This is a pre-print version.

Accepted by IEEE Communications Standards Magazine under COMSTD-17-00031 1

Applying Device-to-Device Communication to

Enhance IoT Services
Ji Lianghai, Student Member, IEEE, Bin Han, Member, IEEE, Man Liu and Hans D. Schotten, Member,
IEEE

 low latency, they experience technical challenges to meet the

Abstract—Massive Machine Type Communication (mMTC) to requirements of mMTC services (e.g., low device cost, long
serve billions of IoT devices is considered to open a potential new device battery life and high service availability). In order to
market for the next generation cellular network. Legacy cellular obtain a deep market penetration in exploiting 5G to support
networks cannot meet the requirements of emerging mMTC
mMTC services, the Third Generation Partnership Project
applications, since they were designed for human-driven services.
In order to provide supports for mMTC services, current research (3GPP) has conducted studies to adapt and evolve legacy
and standardization work focus on the improvement and networks. For instance, to reduce device complexity, a new
adaptation of legacy networks. However, these solutions face type of user equipment (UE) is introduced as the category 0 in
challenges to enhance the service availability and improve the [3], which has a reduced peak data rate, a single antenna design
battery life of mMTC devices simultaneously. In this article, we and reduced bandwidth. Further cost reduction can be gained
propose to exploit a network controlled sidelink communication
by reducing the maximal transmission power [2] to simplify the
scheme to enable cellular network with better support for mMTC
services. Moreover, a context-aware algorithm is applied to integration of power amplifier (PA). However, it reduces the
ensure the efficiency of the proposed scheme and multiple context network coverage in uplink as a trade-off. Besides, since some
information of devices are taken into account. Correspondingly, mMTC devices are deployed in deep indoor scenarios, an extra
signaling schemes are also designed and illustrated in this work to penetration loss up to 20 dB can be foreseen [4]. In order to
facilitate the proposed technology. The signaling schemes enable maintain the uplink coverage, both narrow band transmission
the network to collect required context information with light
and massive transmission time interval (TTI) bundling [4] can
signaling effort and thus network can derive a smart
configuration for both the sidelink and cellular link. In order to help in serving deep coverage area, but lead to huge time
demonstrate the improvements brought by our scheme, a resource usage on the system level and battery drain on the
system-level simulator is implemented and numerical results show device level. It is, therefore, important to have a mechanism
that our scheme can simultaneously enhance both the service that is more efficient in power and resource to support wide
availability and battery life of sensors. deployment scenarios.
As one of the critical technical enablers for 5G networks,
Index Terms—5G, mMTC, IoT, D2D, cellular network,
signaling schemes, system level simulation device-to-device (D2D) communication [5] provides an
efficient alternative to cope with the requirements of mMTC
services. The original motivation of exploiting D2D
I. INTRODUCTION communication was to enable local information exchange for
high reliability and low latency [5]-[7]. In 3GPP, D2D
M ASSIVE Machine Type Communications (mMTC) [1],
[2] is widely considered as an important service to be
offered by the upcoming fifth generation (5G) cellular
communication has been standardized to enable discovery and
communication between two D2D devices [8]. Specifically,
networks. mMTC refers to a typical Internet-of-Things (IoT) UE-to-network relaying is introduced in LTE release 13 to
scenario, where a large amount of static sensors are deployed extend network coverage in public safety scenario, to assist
and report sporadically to an application server in the cloud automatic driving and achieve high transport efficiency. In that
(e.g., to enable environment monitoring and object condition scheme, base station (BS) can control the selection of relay UEs
tracking). While opening a new potential market, mMTC also by setting requirements on radio quality that a relay UE should
poses different requirements on network. Since legacy cellular fulfill. Moreover, the ranking of different UE-to-network relays
networks were designed for services with high data rate and is performed locally at remote UE and the remote UE may
select the relay with the best radio quality to itself. Since the
features above are designed to provide public safety, they are
L. Ji, B. Han, M. Liu and H. D. Schotten are with Institute for Wireless
Communication and Navigation, Department of Electrical and Computer
not optimized for the mMTC. Recently, this technology has
Engineering, University of Kaiserslautern, 67663 Kaiserslautern, Germany. been studied to enhance the mMTC services. In [9] [10], it is
Emails: {ji,binhan,manliu,schotten}@eit.uni-kl.de proposed that one cellphone can act as a relay with D2D links to
This work is copyrighted by the IEEE. Personal use of this material is other sensors. Thus, packets generated by those sensors can be
permitted. However, permission to reprint/republish this material for
advertising or promotional purposes or for creating new collective works for transmitted to BS through the relay. In 3GPP, the possibility to
resale or redistribution to servers or lists, or to reuse any copyrighted exploit D2D communication for mMTC applications is also
component of this work in other works must be obtained from the IEEE.
This is a pre-print version. Accepted by IEEE Communications Standards Magazine under COMSTD-17-00031 2

