DCN 16-13-0165-00-Gcon
Project
IEEE 802.16 Broadband Wireless Access Working Group <http://ieee802.org/16>
Title
Cooperative Transmission Schemes In D2D Group Communications in Cellular Networks
Date
Submitted
2013-11-5
Voice: +8613522359324
E-mail: [email protected]
Source(s)
Lihua Li, Xin Wang.
*
State Key Laboratory of Networking and
Switching Technology, BUPT
Re:
None.
Abstract
This article describes different cooperative transmission schemes in D2D communications when
BS acts as the relay.
Purpose
To investigate the potential of D2D communication underlaying cellular network.
Notice
Copyright
Policy
Patent
Policy
This document does not represent the agreed views of the IEEE 802.16 Working Group or any of its subgroups. It
represents only the views of the participants listed in the “Source(s)” field above. It is offered as a basis for
discussion. It is not binding on the contributor(s), who reserve(s) the right to add, amend or withdraw material
contained herein.
The contributor is familiar with the IEEE-SA Copyright Policy
<http://standards.ieee.org/IPR/copyrightpolicy.html>.
The contributor is familiar with the IEEE-SA Patent Policy and Procedures:
<http://standards.ieee.org/guides/bylaws/sect6-7.html#6> and
<http://standards.ieee.org/guides/opman/sect6.html#6.3>.
Further information is located at <http://standards.ieee.org/board/pat/pat-material.html> and
<http://standards.ieee.org/board/pat>.
DCN 16-13-0165-00-Gcon
Cooperative Transmission Schemes In D2D Group
Communications in Cellular Networks
Lihua Li, Xin Wang
State Key Laboratory of Networking and Switching Technology, BUPT
Beijing 100876, P.R.China
I.
INTRODUCTION
With the developments of wireless communication technologies, explosive growth of location-based services provided by
mobile terminals are witnessed. Device-to-Device Proximity Services (D2D), based on the cellular network architecture, have
gained strong attention currently. D2D communication technology is defined as a direct communication scheme between two or
more relatively close terminals, which means that terminals communicate with each other without data forwarding of the base
station, using dedicated spectrum resources or reusing the licensed spectrum resources. Within a small coverage, D2D will increase
the transmission rate and reduce the transmission power of devices, so as to enhance spectrum utilization and improve the local
users’ experience in the communication system.
D2D group communication is an important aspect, especially in public safety section. In a commercial scenario, it is also a
potential technique for offloading the network. On the other hand, network coding is one of the candidate techniques in the new
generation of wireless communication. Since the significance of broadband data service stands out these days, to achieve reliable
communication over a wireless network of packet erasure channels with high bandwidth efficiency and low overhead is necessary
and urgent. The outstanding performances of network coding in this area make more and more scholars interested in it. Here,
network coding is also introduced in D2D group communication.
In this article, we propose a novel cooperative transmission scheme in D2D groups based on network coding. Uplink resources
are multiplexed for D2D communication to increase the resource utilization in cellular communications. Another advantage is that
interference can be eliminated by the control of the base station.
II.
PROBLEM STATEMENT
Consider a multicast communication system depicted in Figure.2. The transmission method is assumed to be half-duplex and
channel state information (CSI) is only exploited at the receiver side. Each UE is assumed to be equipped with single antenna for
simplicity. Time is divided into uplink and downlink periods. The signals to be transmitted are independently identical distributed,
denoted as x[N] or {x1 , x2 ,..., xN } , which represent a series of modulated symbols of length N. The BS-UE and UE-UE links are
independent fading Rayleigh channels where the channel gain keeps constant during every symbol. In a specific D2D data sharing
scenario, all these UEs are cellular UEs initially. When UEs in close proximity subscribe to the same service from the BS, the BS
would multicast the data via 3G/4G links during the downlink period. Since the diversity of the wireless channels, some UEs
receive and decode it successfully, while others don’t. Duplicate transmission may burden the BS a lot, especially with more UEs
and data requests. Since the subscribers are in proximity enough, when some of them fail to receive and decode the data packets
from the BS, others would forward the specific data packets to them via D2D links during the uplink period. So they transfer from
cellular mode to D2D mode.
DCN 16-13-0165-00-Gcon
Data Cloud
3G/4G link
BS
D2D UEs
Cellular UE
Fig.1. The cooperative transmission scheme in D2D group communication
The multicast transmission proposed here is divided into three phases, that is, D2D group initialization, BS multicasting and
D2D cooperative transmission.

