In the proposed FFR-ABS mode, DUE-Rx does not suffer from CUEs whose distance to DUE-Rx is
ABS because those do not transmit data simultaneously through the ABS method. Thus, the
less than Dmax
SINR of DUE-Rx j from DUE-Tx i in the FFR-ABS can be defined as follows:
Pi,jD2D
SI NR D2Dn
ABS = . (5)
γ2 + ∑k∈Ktx
n ,k 6 =i P
D2D +
k,j ∑k∈C n ,k∈/ ABS CUE PCUE
k,j i,j∈ Fn
j
Note that CUE k does not interfere with DUE-Rx j in ABS CUE
j . In the same way, the SINR of CUE in
the FFR-ABS can be described as follows:
Pi,cCUE
SI NRCUEn
ABS = D2D + CUE
. (6)
γ2 + ∑k∈Kn ,k∈/ ABS DUE Pk,c ∑k∈C n Pk,c
i ∈ Fn
tx c
Rlink link
mode = log2 (1 + SI NRmode ), (7)
where a link can be a D2D or cellular link. To be short, a cellular link can be denoted as CUE and
be applied to FFR or FFR-ABS. FFR-ABS is shortened as ABS. From Equation (7), we define the total
Sensors 2017, 17, 1088 10 of 18
network throughput to maximize the data rate of CUE and DUE. Firstly, the total network throughput
of conventional FFR can be expressed as the following:
FFR
Rtotal = ∑ RD2D
FFR + ∑ R FFR .
CUE
In FFR-ABS, a part of CUEs have ABS applied and others operate by the FFR method. The nodes
that are on the ABS scheme have to transmit data on their subframes only. Thus, the total data rate of
FFR-ABS is as follows:
ABS CUE )
S
n(
S
ABS D2D ) n( j∈K
where Pr D2D
ABS =
c∈C
n(K)
c
, and PrCUE
ABS = n(C)
j
. These are the ratios of DUE and CUE that
have the ABS scheme applied. β is an ABS ratio. It means how many subframes are used as ABS
in a frame. We can calculate the instantaneous data rate with Shannon capacity equation, but the
instantaneous data rate can not consider time resource distribution for a node. In order to apply the
distribution of time resources to the data rate, we use an ABS ratio, β. During the ABS period, an eNB
exchanges an ABS ratio via X2 interface. There are two prime interfaces in LTE such as X2 and S1
interfaces. The X2 interface is used to communicate between eNBs. The eNBs share the information for
UEs, hand-over, channel status, and the configuration of eNBs [25].
5. Performance Evaluations
where d is the distance between transmitter and receiver in meters. Channel models usually consider
an urban model with non-line-of-sight (NLOS). However, we additionally consider NLOS 80% + line
of sight (LOS) 20% channel models to analyze the performance of the proposed scheme.
Sensors 2017, 17, 1088 11 of 18
Parameter Assumption/Value
Cellular layout Hexagonal grid, 7 cells sites
Path Loss Model (D2D-NLOS 100%) 24.82 + 35.31×log10 d
Path Loss Model (D2D-NLOS 80%) 28 + 40×log10 d
Path Loss Model (CUE-NLOS 100%) 31.25 + 33.76×log10 d
Path Loss Model (CUE-NLOS 80%) 30.35 + 36.7×log10 d
CUE transmit power 24 dBm
DUE transmit power 20 dBm
Noise power density −174 dBm
Inter-site distance (radius of the cell) 500 m
Carrier frequency 2 GHz
Bandwidth 10 MHz
Number CUEs 10
Number of DUE 20
ABS pattern period 10 ms
Distance between D2D 20–60 m
Traffic patterns Full-buffered
Monte Carlo number 10,000
Figures 6 and 7 depict these scenarios. In Figure 6, there are two significant interferers to ABS
mode coverage like DUE-Tx and CUE. In this case, the interference from CUE can be suppressed
through the FFR-ABS scheme, but interference from DUE-Tx cannot be avoided. We simulate this
model with various transmission powers and channel models. Figure 7 describes a model in which all
of the DUEs and CUEs are randomly distributed. In this model, we compare throughput for each node
with different traffic loads.