Figure 1: Our proposed approach allows the sensors with better propagation channel and enough battery level to relay uplink packets from other sensors.

studied in [11] where one cellphone acts as a relay for a group transmission is used to transmit system and control
of mMTC sensors. Since the D2D pairing procedure is information. Due to a higher available transmission power at
performed in a distributed manner without help from BS, it BS, mMTC services have better coverage in downlink than in
brings a loss in global awareness. Additionally, very few uplink. Therefore, in this work, we ignore the downlink
cellphones will appear in deep indoor or rural areas, especially performance but only inspect the uplink performance of mMTC
at night. Thus, the proposed approaches have limited services. Besides, we consider mMTC sensors as static and able
applicability. to perform sidelink communication with other sensors directly.
In this work, we refer the sensors in cell border as remote Figure 1 shows the proposed scheme by applying sidelink to
sensors, since they experience high pathloss for cellular links. enhance mMTC services. Sensor #2 experiences a better
We inspect on how to improve the availability of the remote propagation channel to BS than sensors #4 and #5, due to a
mMTC sensors and meanwhile also attempt to have more smaller distance to the BS. If the battery level of sensor #2 is
sensors fulfilling the battery life requirement. A scheme is high enough, it can be configured by the BS to set up sidelink
proposed where sidelink communication can be performed communication with sensors #4 and #5, and thus relay their
among mMTC sensors. In this way, remote sensors have the data packets to the BS. In order to improve the performance,
opportunity to efficiently set up sidelinks with relay sensors. It context information should be collected and exploited by the
worths to note that, in 3GPP, the term "sidelink" refers to a network to set up sidelink communication. For instance, the
direct link between two devices over air interface, while the sensors in virtual cluster #m should not have direct sidelinks
term "D2D communication" normally refers to a general with the ones in virtual cluster #n, since they are geometrically
system structure to support the direct link. This article provides separated. As the figure illustrates, three TMs exist for mMTC
some insights for the future standardization work in 3GPP to sensors.
apply D2D communication in mMTC. Our solution differs  Cellular TM, where sensors upload their reports to BS via
from the work of LTE release 13 [8] at following points: cellular links.
 Our focus is the smart transmission mode (TM)  Relay TM, where sensors receive reports from other
configuration of relay and remote UEs in mMTC sensors and then transmit both the received packets and
scenario. So far, the corresponding technology is their own reports to BS.
designed in 3GPP for public safety scenario.  Sidelink TM, where sensors transmit reports to
 In our scheme, D2D groups are pre-selected by BS, while corresponding relay sensors.
the current approach in 3GPP is that the remote UE
locally performs the ranking of the potential III. DEVICE CLUSTERING AND TM SELECTION
UE-to-network relays. As aforementioned, to achieve a high efficiency, sidelink
 mMTC-related context information (e.g., sensor location, communication should only happen between sensors located
battery level) are collected with reasonable signaling closely to each other. An efficient approach here is to perform
overload to improve service availability and power device clustering algorithm and allow only intra-group sidelink
consumption of sensors. communication. Thus, the context-aware sidelink
To describe the proposed scheme, our article is organized as communication scheme comprises of two steps, as:
following. At first, the system model for the proposed scheme 1) clustering of mMTC devices;
is introduced and then a context-aware clustering scheme is 2) selection of TM.
provided to guarantee the sidelink transmission efficiency. Once a sensor is initially attached to the network or the network
Following that, a smart TM selection procedure is also given, in needs to update the TMs of the serving sensors, the above two
order to configure each sensor properly. Afterward, signaling steps should be carried out in BS.
schemes to support the proposed technology are illustrated.
Additionally, to compare the performance of the new scheme A. Clustering of mMTC devices
with LTE network, a system level simulator is implemented Since a sidelink is applied for two nearby devices, an
and numerical results are shown. efficient clustering algorithm can assist remote sensors to find
proper relay sensors. The design and implementation of an
II. SYSTEM MODEL efficient clustering should take account of useful context
Many mMTC services request sensors to sporadically report information, which refers to the information (e.g., location
their status with a short payload in uplink. Downlink information, traffic type, battery life requirement, etc) that can
help to improve system performance. For instance, remote
This is a pre-print version. Accepted by IEEE Communications Standards Magazine under COMSTD-17-00031 3