Phase 1: D2D group initialization. As all UEs are cellular UEs initially, some steps should be made before forming a D2D
group. D2D discovery is executed by each UE to achieve the surrounding D2D communication opportunities. All these
information should be reported to the BS as a reference for mode assignments. Because of the mobility of the UEs, the BS
should select the D2D group leader ( GL ) and members ( GM ) within Tpr to reach a stable state.

Phase 2: When the D2D group is established successfully, the BS would muticast the packets
during the downlink period and the UEs should receive and decode them .

Phase 3: After the muticast processing, both the D2D GMs and the cellular UEs are assumed to send ideal ACK/NACK
feedbacks to the D2D GL and the BS, respectively. According to the feedbacks, D2D GL will schedule a network coded
D2D cooperative transmission scheme with the assist of the BS via D2D link, multiplexing the uplink resources.
{x1 , x2 ,..., xN } to the group
In the next section, the proposed transmission scheme will be described in details.
III.
A NETWORK CODED D2D COOPERATIVE TRANSMISSION SCHEME
The BS multicasts {x1 , x2 ,..., xN } to all UEs who subscribe to the service, including cellular UEs and D2D UEs during the
downlink period. After a round of transmission, cellular UEs send the feedbacks in a traditional way and the D2D GL collects
feedbacks from D2D GMs. Now let’s focus on the D2D group.
A. Selection Method of D2D TX
In phase 3, when the D2D GL collects all feedbacks, it is his duty to select a D2D TX to be the cooperative relay. The
selection method is based on the feedback list. According to the feedbacks, four cases are concluded.
Case 1: All of the D2D UEs, including the D2D GL and GMs, receive and decode the packets successfully, so the D2D GL
sends ACK feedbacks to the BS to inform the success of transmission this round.
Case 2: m1 D2D UEs, including the D2D GL, receive and decode the packets successfully, so the GL chooses himself as the
transmitter in the group (D2D TX) during D2D communication.
Case 3: m2 D2D UEs, excluding the D2D GL, receive and decode the packets successfully, so the D2D GL chooses one of
them randomly as the transmitter in the group during D2D communication.
Case 4: None of the D2D UEs receive and decode all the packets, so the D2D GL selects the one with the most possibility
for decoding.
For D2D GL, measures would be taken as below to complete the selection of D2D TX. All these should be done right after
the multicast phase.
DCN 16-13-0165-00-Gcon
D2D GL scans the
feedback list.
All ACK?
NO
NO
YES
REPORT to the BS.
MYSELF ACK?
SELECTS a ACK D2D
GM randomly &
SENDS him the
feedback list.
YES
D2D cooperative
transmission.
Fig.2. The judging processing flow of D2D GL in selecting D2D TX
B. Opportunistic Network Coding Scheme
According to the feedback list, the selected D2D TX should encode the erased packets according to the proposed
opportunistic network coding scheme and multicast them to the target D2D CMs. To begin with, a simple example is introduced
that the feedback list of 3 D2D UEs is shown as below to describe the encoding method.
UE_ID
U1
U2
U3
PACK_N
7-13-25-27
2-13-18-23
3-5-6-7-11
Fig.3. An example of feedback list
UE_ID denotes the target D2D UEs, and the PACK_N is the number of the missing packets. In this case, U1, U2, U3 are the
target D2D CMs and P7, P13, P25, P27 are the missing packets of U1. D2D TX combines one missing packets of each UE in
sequence, that is P7  P2  P3 , P13  P5 , P25  P18  P6 , P27  P23  P7 and P11 . Duplicate missing packets would appear only once in
the same encoded packet. For more common cases, the criterion is min N D 2 D , where ND2D is the number of packets transmitted
during D2D communication period. J D2D UEs are assumed to have n j , j  1, , J  missing packets, respectively. D2D TX
would combine the ith ( i  1, , max  n j  ) missing packets to form min N D 2 D encoded packets.


