5g challenges

IEEE Communications Magazine • May 201436 0163-6804/14/$25.00 © 2014 IEEE

Shanzhi Chen and JianZhao are with DatangTelecom Technology &Industry Group.

INTRODUCTION

International Mobile Telecommunications(IMT)-Advanced Specifications of Fourth Gen-eration (4G) Terrestrial Mobile Telecommunica-tion were approved by the InternationalTelecommunication Union Radio Standards Sec-tor (ITU-R) in January 2012. Meanwhile, thedramatic growth of mobile data services drivenby wireless Internet and smart devices has trig-gered the investigation of 5G for the next gener-ation of terrestrial mobile telecommunications.

Mobile traffic requirements have shown dif-ferent features that introduce significant impacton future mobile system architectures, technolo-gy developments, and evolution. These featuresinclude:Big traffic volume: increases on the order of

several magnitudes. The compound yearlygrowth rate for the period 2012–2016 is 78percent [1]. Now the industry is preparingfor an astounding 1000-fold of data trafficincrease for 2020 and beyond.

Increased indoor or hotspot traffic: dominatesmobile traffic. Currently, 60 percent voicetraffic and 70 percent data traffic happensindoors; in the future, indoor/hotspot trafficmay approach 90 percent.

Higher traffic data asymmetry: The ratio ofdownlink to uplink traffic was around 6:1 in2010, and the ratio could rise to 10:1 overthe next five years [2] as the proportion ofvideo traffic in the total mobile trafficgrows.

Huge numbers of subscribers: will be createdfrom machine-to-machine (M2M) applica-tions, which are estimated at 100-fold ormore of mobile phone subscribers.

Energy consumption: for future mobile net-works may need to be reduced on the orderof several magnitudes.Future mobile networks will face great chal-

lenges, including higher capacity, higher perfor-mance, lower power consumption, higherspectrum efficiency, more spectrum resource,and lower cost.

The legacy architecture of terrestrial mobilesystems will hardly meet those requirements andchallenges in 5G. We predict that a step changewill eventually happen in 5G.

STEP CHANGE: FROM MACRODOMINATED EVOLUTION TO

MACRO-LOCAL COEXISTING ANDCOORDINATING EVOLUTION

A concept of nomadic/local area wireless accesshas been mentioned early in the IMT frameworkmodel in ITU-R. But in the actual 3G/4G sys-tem, the system designs are mainly based on theperspectives and requirements of the macrocell.The first reason is that the main goal of 3G/4Gis to achieve consistent coverage for the sameservices in both outdoor and indoor scenarios.The second reason is to keep backward compati-bility for devices both outdoors and indoors.Thus, those designs based on the macro scenarioare not always good for the indoor scenario. In3G, those considerations were reasonable. But in5G, where conditions will be completely differ-ent, the services are layered and inconsistent.The voice services and moderate data servicesare always consistent in the whole system, butmost high-data-rate services are always isolated

ABSTRACT

In this article, we summarize the 5G mobilecommunication requirements and challenges.First, essential requirements for 5G are pointedout, including higher traffic volume, indoor orhotspot traffic, and spectrum, energy, and costefficiency. Along with these changes of require-ments, we present a potential step change forthe evolution toward 5G, which shows thatmacro-local coexisting and coordinating pathswill replace one macro-dominated path as in 4Gand before. We hereafter discuss emerging tech-nologies for 5G within international mobiletelecommunications. Challenges and directionsin hardware, including integrated circuits andpassive components, are also discussed. Finally,a whole picture for the evolution to 5G is pre-dicted and presented.

5G WIRELESS COMMUNICATIONS SYSTEMS:PROSPECTS AND CHALLENGES

Shanzhi Chen and Jian Zhao

The Requirements, Challenges, andTechnologies for 5G of Terrestrial Mobile Telecommunication

ZHAO_LAYOUT_Layout 5/7/14 11:58 AM Page 36

IEEE Communications Magazine • May 2014 37

and lie in only indoor or hotspots. Furthermore,5G will be a heterogonous framework, and back-ward compatibility will not be mandatory bothindoors and outdoors. The improvement of userequipment (UE) provides the ability to supportsimultaneous connections both indoors and out-doors. The previous barriers will be broken.

Thus, a fundamental change will occur in 5G.It is predicted that there will be two coexistingand coordinating evolution paths of a macro-based IMT path and a local-based IMT pathinstead of one macro-dominated path as in 4Gand before. A general profile for the change isshown in Fig. 1, and the reasons are analyzedbased on the following key aspects.

REQUIREMENT ON HEAVY DISTRIBUTION OFMOBILE DATA SERVICES

In 5G, high data rate services indoors and athotspots occupy the main traffic volume ofmobile services. The previous macro-dominateddesign will not be suitable to meet this change.As a result, a new design, local-based IMT, opti-mized for both indoors and hotspots, will be nec-essary. Local-based IMT needs to coexist andcoordinate with macro-based IMT.