sensors in rural area are located far away from the BS. In this 4) The configuration message is sent to both relay and
sense, by analyzing the location information of sensors, a remote sensors to trigger the sidelink discovery
sensor located between the remote sensor and the BS can be procedure.
recognized and considered as a relay. In contrast, in a dense 5) In sidelink discovery procedure, a remote sensor is
urban scenario, sensors located deep indoor or in basements are configured to sidelink TM, if its sidelink channel
seen as remote sensors due to the high penetration loss. pathloss is below a threshold. Otherwise, it remains in
Therefore, the relay sensors, in this case, may be the sensors cellular TM.
located on higher building floors. In this work, our proposed 6) Other sensors not involved in sidelink communication are
scheme is inspected in a rural area scenario. Thus the clustering configured to cellular TM.
algorithm is adapted correspondingly. At first, BS converts the To be noticed, in rural areas, the remote sensors mentioned in
location of sensors to their reference angles. And then the K step 1) are located far away from the BS. Thus, relay sensor
-means clustering algorithm w.r.t. the reference angles of candidates with distance to BS larger than a given radius, as
sensors is performed in BS. The basic steps of the K-means stated in step 3), can ensure them to be in the proximity of the
algorithm are listed below: remote sensors. Taking its power consumption and signaling
1) Initially, BS selects K sensors with reference angles effort into account, the period of sidelink update procedure
separated from each other as far as possible, these should be larger than the minimal uplink transmission period of
sensors are considered as centroids of K clusters. sensors. Otherwise, multiple sidelink update procedures may
2) Take another sensor and associates it to the cluster, take place between two consecutive reports of a sensor, but
whose centroid is the closest to the selected sensor only the last sidelink update procedure makes sense. On the
w.r.t. reference angle. other hand, increasing the update period reduces the network
3) Calculate the mean reference angle of the updated flexibility to respond to condition changes. For instance, since a
cluster, and select the sensor in this cluster nearest to relay sensor forwards the packets of a group of sensors, it
the mean value as the new centroid. experiences a high battery power drain. Therefore, the network
4) Repeat steps 2) and 3) until every sensor is associated needs to update the sidelinks in a timely manner, to prevent
with a cluster. from the case where a relay sensor powers out and the remote
Since sensors are assumed to be static in our considered sensors will lose connections until the next sidelink update
mMTC scenario, BS can acquire their location at their initial procedure takes place.
attachments to the network, as stated in the next section.
IV. RADIO LINK ENABLER TO SUPPORT THE CORRESPONDING
B. TM selection
SCHEME
Once a cluster is formed, BS configures every sensor in this
To support the proposed scheme, corresponding signaling
cluster to a proper TM. Context information is helpful to
procedures are introduced in this section. Three important
optimize the TM of each device. For instance, a higher pathloss
procedures, including initial attachment of sensors, the update
of a cellular link introduces a higher power consumption per
of TM and uplink transmission exploiting sidelink
uplink transmission. Thus, the battery life requirement might
communication, are provided with details.
not be met for this sensor and BS tries to assign a relay to
improve its battery usage. Moreover, some sensors even cannot
reach the BS in uplink due to their extreme high pathloss Figure 2: Initial attachment of a new sensor
values, therefore they are also assigned to a relay sensor which
can contribute to improving the service availability. From the
efficiency perspective, not all sensors with good cellular links
are suitable to act as relays, since the sidelink radio quality also
plays a critical role in power consumption. Besides, since each
relay multicasts a sidelink discovery message to neighbor
nodes during the sidelink setup or update procedure, the
number of relay sensors should not be too high. Otherwise, A. Initial attachment of a sensor
more time and frequency resource are required for the sidelink
Once a sensor is deployed and powered on, an initialization
discovery process. In this work, with the help of context
procedure shown in Fig. 2 is required. The signaling diagram is
information, the transmission mode selection (TMS) of sensors
illustrated with more details in the following:
in each cluster is performed as following:
1) The sensor receives the sidelink system information
1) Sensors out of the uplink coverage and sensors with
blocks (S-SIBs), then reports its channel state
battery life lower than requirement are configured to the
information (CSI), location and battery information to
sidelink TM.
the BS. Since the deployment of the new sensor is
2) Sensors with good cellular links and enough battery
performed manually by technicians, such information
capacity are considered as relay sensor candidates.
can be submitted by equipment of technicians though
3) A relay sensor candidate with distance to the BS larger
the device might be in the uplink outage. Other context
than a given threshold is selected as the relay sensor.
information such as traffic type can be requested by the
This is a pre-print version. Accepted by IEEE Communications Standards Magazine under COMSTD-17-00031 4