C. Corresponding Decoding Scheme
In a retransmission round, min N D 2 D encoded packets are sent from D2D TX to those who fail to receive or decode all
packets via D2D links. Others will keep silent at the moment and wait for the next downlink multicast. In our case, U1 have P2 ,
P3 ,P5 ,P6 ,P7 ,P18 and P23. When U1 receives the 5 encoded packets, it will get P7 by XOR P2 and P3 with P7  P2  P3 . A formal
description of the corresponding decoding scheme is presented here. The target D2D UEs receive a block of encoded packets,
min N D 2 D in the first round of transmission and encode them with those packets they already have accordingly. In an ideal D2D
scenario, all encoded packets would be received successfully, so after a round of transmission, all erased packets would be
recovered. While the wireless channels are varying, even D2D communication may suffer from packets loss. So other effective
measures may be introduced to guarantee the efficient transmission, HARQ et al.
D. Analysis of Signalsl
A specific scenario of data sharing is analyzed, where there is a BS, N D2D UEs including a group leader (GL) and N-1
group members (GM). Signaling overheads in our scheme mainly arise after the DL muticast period. D2D members who fail to
receive and decode the packets successfully, NACK UEs for short, will transmit the HARQ feedback information nUL to GL via
D2D links, containing the user ID (UE_ID) and loss packet number (PACK_N). GL receives and decodes nUL to manage D2D
retransmission in the coming uplink period. Specifically, GL integrates all feedback information into a feedback list including the
retransmission data packet number (RE_NUM), number list of original data packets involved in the retransmission data packets
(PAC_LIST), the D2D TX UE ID (TX_ID), cooperative transmission target UE ID (OBJ_ID). This information is sent via D2D
DCN 16-13-0165-00-Gcon
link to the specified user (GL himself or the selected D2D TX UE). After the D2D cooperative transmission during uplink period,
NACK UEs still send feedbacks to GL, and GL collects and sends them to the BS. This is a transceiver cycle.
IV.
NUMERICAL EXAMPLES
This section demonstrates the effect of the proposed schemes by comparing with traditional schemes.
Firstly, transmission bandwidth  is defined as an exponent to evaluate the overhead of transmission. It is defined as followed:

ntotal
n
 1  retr
N
N
Where ntotal is the total number of transmission packets that BS has transmitted, nretr is the number of packets transmitted during
  1 , when it is an ideal error-free
the D2D communication period, and N is the number of packets should be transmitted.
system, equality holds.
According to current literature, we have got
M
 i j 1
(1) j1

 ARQ 
M
i1 ,i2 ,...,iM
1   p jj
i
j 1
Where i1 , i2 ,..., iM {0,1}, i j  0 .

1
1  max {Pi }
i 1,..., M
Equation applies to both  NCWBR and  IAA .
For simplicity, supposing that BS-to-UE links are erasure channels with erasure probabilities p1 , p2 ,..., pM during
transmission phases and error-free during retransmission phases. The number of transmitted packets in proposed scheme is
N  (1  max pi ) in any case. While in NCWBR scheme, the number of transmission packets starts as N  (1  max pi )
i 1,..., M
i 1,..., M
and increases
with the failure of recovering the erased ones when the retransmission packets combine more than one
erased
packet of a UE. For IAA, N  max pi is the lower bound of encoding sets.
i 1,..., M
2
ARQ
1.7
1.9
NCWBR
ARQ
Proposed Scheme
1.6
NCWBR
1.8
Index Allocation Scheme
Proposed Scheme
Theoretical Lower Scheme
1.7
Index Allocation Scheme
Theoretical Lower Bound
1.6

1.5

1.5
1.4
1.4
1.3
1.3
1.2
1.2
1.1
1.1
2
3
4
5
6
M
7
8
9
1
0.02
10
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
P


Fig. 4. versus M, when p=0.1,N=100.
Fig. 5. versus P, when M=5,N=100.
Figures are shown to prove advantage of proposed scheme over other coding schemes. The impacts of p, M, N on
transmission bandwidth  are considered, respectively. We assume all channels have equal erasure probabilities. A packet loss at
U i follows a Bernoulli trial with parameter pi . Fig.5. shows  increases almost linearly with the value of pi in case M=5,
N=100. Compared with current methods, our proposed scheme asymptotically coincides with the theoretical lower bound. When
pi =0.1, M=5, retransmission bandwidth of all the related schemes decrease with increasing N, a steep drop in the first 50 packets
1.55
1.5
1.45
NCWBR
Proposed Scheme
IAA
1.4
Theoretical Lower Bound

1.35
1.3
1.25
1.2
1.15
1.1
0
50
100
150
200
250
N
300
350
400
450
DCN 16-13-0165-00-Gcon
happens
Fig. 6.

versus N, when M=5,p=0.1.
after which the curve has slowed down, as Fig.6.presents. If N is larger than 300, proposed scheme performances even better.
Consider more and more active UEs joining the system, expressed as the growth of M, system performance suffers as a whole,
while proposed scheme achieves quite a stable performance in fig..
As is shown, increasing packet loss and number of UEs impose a negative influence on system efficiency which is
reasonable. On the other hand, as the number of packets grows, all network coding schemes harvest better performance,
manifesting the superiority of network coding in retransmission.
V.
CONCLUSION
A novel cooperative D2D communication scheme is proposed in this article when network coding is applied. The selection of
D2D TX, encoding and decoding methods are presented in details. Simulation results reveal that with the proposed scheme, the
D2D transmission efficiency can be improved a lot.