Professional Documents
Culture Documents
S M A RT B A C K H A U L I N G A N D F R O N T H A U L I N G
F O R 5G N E T W O R K S
DMRS
frequency
•••
trolled. This gives an edge to cellular over ad
hoc networks, which utilize license-exempt fre-
quencies. To achieve reliability in time-varying
Short subframe length communication channels, either very rapid feed-
GP back or overprovisioning of resources is required.
Highly reliable transmission requires more band-
Figure1. TDD optimized subframe structure. width than less reliable transmission. Ultimately,
in URC one is interested in a delay-limited out-
age capacity. For factory automation, this calls
self-backhauling and multihop, control informa- for high-capacity indoor cellular networks, where
tion needs to be signaled not only between access capacity can be traded for reliability.
nodes and terminals but also among the access For self-backhauling with controlled reliabili-
nodes in the self-backhauling chain. Since in ty, we consider decode-and-forward (DF) relay-
half-duplex (HD) communication, transmission ing. In DF-multihop networks, end-to-end
and reception cannot be done simultaneously, a throughput decreases rapidly when the number
TDD problem arises where two nodes with the of hops increases due to HD losses. By making
same TX/RX timing cannot communicate with relays capable of FD forwarding, throughput can
each other. To enable bidirectional communica- be significantly increased if self-interference can
tion in a TDD environment where all nodes are be cancelled [8]. In practice, FD devices with
able to listen to each other, TX/RX patterns and compact form factors suffer from self-interfer-
pattern groups [4] can be utilized. In a flexible ence caused by the coupling of the transmitted
frame structure according to Fig. 1, the TX and signal to the receiver chain, which decreases the
RX parts of the control plane, and the commu- gains from FD operation.
nication direction in the data plane would thus In this article, we discuss in-band self-back-
switch accordingly on a per node basis. If self- hauling for dense indoor 5G deployments. A
backhauling nodes operate in full-duplex (FD) new category of low-cost relaying APs (R-APs)
mode (i.e., they are capable of transmitting and are considered for 5G small-cell networks. These
receiving simultaneously in the same resource operate in a TDD fashion, and serve as APs
block), problems related to control signal com- toward clients while relaying traffic between fully
munication restrictions are mitigated. To reach equipped APs (F-APs) and client devices. Both
5G targets on 1 ms end-to-end latency in a mul- FD and HD operation will be addressed. Mecha-
tihop system, potentially with a hybrid automatic nisms for canceling self-interference in FD relays
repeat request (HARQ) retransmission protocol, are considered to set realistic performance limits
the subframe has to be short. A subframe of 0.25 for FD relays. Despite using projected state-of-
ms is considered feasible in terms of control the-art technologies, not all self-interference can
overhead in local area networks [3]. Time-multi- be canceled. The residual self-interference may
plexed control and data planes enable the receiv- be on the level of other interference in the net-
er to process its dedicated control information work (e.g., in two-hop self-backhauling) to the
while transmitting/receiving data, which further interference from the source transmission to the
reduces latency. The time-multiplexed structure destination. To balance the self-interference
also reduces energy consumption, because against other interference, link adaptation and
receivers do not need to receive and buffer data radio resource management (RRM) techniques
that is not targeted to them. Note that this is dif- can be used. Power control, resource allocation,
ferent from Long Term Evolution (LTE), where and multiple-input multiple-output (MIMO)
the control plane is multiplexed in time and fre- techniques can be used to mitigate interference.
quency, and receivers have to detect both data This can be with or without coordination
and control planes [5]. between the backhaul and access links. Tight
Another aspect of prospective 5G networks is coordination enabled by RRM signaling provides
related to ultra-reliable communication (URC) significant gains for self-backhauling systems.