BS from the corresponding application server.

Figure 3: Procedure for TM update
Afterward, BS will perform the sidelink clustering and
TMS algorithm, taking account of the collected context
information.
2) In case the sensor is configured to either relay or sidelink
TM, the BS transmits the ID of its virtual cluster and
other control information for sidelink discovery and
communication to this sensor.
3) If the new sensor is configured to sidelink TM with sensor in Fig. 3 where two sensors are to be updated with the sidelink
#m as the corresponding relay, sensor #m sends a TM. This scheme can be easily extended to the case where
sidelink discovery announcement message containing more sensors are assigned with sidelink TM. The signaling
the ID of the new sensor. In another case, if the new diagram is illustrated with more details in the following:
sensor is configured to relay TM for sensor #m, it 1) BS performs the context-aware TMS algorithm.
transmits the sidelink discovery message to sensor #m, 2) BS pages the corresponding sensors and configures
where the ID of sidelink counterpart is also carried. The them with their new TMs. Besides, other dedicated
time and frequency resources used to transmit the control information is also transmitted (e.g., IDs of
discovery message are indicated in step 2). In this step, the sensors involved in sidelink, resource for sending
reference symbols to estimate sidelink channel will also sidelink discovery message). In Fig. 3, sensors #m
be transmitted. and #n are configured to sidelink TM and another
4) Upon receiving the sidelink discovery message, the sensor is configured to relay TM.
receiver sensor determines whether the discovery 3) The relay sensor sends a discovery announcement
request should be accepted, based on the estimated message to the sensors in sidelink TM, with the IDs of
sidelink reference signal received power (S-RSRP). An itself and the conveyed target sensors. Reference
acknowledgment (ACK) / non-acknowledgment signals are embedded in this message.
(NACK) message is transmitted back. 4) Based on the S-RSRP, receivers of the discovery
5) If the discovery procedure is acknowledged, a security message (i.e. sensors #m and #n) determine whether
association between the two sidelink ends can be the sidelink request is accepted. The decision will be
established by exchanging security-related messages. fed back to the relay sensor.
6) The discovery decision is further forwarded to the BS by 5) If the request is accepted, a security association is
the relay sensor, thanks to its good cellular channel established.
condition. In case a NACK message is received, the BS 6) The results of TM update are further transmitted from
will configure the remote sensor to a cellular TM and the relay sensor to the serving BS. If a sensor fails in
avoid setting up the same sidelink ends in future. establishing sidelink with the relay sensor, BS
In the procedure above, the S-SIBs provide the less configures it to cellular TM and avoids to pair these
frequently changed configuration information for sidelink two sensors for sidelink communication in future.
(e.g., resource pools used for sidelink discovery and As the relay sensor transmits a discovery announcement to a
communication). Thus, the relay sensor transmits the group of remote sensors in step 3), each sensor in step 4) can
discovery announcement message on the configured respond by picking up a resource from the resource pool, which
resource so that the monitoring sensors can receive and is indicated in the configuration information in step 2).
process the messages from the same resource. Moreover, Moreover, if a sidelink is successfully established in step 5), the
since both the sidelink ends are already synchronized to the sensors in sidelink TM should be aware of the discontinuous
same access node in downlink and they are statically located reception (DRX) cycle of the relay sensor, in order to derive the
within the proximity of each other, no synchronization time instance that the relay sensor wakes up to receive the
procedure between them is needed. Last but not least, certain packets.
context information (e.g., security information) should be
stored at both the sidelink ends once the sidelink connection C. Reports transmitted by sidelink communication
is established. In this way, a fast sidelink connection In case certain sensors in sidelink TM are paged by network
resumption can be obtained with less signaling effort, in or have data in their buffer to transmit, the established sidelinks
order to efficiently support the sporadic traffic of mMTC are exploited to relay the data in the uplink. Figure 4 shows this
services. transmission procedure and it is further illustrated in the
following:
B. TM update procedure
In case some conditions are no more fulfilled for sidelink
communication (e.g., the battery level of a relay sensor
decreases severely and is not sufficient to support the relay
functionality anymore), TMs of the sensors in that cluster need
to be updated. The corresponding signaling diagram is provided
This is a pre-print version. Accepted by IEEE Communications Standards Magazine under COMSTD-17-00031 5