Figure 6. Scenarios 1 and 2: Isolated D2D user equipment (DUE)-pairs and cellular user equipment
(CUE) model.
Sensors 2017, 17, 1088 12 of 18
Figure 8. Combined graphs of signal to interference and noise ratio (SINR)-FFR and SINR-ABS received
at the target DUE-Rx.
Figure 9. Combined graphs of FFR and ABS throughput received at the target DUE-Rx.
5.2.2. Scenario 2: Different Transmission Power Levels for CUE and DUE
Figures 10 and 11 show the SINR and throughput at DUE-Rx when aggregate interference from
all interfering nodes are present in the network. In this scenario, all DUEs use the same transmission
power levels (100 mW or 20 dBm), even those that are located at different distances; the CUEs also
exhibited the same transmission power levels (250 mW or 24 dBm). Figures 10 and 11 show that the
received SINR and throughput of the target DUE-Rx are higher in the proposed FFR-ABS compared to
those of the conventional FFR at the cell edge area.
Sensors 2017, 17, 1088 14 of 18
Figure 10. SINR received at the target DUE-Rx with constant CUE-Tx and variable DUE-Tx
power levels.
Figure 11. Channel capacity received at the target DUE-Rx with constant CUE-Tx and variable DUE-Tx
power levels.
Figure 12 shows the throughput received at the target DUE-Rx with respect to the variable DUE
transmission power. It is obvious from Figure 12 that the proposed scheme achieves higher throughput
for the target DUE-Rx in the cell edge area compared to the conventional scheme. As shown in
Figure 12, CUEs experience much higher interference levels compared to DUE pairs. Even at the
same transmission power levels, CUEs outperform the DUE pairs. Thus, the use of the proposed
hybrid FFR-ABS can reduce aggregate interference that could reach the target DUE-Rx. Once aggregate
interference to the target DUE-Rx is significantly reduced, SINR and throughput can be significantly
increased for any DUE transmission power level, as shown in Figure 12.
Sensors 2017, 17, 1088 15 of 18
Figure 12. Interference levels from both interferers for variable Tx power.
We also assume equal transmission power levels for the target transmitter (DUE-Tx) and all
interferers (DUEs and CUEs). As shown in Figure 12, when the same transmission power levels are
applied, aggregate interference increases due to a high level of interference from the CUEs. Therefore,
for the same transmission power levels, the interference of target DUE-Rx would increase depending
on the location of the interferers. The closer the interferer, the higher the aggregate interference,
resulting in a reduction to the SINR and throughput. By applying the ABS, interference from the CUEs
can be suppressed during ABS time. Due to non-CUE transmission, aggregate interference is reduced,
but SINR and throughput are increased as shown in Figures 13 and 14.
Figure 13. SINR received at the target DUE-Rx under the same CUE-Tx and DUE-Tx power levels.
Sensors 2017, 17, 1088 16 of 18
Figure 14. Channel capacity received at the target DUE-Rx under the same CUE-Tx and DUE-Tx
power levels.
Figure 15. Cumulative distribution function (CDF) of throughput based on traffic loads.
Sensors 2017, 17, 1088 17 of 18
6. Conclusions
In this work, we proposed a hybrid mechanism with regard to FFR and ABS for D2D
communication underlying the LTE-A network. The hybrid mechanism provided a solution toward
mitigating serious interference caused by CUEs of neighboring cells to D2D receivers in the cell edge
area. The use of ABS in the conventional FFR provided an effective method with regard to decreasing
inter-cell interference levels to victims (DUE-Rx) near the cell edge area. Based on our mathematical
analysis, we determined that the ABS ratio that could provide balance between CUEs and DUEs in
terms of Quality-of-Service. Performance evaluation results of the proposed scheme showed that the
introduction of the FFR-ABS scheme could significantly improve system throughput and guarantee
Quality-of-Service to both CUE and D2D pairs in the cell edge area. Further studies should be extended
from the scheme to minimize intra-cell interference caused by co-channel D2D pairs.