[6], where highly reliable and low-latency com- Assuming a reasonable residual self-interfer-
munication is required, primarily for machine- ence level, achievable with state-of-the art tech-
type communication. URC is a key enabler for nologies, a multihop network with FD relays and
enhanced wireless automation (e.g., in factory RRM coordinating backhaul and access links sig-
environments). To replace wires in factories, nificantly outperforms a network with HD relays,
near-real-time highly reliable wireless communi- and a conventional single-hop cellular network.
cation technologies are needed. Wireless com-
munication provides clear advantages for flexible
realization of wireless automation: ease of SELF-BACKHAULING NETWORKS WITH
deployment, mobility, and scalability. Despite FULL-DUPLEX RELAYS
these benefits, compromised reliability and low
capacity [7] have kept wireless solutions in a To enable optimization of frequency and time
marginal role. This will change in 5G, with usage in a self-backhauling 5G network, we con-
natively designed support for URC. Both cellu- sider APs that act as in-band relays. Different
lar and D2D/M2M communications are foreseen from conventional relaying in cellular systems,
in this context. where carefully placed relay nodes mitigate cov-
MBB:48
same placement of F-APs and R-APs as the With 80 dB SI cancela-
F-APs in the network buildings in the MBB scenario, except that there
URC: 16
are no inner walls. In the factory, non-line-of tion, we see a clear
sight is assumed, with path loss exponent 3.
R-APs in the network
MBB:480
Some of the most important simulation parame-
gain from coordinated
URC: 160
ters can be found in Table 1. RRM with FD. If RRM is
We consider downlink flows. Each R-AP is
Floors in the buildings
MBB: 3
associated with an F-AP with the smallest direct enabled to choose
URC: 1
link path loss, thereby dividing the network into between HD and FD,
cells. Cell selection for the UEs is similarly
Boundary conditions
Wrap-around in
based on average path loss. UEs are distributed some additional gains
XY-direction
uniformly in the buildings. It is assumed that
there is an active UE in 20 percent of the cells,
would be realized. Direct
Maximum Tx power
(F-APs and R-APs)
24 dBm and in each cell there is only one active UE. The link cancelation/combi-
MIMO channels between the transmitters and
receivers, as well as between the transmitters nation provides little
Carrier frequency 3.6 GHz
and all interference victims in the network, are gain with coordinated
modeled taking distance-dependent path loss,
Bandwidth
MBB: 100 MHz shadow fading, and Rayleigh-distributed uncor- RRM. However, with
URC: 5 MHz
related fast fading into account. After RRM, the
signal-to-interference-plus-noise ratios (SINRs)
hypothetical 100dB SI
Bandwidth efficiency 0.9
MIMO configuration 1 × 1, 2 × 2, 4 × 4
of the receivers are calculated for each MIMO cancelation, when the
layer transmitted.
For MBB, we are interested in high data direct link interference is
MBB: 5 dB
Inner wall loss
URC: 0 dB
rates, and rely on end-to-end HARQ for reliabil- the bottleneck, direct
ity. This is realized by applying a margin when
Table 1. System-level simulation parameters. controlling the SINR of the self-backhauling link IC provides
hop. Transmission packets to the users are trans-
mitted in subframes with segments of equal size,
noticeable gains.
and the flow is operating for tens of subframes.
SIMULATION RESULTS Physical layer throughput is modeled by using
SIMULATION ASSUMPTIONS adaptive modulation and coding with a through-
put close to Shannon’s law, with an implementa-
For mobile broadband services, the highest tion margin of 2 dB. The largest SINR allowed
capacity requirements are expected in indoor by the RF impairments is 25 dB, which sets the
locations, and URC networks for factory highest transmission rate in the system. Band-
automation are also predominantly indoors. To width efficiency is 90 percent, as in LTE, which
address the potential of FD multihop relaying, decreases achievable rates in linear manner.