1) Sensors involved in sidelink communication receive the performed as a normal cellular uplink transmission
S-SIBs. where a control plane (CP) connection needs to be
2) In the mobile terminated case, one or multiple sensors in established. Further, network can configure the relay
sidelink TM will be paged to trigger their reporting sensor to compress its own uplink packet together with
procedure by the BS. In the paging message, dedicated the packets from remote sensors and send them together
resource for the sidelink communication can be to the BS. In this case, less power is consumed at relay
assigned. In Fig. 4, sensor #m, sensor #n and the sensor to transmit both its own packet and the packets
corresponding relay sensor are paged by network to from remote sensors, since the relay sensor needs to
transmit their data reports. wake up and perform the CP connection establishment
3) In the mobile originated case, sensors in sidelink TM (i.e., procedure only once.
sensor #m and sensor #n in Fig. 4) try to transmit their 6) Upon successfully receiving packets from the relay
data packets through sidelink. There are two options to sensor, BS sends an ACK message back. Otherwise, a
obtain resource for the sidelink transmission. One is that re-transmission procedure is triggered by a NACK
multiple remote devices can randomly access to the message.
relay sensor by sending different preambles and the In order to update the context information at BS, real time
relay sensor replies with the assigned resource. Another sensor information such as battery level should be transmitted
option refers to a semi-persistent resource allocation in the data reports from both the relay and remote sensors.
scheme where the time and frequency resource for the
sidelink transmission of each sensor is pre-allocated. V. EVALUATION AND NUMERICAL RESULTS
4) After successfully receiving packets from remote sensors
(i.e. sensors #m and #n in Fig. 4), relay sensor replies
with ACK messages back. Otherwise, a NACK message
is transmitted back and triggers the re-transmission
procedure.
5) Relay sensor further forwards the successfully received
packets to the serving BS. This process can be

Figure 4: Uplink report procedure

This is a pre-print version. Accepted by IEEE Communications Standards Magazine under COMSTD-17-00031 6

TABLE I: Device power consumption parameters

(a) CDF of served days of all deployed mMTC devices (b) CDF of served days of mMTC devices from two different sets (in the coverage or outage of LTE)
Figure 5: System performance of mMTC devices w.r.t. their served days in uplink