Acknowledgments: This work was supported by the Korea Institute of Energy Technology Evaluation
and Planning (KETEP) and the Ministry of Trade, Industry and Energy (MOTIE) of the Republic of Korea
(No. 20161520302230), and supported by the research fund of Hanyang University (HY-2015-N).
Author Contributions: J. Kim contributed to design the proposed scheme and perform computer simulation and
major revision of this paper. N. A. Karim contributed to design the proposed scheme and write the 1st draft of
this paper. S. Cho is a correspondence of this paper and the proposed research results.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Ericsson. More Than 50 Billion Connected Devices; Ericson White Paper; Ericsson: Stockholm, Sweden, 2011.
2. Cho, S.; Choi, J.W.; You, C. Adaptive multi-node multiple input and multiple output (MIMO) transmission
for mobile wireless multimedia sensor networks. Sensors 2013, 13, 13382–3401.
3. Evans, D. The Internet of Things How the Next Evolution of the Internet is Changing Everything; Cisco White
Paper; Cisco Systems: San Jose, CA, USA, 2011.
4. Mumtaz, S.; Rodriguez, J. Smart Device to Smart Device Communication; Springer: Cham, Switzerland, 2014.
5. LG Electronics. Discussion on D2D Discovery Physical Layer Design; LG Electronics: Seoul, Korea, 2014.
6. Swetina, J.; Lu, G.; Jacobs, P.; Ennesser, F.; Song, J. Toward a standardized common M2M service layer
platform: Introduction to oneM2M. IEEE Wirel. Commun. 2014, 21, 20–26.
7. OneM2M Alliance. Available online: http://www.onem2m.org/ (accessed on 10 May 2017).
8. Feng, J. D2D Communication in LTE-Adavanced Network. Ph.D. Dissertation, Universite de Bretagne-Sud,
Lorient, France, 2013.
9. The Mobile Broadband Standard. Available online: http://www.3gpp.org/DynaReport/
FeatureOrStudyItemFile-580038.htm (accessed on 10 May 2017).
10. Lin, M.; Ouyang, J.; Zhu, W.P. Joint Beamforming and Power Control for Device-to-Device Communications
Underlaying Cellular Networks. IEEE J. Sel. Areas Commun. 2016, 34, 138–150.
11. Ma, C.; Liu, J.; Tian, X.; Yu, H.; Cui, Y.; Wang, X. Interference Exploitation in D2D-Enabled Cellular
Networks-A Secrecy Perspective. IEEE Trans. Commun. 2015, 63, 229–242.
12. Chae, H.S.; Gu, J.; Choi, B.G.; Chung M.Y. Radio Resource Allocation Scheme for Device-to-Device
Communication in Cellular Networks Using Fractional Frequency Reuse. In Proceedings of the 17th
Asia-Pacific Conference on Communications (APCC), Sabah, Malaysia, 2–5 October 2011; pp. 58–62.
13. Zhang, Z.; Hu, R.Q.; Qian, Y.; Papathanassiou, A.; Wu, G. D2D communication underlay uplink cellular
network with fractional frequency reuse. In Proceedings of the 11th International Conference on the Design
of Reliable Communication Networks (DRCN), Kansas City, MO, USA, 24–27 March 2015; pp. 247–250.
14. Wu, W.; Zhang, Y. Dedicated resource allocation for D2D communications in cellular systems employing
FFR. In Proceedings of the 6th International Conference on Wireless Communications and Signal Processing
(WCSP), Hefei, China, 23–25 October 2014; pp. 1–6.
15. Shah, S.T.; Gu, J.; Hasan, S.F.; Chung, M.Y. SC-FDMA-based resource allocation and power control scheme
for D2D communication using LTE-A uplink resource. EURASIP J. Wirel. Commun. Netw. 2015, 1, 1–15.
16. 3GPP TR 23.703. Available online: http://www.arib.or.jp/english/html/overview/doc/STD-T63v11_00/5_
Appendix/Rel12/23/23703-c00.pdf (accessed on 10 May 2017).
Sensors 2017, 17, 1088 18 of 18
c 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).