we thus consider indoor small cell networks con- For URC, HARQ is not applied against
sisting of F-APs that have a wired connection to SINR variations arising from imperfect precod-
the Internet, and R-APs connected by self-back- ing, and variations in interference. Note that the
haul to an F-AP. The APs collaborate to serve a self-interference experienced in the R-AP is a
population of users or machines. random variable, which causes fading in the F-
To address 5G objectives for both MBB and AP to R-AP communication in this scenario. In
URC, we perform simulations in two scenarios. addition, mobility causes errors in precoder
For MBB, we consider an indoor local area selection and CQI. For each flow toward a user,
deployment in a dense urban environment. A we consider an implementation limited outage
self-backhauling network is deployed in a Man- capacity. A rate that is achievable with 99.9 per-
hattan grid of multi-floor buildings. The simula- cent reliability over the self-interference and
tion model is discussed in [9], and is based on mobility-induced fading is considered. In the
Winner path loss models [15]. The buildings are simulation, we consider an environment where
according to the indoor-office A1-model of [15]. machines are moving with 20 km/h velocity.
The basic unit of a building is a 10 × 10 m room, With a 0.25 ms subframe and two-hop communi-
and there are two 5-m-wide corridors between cation, a CQI delay of 0.5 ms may be expected.
rows of rooms. The size of the building is 50 × Accordingly, measured channels and channel
100 m consisting of 40 rooms per floor. There realizations at the time of transmission are 99
are three floors in each building, and four F-APs percent correlated. We use an eigenbeam trans-
per floor located in the corridors. The two F- mitter and receiver, with separate coding per
APs in the same corridor are 50 m apart. One MIMO layer. To decrease outage probability, a
R-AP is placed in a random position in each margin is applied to measured CQIs. With a first
room. There are four buildings in the model order approximation to estimate the statistics of
placed in a Manhattan grid 20 m apart from realized channels, a 12 dB margin is sufficient to
each other. Wrap-around boundary conditions guarantee target reliability. For URC, short sub-
are applied in the two horizontal directions. Fol- frames are highly important to increase the relia-
lowing the indoor-to-outdoor model A2 in [15], bility of feedback information, especially CQI.
the loss in the outer walls of the buildings is For self-interference cancellation at the FD
between 14 and 29 dB depending on the direc- R-AP, we consider two values, 80 and 100 dB.
tion. For URC, we consider a factory scenario The former may be realistic in future wideband
consisting of a Manhattan grid of one-floor fac- transmissions as foreseen in MBB, and the latter
tories. Each factory has the same size, and the may be realistic in narrower bands [10, 11], per-
500 1
HD
450 0.9 1x1
400 FD 80 dB SI
uncoordinated 0.8
CDF
250 0.5
FD 100 dB SI
uncoordinated
200 0.4
Figure 4. MBB scenario: Left: Throughput at 5 percent of user throughput CDF, gain [%] relative to direct communication with
F-AP only, that is, (MultiHop- Direct)/Direct in percent. Right: CDF of user throughput [Gb/s], 1 × 1, 2 × 2 and 4 × 4 MIMO.
Self-interference cancelation 80 dB at R-AP.
tinent for URC. Coordinated RRM over the two to the required reliability, the gains from routing
hops is compared to uncoordinated RRM, based diversity against shadow fading persist even with
on precoder and power control. For HD trans- MIMO transmissions. This lays the foundation
mission, time domain resource optimization is for the other gain mechanisms.
performed. Direct link IC is used based on a
time-reversal strategy, with a frame consisting of CONCLUSION
15 subframes.
The statistics in the simulations are collected We consider in-band backhauling for 5G small-
over 1000 network instances. In the simulations, cell networks. A flexible frame format enables
cross-interference between the cells is taken into effective use of time and frequency resources for
account. Detailed system simulation parameters uplink, downlink, and backhaul transmissions.
are shown in Table 1. Relay nodes are densely deployed in the net-
work to provide routing diversity for two-hop
PERFORMANCE RESULTS transmissions. A two-hop FD self-backhauling
Throughput performance in the MBB scenario flow has two possible bottlenecks, either self-
can be found in Fig. 4. The cumulative distribu- interference in the FD relay or interference
tion function (CDF) of user throughput is depict- from the direct transmission to the destination.