For evaluation, a system level simulator is used where the enough relay candidates. In LTE network, if a radio link
service availability and battery life of mMTC sensors are taken experiences a poor SNR (e.g., lower than -7 dB), no data
as key performance indicators. In the simulation, one single BS transmission is possible on this link and the sensor is in the
is deployed in rural area with a cell radius of 2500 meters, as outage of network. In Fig. 5a, the performance of all deployed
defined in [12]. One hundred thousand sensors are uniformly sensors is given. As it shows, more than six percent of sensors
distributed inside the cell at a height of 0.5 meter and each are in the outage of the BS in LTE release 12 and unable to
sensor has a maximal transmission power of 20 dBm, as transmit data to the BS in uplink. The discrete steps in the curve
proposed for LTE-M [2]. 900 MHz is assumed to be the carrier are caused by the fixed modulation and coding schemes (MCS)
frequency for the air interface. Moreover, since different applied by LTE technology. Compared with the performance of
subbands are used for cellular link and sidelink, there is no LTE release 12, both LTE release 13 and our proposed
mutual interference between them. The model proposed in [13] context-aware D2D scheme improve the service availability so
is implemented for the sidelink channel. Regarding the traffic that all sensors can be served by either cellular link or sidelink.
model, sensors are required to periodically transmit their Additionally, 60 percent of the sensors have service life over 10
reports with a payload of 1000 bits, at a frequency of one report years in LTE release 12 and it can be improved to 78 percent by
per 150 seconds. BS will consider a sensor as a relay candidate, LTE release 13, while our proposed scheme raises this ratio to
only if this sensor has a distance to the BS larger than 1500 95 percent. As aforementioned, relay sensors experience high
meters and meanwhile also a battery life which is 20 percent power consumption and this is the reason that both the
longer than the service requirement. Moreover, both cellular context-aware D2D and LTE release 13 schemes have fewer
link and sidelink apply open power control scheme to achieve sensors with long serving days. For more detailed inspection on
target signal-to-noise ratio (SNR) (3 dB for cellular link and 10 our proposed scheme, sensors are categorized into two different
dB for sidelink). To evaluate battery life performance of sets based on whether they are in the coverage of LTE network.
mMTC sensors, parameters related to power consumption are In Fig. 5b, the system performances of two different sets are
given in Table I. Note that 10 ms is assumed to be the duration provided separately. Regarding the performance of LTE release
for CP connection establishment procedure. This is the value 13, since no UE battery information is presented and controlled
targeted by the 5G network while the legacy 4G network at the BS, relay UEs can run out of their battery very quickly.
requires approximately 50 ms to set the CP connection. Other Therefore, it can be seen that the UEs in the coverage of the BS
parameters (e.g., channel model for cellular link, noise figure, will have worse performance compared with the LTE release
etc.) are aligned with the ITU-R document [14]. Last but not 12. On the other hand, due to the performance sacrifice of the
least, LTE technology is used to model radio link performance. relay UEs, the UEs out of BS coverage can be well supported in
In Fig. 5, cumulative distribution function (CDF) of served time LTE release 13. Compared with the scheme in LTE release 13,
of mMTC sensors is shown. As a baseline scheme, the the proposed context-aware D2D scheme contributes to a better
performance of LTE release 12 is given where relay is not control of both the relay (i.e., the dark blue curve) and remote
applied. Moreover, since LTE release 13 has already UEs (i.e., the black curve). With respect to sensors in the outage
standardized the UE-to-network relaying for public safety of the BS of LTE release 12, they can be served by sidelink
scenario, its performance on the considered mMTC scenario is communication and 80 percent of them can even be served for
also provided. An SNR value of 6 dB in uplink should be more than 10 years. It should be noticed that, though these
fulfilled for being a relay node in both the LTE release 13 and sensors are served by sidelink communication which consumes
our proposed context-aware D2D schemes. Additionally, both less power for data transmission than the cellular link, there are
schemes schedule one hundred relays at each time, if there are still 20 percent of users that cannot achieve the battery life
This is a pre-print version. Accepted by IEEE Communications Standards Magazine under COMSTD-17-00031 7