ed. In addition, relative gains of different The dominant bottleneck depends on the link
strategies over direct transmission are given at distances and the efficiency of self-interference
the 5 percent point of the user throughput. This cancellation at the FD relay. Coordinated RRM
reflects the service offered to cell edge users. between the self-backhauling and access hops
With 80 dB SI cancelation, we see a clear proves to be efficient. The drawback of the dis-
gain from coordinated RRM with FD. If RRM cussed solution is increased infrastructure costs
is enabled to choose between HD and FD, some due to the numerous relay nodes. The price of
additional gains would be realized. Direct link these can, however, be kept low by restricting
cancellation/combination provides little gain with the relays to operate as forwarders of coded bit
coordinated RRM. However, with hypothetical transmissions without L2 functionalities. We
100 dB SI cancellation, when the direct link conclude that FD self-backhauling is a valuable
interference is the bottleneck, direct link IC pro- technology component when striving for the
vides noticeable gains. Sequential direct link gigabit experience in MBB and ultra-reliability
combination provides worse performance and is in URC.
not depicted.
Outage capacity performance in the URC sce- ACKNOWLEDGMENT
nario can be found in Fig. 5. The CDF of the The authors would like to thank Mr. Karol
rate a receiver can enjoy with 99.9 percent proba- Schober and Mr. Ilkka Harjula for participating
bility at a given location is depicted. In addition, in the development of the simulator.
relative gains over direct transmission are given
at the 5 percent point of the user outage capacity REFERENCES
CDF. These represent the gains in the rate that [1] Nokia, “FutureWorks Looking ahead to 5G: Building a
can be offered to 95 percent of the locations in Virtual Zero Latency Gigabit Experience,” white paper;
the factory, with 99.9 percent reliability against http://networks.nokia.com/sites/default/files/docu-
ment/5g_white_paper_0.pdf.
changing fading conditions. With 100 dB SI can- [2] Cisco, “Cisco Visual Networking Index: Global Mobile
cellation, remarkable gains from relaying, FD Data Traffic Forecast Update, 2011–2016,” white
operation, and direct link IC are observed. Due paper, 2012.
1200 1
HD
0.9
1000 FD 80dB SI
uncoordinated 0.8
0.7
800 FD 80dB SI
coordinated 0.6
Direct
CDF
600 0.5 HD
FD 100dB SI FD 80dB SI
uncoordinated uncoordinated
0.4
FD 80dB SI
400 coordinated
FD 100dB SI 0.3
coordinated FD 100dB SI
uncoordinated
0.2 FD 100dB SI
200 coordinated
FD 100dB SI
uncoordinated, 0.1 FD 100dB SI
direct link IC uncoord, DL IC
0 0
0 1 2 3 4 5
1x1 MIMO 2x2 MIMO 4x4 MIMO
Outage capacity [10 Mbps]
Figure 5. URC scenario: 0.1 percent Outage Capacity. Left: Outage capacity at 5 percent of CDF, gain [%] relative to direct com-
munication with F-AP only. Right: CDF of outage capacity [10 Mb/s], 2 × 2 MIMO.
[3] P. Mogensen et al., “Centimeter-Wave Concept for 5G D.Sc. in 2009 and 2014 from Aalto University, Finland, all
Ultra-Dense Small Cells,” IEEE VTC-Spring 2014, in electrical engineering. In 2014, he was a consultant for
MWC2020 workshop, May, 2014. Nokia Networks. He is currently a senior researcher at
[4] E. Tiirola et al., “ On the Design of Discovery Patterns Huawei Technologies Sweden. His research interest includes
for Half-Duplex TDD Nodes Operating in Frame-Based multi-antenna techniques and transceiver design.
Systems,” Future Network and Mobile Summit, 2013,
July 2013, pp. 1–9. OLAV TIRKKONEN (olav.tirkkonen@aalto.fi) received his M.Sc.