requirement of 10 years. This is due to the fact that, since relay [6] J. M. B. da Silva, G. Fodor, and T. F. Maciel, “Performance Analysis of
Network-Assisted two-hop D2D Communications,” in Globecom
sensors experience high battery drain from forwarding the Workshops (GC Wkshps), 2014. IEEE, 2014, pp. 1050–1056.
packets from remote UEs, they might lose the ability to be a [7] A. Osseiran, J. F. Monserrat, and P. Marsch, 5G Mobile and Wireless
relay after some time. Thus, after some time, there are not Communications Technology. Cambridge University Press, 2016.
[8] “Proximity-Based Services (ProSe) (Release 14),” Third Generation
enough suitable relay sensors in the system and some remote
Partnership Project (3GPP), Tech. Rep. TS 23.303, Dec. 2016.
UEs cannot be served by sidelink communication any more. [9] N. K. Pratas and P. Popovski, “Underlay of Low-Rate Machine-Type
Additionally, performance comparison of the sensors in the D2D Links on Downlink Cellular Links,” in 2014 IEEE International
coverage of the BS is also shown in this figure, and we can see Conference on Communications Workshops (ICC), IEEE, 2014, pp.
423–428.
that the ratio of sensors meeting the battery life requirement is [10] ——, “Low-Rate Machine-Type Communication via Wireless
increased from 76 percent in LTE (i.e., the red curve) to 99 Device-to-Device (D2D) Links,” arXiv preprint arXiv:1305.6783, 2013.
percent in the context-aware D2D scheme (i.e., the dark blue [11] “New WI proposal: D2D Based MTC,” Third Generation Partnership
Project (3GPP), Tech. Rep. RP-151948, Dec. 2015.
curve). Please note that, since the proposed context-aware D2D [12] “Study on Scenarios and Requirements for Next Generation Access
scheme targets at a battery life requirement of 10 years, the Technologies (Release 14),” Third Generation Partnership Project
remote UEs will not be served by relay UEs any more, once this (3GPP), Tech. Rep. TR 38.913, Oct. 2016.
[13] “Discussion on UE-UE Channel Model for D2D Studies,” Third
requirement has been fulfilled. This is the reason why a step Generation Partnership Project (3GPP), Tech. Rep. R1-130092, Jan.
from 1 percent to 24 percent occurs for the context-aware D2D 2013.
scheme (i.e., the dark blue curve) at the point of 10 years, which [14] “Guidelines for Evaluation of Radio Interface Technologies for
IMT-Advanced,” ITU Radiocommunication Sector (ITU-R), Tech. Rep.
shows that 23 percent of UEs are served exactly for 10 years by M.2135, 2008.
the context-aware D2D scheme.

VI. CONCLUSION AND FUTURE WORK

In this work, mMTC services are enhanced by a
context-aware sidelink communication scheme. Three critical Ji Lianghai (S’17) received his B.Sc degree from the
signaling procedures to support the proposed scheme are Shandong University, China in 2010 and his M.Sc. degree from
described, with respect to the initial attachment of sensors, the University of Ulm, Germany in 2012. He is at this moment
update of TM and sidelink communication procedure. working for his Ph.D degree at the institute of Wireless
Meanwhile, the proposed signaling schemes enable the Communication and Navigation, University of Kaiserslautern,
collection of required context information without a heavy Germany. He worked for European 5G flagship project METIS
signaling overload. Moreover, the numerical results gained and some other 5G projects with industrial partners. He is
from a system level simulation show that our scheme improves currently representing University of Kaiserslautern in 5GPPP
both the service availability and the battery life of mMTC project METIS-2.
devices simultaneously.
Bin Han (M’17) received his B.E. degree in 2009 from the
ACKNOWLEDGEMENT Shanghai Jiao Tong University, China, and his M.Sc. degree in
Part of this work has been performed in the framework of 2012 from the Technische Universität Darmstadt, Germany. In
H2020 project METIS-II, which is funded by the European 2016 he was granted the Ph.D. (Dr.-Ing.) degree from the
Union. The views expressed are those of the authors and do not Kalsruhe Institute of Technology, Germany. He is currently a
necessarily represent the project. The consortium is not liable postdoctoral researcher at the Institute of Wireless
for any use that may be made of any of the information Communication and Navigation, University of Kaiserslautern,
contained therein. Germany. His research interests are in the broad area of
wireless communication systems and signal processing.

REFERENCES
Man Liu received her Diplom degree from the University of
Kaiserslautern, Germany in 2010. She is at this moment
[1] S.-Y. Lien, K.-C. Chen, and Y. Lin, “Toward Ubiquitous Massive
Accesses in 3GPP Machine-to-Machine Communications,” IEEE
working for her Ph.D degree at the institute of Wireless
Communications Magazine, vol. 49, no. 4, 2011. Communication and Navigation, University of Kaiserslautern,
[2] “Study on Provision of Low-Cost Machine-Type Communications Germany. Her research interests are on the application of 5G
(MTC) User Equipments (UEs) Based on LTE (Release 12),” Third network for machine type communication.
Generation Partnership Project (3GPP), Tech. Rep. TR 36.888, Jun. 2013.
[3] “User Equipment (UE) Radio Access Capabilities (Release 13),” Third
Generation Partnership Project (3GPP), Tech. Rep. TR 36.306, Dec.
2016. Hans D. Schotten (S’93-M’97) received the Diplom and Ph.D.
[4] “Cellular System Support for Ultra-Low Complexity and Low degrees in Electrical Engineering from the Aachen University
Throughput Internet of Things (CIoT) (Release 13),” Third Generation
Partnership Project (3GPP), Tech. Rep. TR 45.820, Nov. 2015.
of Technology RWTH, Germany in 1990 and 1997,
[5] K. Doppler, M. Rinne, C. Wijting, C. B. Ribeiro, and K. Hugl, respectively. Since August 2007, he has been full professor and
“Device-to-Device Communication as an Underlay to LTE-Advanced head of the Institute of Wireless Communication and
Networks,” IEEE Communications Magazine, vol. 47, no. 12, 2009. Navigation at the University of Kaiserslautern. Since 2012, he
This is a pre-print version. Accepted by IEEE Communications Standards Magazine under COMSTD-17-00031 8

has additionally been Scientific Director at the German

Research Center for Artificial Intelligence heading the
"Intelligent Networks" department.
This is a pre-print version. Accepted by IEEE Communications Standards Magazine under COMSTD-17-00031 9

Figure 1: Our proposed approach allows the sensors with better propagation channel and enough battery level to relay uplink packets from other sensors.
This is a pre-print version. Accepted by IEEE Communications Standards Magazine under COMSTD-17-00031 10

Figure 2: Initial attachment of a new sensor

This is a pre-print version. Accepted by IEEE Communications Standards Magazine under COMSTD-17-00031 11

Figure 3: Procedure for TM update

This is a pre-print version. Accepted by IEEE Communications Standards Magazine under COMSTD-17-00031 12

Figure 4: Uplink report procedure

This is a pre-print version. Accepted by IEEE Communications Standards Magazine under COMSTD-17-00031 13

(a) CDF of served days of all deployed mMTC devices (b) CDF of served days of mMTC devices from two different sets (in the coverage or outage of LTE)

Figure 5: System performance of mMTC devices w.r.t. their served days in uplink
This is a pre-print version. Accepted by IEEE Communications Standards Magazine under COMSTD-17-00031 14

TABLE I
DEVICE POWER CONSUMPTION PARAMETERS
Parameter Description Value Time duration if
applicable
power consumption during 45% PA efficiency plus 60 MCS and packet size
transmission mW/s for other circuitry related
power to receive packets from 100 mW/s MCS and packet size
remote sensors related
power to receive paging 100 mW/s 10 ms
command
power to obtain 100 mW/s 10 ms
synchronization
power consumption during 200 mW/s 10 ms
the control plane
establishment procedure
power consumption in 0.01 mW/s time of sensor staying
sleeping mode in sleeping mode
number of DRX cycles per day 4 times/day
battery capacity 5 Wh
data reporting period 150 s
packet size 1000 bits
length of DRX cycle 6 hours
periodicity of TM update 1 day