[5] H. Holma, and A. Toskala, Eds., LTE for UMTS: OFDMA and D.Sc. (Tech) degrees in theoretical physics from Helsin-
and SC-FDMA Based Radio Access, Wiley, 2009. ki University of Technology, Finland, in 1990 and 1994,
[6] P. Popovski, “Ultra-Reliable Communication in 5G Wire- respectively. Currently he is an associate professor in com-
less Systems,” Proc. 5GU Conf., 6 pp., Nov. 2014. munication theory at the Department of Communications
[7] G. Gaderer, P. Loschmidt, and A. Mahmood, “A Novel and Networking in Aalto University, Finland. His research
Approach for Flexible Wireless Automation in Real-Time interests are in coding theory, multi-antenna techniques, as
Environments,” IEEE Int’l. Wksp, Factory Commun. Sys- well as cognitive and heterogeneous cellular systems.
tems, 2008, pp. 81–84.
[8] H. Ju et al., “Full Duplexity in Beamforming-Based RISTO WICHMAN (risto.wichman@aalto.fi) received his M.Sc.
Multi-Hop Relay Networks,” IEEE JSAC, vol. 30, no. 8, and D.Sc. (Tech) degrees in digital signal processing from
Sept. 2012, pp. 1554–65. Tampere University of Technology, Finland, in 1990 and
[9] I. Harjula et al., “ Full Duplex Relaying for Local Area,” 1995, respectively. Currently he is a professor in the
Proc. IEEE PIMRC 2013. Department of Signal Processing and Acoustics, Aalto Uni-
[10] S. Venkatasubramanian et al., “Impact of Neutralization versity. His research interests include signal processing
on Isolation in Co-Planar and Back-to-Back Antennas,” techniques for wireless communication systems.
Euro. Conf. Antennas and Propagation, Apr. 2015.
[11] A. Sabharwal et al., “In-Band full-Duplex Wireless: K ARI P AJUKOSKI (kari.pajukoski@nokia.com) received his
Challenges and Opportunities,” IEEE JSAC, vol. 32, no. B.S.E.E. degree from Oulu University of Applied Sciences in
9, Oct. 2014, pp. 1637–52. 1992. He is a senior research specialist with Nokia Net-
[12] D. Bharadia, E. McMilin, and S. Katti, “Full-Duplex works, Oulu, Finland, and has broad experience in cellular
Radios,” ACM SIGCOMM, Aug. 2013. standardization, link and system simulation, and algorithm
[13] D. W. K. Ng, E. S. Lo, and R. Schober, “Dynamic development for products. Currently he leads a team
Resource Allocation in MIMO-OFDMA Systems with focusing on key physical layer technologies for 5G.
Full-Duplex and Hybrid Relaying,” IEEE Trans. Commun.,
vol. 60, no. 5, May 2012, pp.1291–1304. EEVA LÄHETKANGAS (eeva.lahetkangas@nokia.com) received
[14] T. Riihonen, S. Werner, and R. Wichman, “Hybrid Full- her M.Sc. degree in communication engineering in 2005
Duplex/Half-Duplex Relaying with Transmit Power Adap- from the University of Oulu. She has been with Nokia Net-
tation,” IEEE Trans. Wireless Commun., vol 10, no. 9, works since 2005, where she first concentrated on digital
Sept. 2011. signal processing for WCDMA, WiMAX, and LTE. Currently
[15] P. Kyösti et al., “Winner II Channel Models,” tech. rep. she is concentrating on 5G physical layer research.
D1.1.2 V1.2, 2007, http://www.ist-winner.org.
ESA TIIROLA (esa.tiirola@nokia.com) received his M.S.E.E. in
1998 from the University of Oulu. He is currently with
BIOGRAPHIES Nokia Networks, where he is working on various topics
R ENAUD -A LEXANDRE P ITAVAL (renaud.alexandre.pitaval@ related to radio research and standardization. His current
alumni.aalto.fi) received an M.Sc. in 2008 from the Greno- research interests include transceiver algorithms, multi-
ble Institute of Technology, France, and an M.Sc. and a antenna techniques, and cellular system design.