STUNNING MOBILE DATA CAPACITY(1000-FOLD) REQUIREMENT

The increasing of cell numbers will be a veryefficient way to improve the system capacity inan almost linear ratio [3] when the signal-to-interference-plus-noise ratio (SINR) is guaran-teed. However, it is impossible to increase thenumber of the current small cells by orders ofmagnitude because of the limitations from cur-rent macro-based design such as backward com-patibility, cost, interference, cell management,and cell sites. But in 5G, local-based IMT can beintroduced, and small cells can be designed fromlocal perspectives. It is completely possible toincrease the number of small cells by orders ofmagnitude when new architecture and new tech-nologies are adopted. According to [3], theincrease of small cell numbers in the future mayfar exceed 100 times the current number.

ENERGY CONSUMPTION REQUIREMENTEnergy consumption shall be decreased signifi-cantly in orders of magnitudes in 5G. But it isimpossible in the current framework and currentbase stations [4]. For example, the goal of ECFP7 EARTH is to decrease energy consumptionfor base stations by 50 percent, mainly based inthe current framework. However, it is possible toreduce power down to orders in magnitude byoffloading majority data to local small cells, let-ting inactive cells sleep, in the macro-local coex-isting and coordinating framework.

NEW AVAILABLE SPECTRUM REQUIREMENTFor 2G/3G and even 4G, most allocated spec-trums are mainly below 3 GHz. In addition, thespectrum usage reaches almost the maximumefficiency. A new work item was thus establishedto seek new potential spectrum for the futureterrestrial mobile communication. It seems thatthe new available spectrum for the future gener-

ation is mainly above 3GHz or even higher.Those spectrums will be suitable for local sce-narios to improve capacity rather than for amacro scenario to improve coverage. This pro-vides the fundamental condition to the newdesign of local-based IMT.

SIGNIFICANTLY RELAX THEREQUIREMENT ON COST

Since the RF and baseband requirements on thecurrent terminal chipset are based on the physi-cal parameters meeting high mobility and largecell radius, the cost is hard to be significantlyreduced. When local-based IMT is introduced,the cost of the RF and baseband chipsets for thelocal scenario can be decreased because the RFand baseband performance requirements arerelaxed to meet low mobility and small cellradius. Moreover, the cost of a small cell, includ-ing the cost of maintenance and a cell site, willalso be significantly reduced.

From 2G to 3G and to 4G, the design princi-ple is macro-based for the design of either theprimary physical parameters or the framework.In order to change the situation to some extent, awork item for small cells, probably using a newtype of carrier, is introduced in 3GPP Release12. However, great changes will be very difficultin 4G due to the limitations in the frameworkand backward compatibility. This work item wasterminated at the RAN #61 meeting in Septem-ber 2013. That is why only a small branch hasspun off from the main macro-dominated branchin Fig. 1.

Together with the requirements for 5G, thisstep change will bring in some technical trends,described in the following sections.

Figure 1. Evolution change for 5G.

CDMA

TDMA

Macro-dominated evolution before 5G

OFDM

2G 3G 4G 5G

IMT/IMT-A macro-basedevolution

New type carrier forlocal area in rel 12and beyond

CDMATDMA

Macro-local-coordinated evolution in 5G

OFDM

2G 3G 4G 5G

IMT/IMT-A macro-based evolution

IMT/IMT-A local-based evolution


IEEE Communications Magazine • May 201438

POSSIBLE TECHNOLOGIES IN 5GIn this section, several technical trends are dis-cussed for 5G mainly based on IMT, includinglocal IMT small cells, heterogonous layer coordi-nation, flexible separation and combination ofthe C/U plane, technologies for asymmetric traf-fic, hybrid networking topology, and signal pro-cessing technologies.

LOCAL IMT SMALL CELLSIt is believed that small cells will play a veryimportant role in 5G to meet the 5G require-ments in traffic volume, frequency efficiency,and energy and cost reduction. It is predictedthat small cells will be designed differently from4G to 5G. Local IMT small cells is a branch ofIMT local-based evolution. Local IMT smallcells will be designed based on the indoorsand/or hotspot scenario. There are significantdifferences between the indoors or hotspot sce-nario and the macro scenario. First, the indooror hotspot scenario supports only nomadic orpedestrian speeds (e.g., 0–30 km/h) and smallcell radius (e.g., 150 m) compared with highmobility (e.g., 350 km/h or higher) and large cellradius (e.g., > 30 km) in macro scenarios. Thesedifferences will impact the principal parameterdesign. Second, the propagation characteristicsof the indoor or hotspot scenario are quite dif-ferent from those of the macro scenario. Thoselead to different designs for multi-antenna tech-nology and bring more performance benefits tolocal IMT small cells. Third, the transmissionpower of local IMT small cells will be orders ofmagnitude less than that of today’s macrocells.This provides more flexible spectrum usage andenergy savings.

Design of the Principal Physical LayerParameters — Since the target cell sizes andmaximum speeds are significantly differentbetween the macro and the indoor or hotspotscenarios, the principal physical layer parame-ters are supposed to be quite different for thosetwo kinds of scenarios. However, in currentmobile systems, those principal physical layerparameters are mainly determined according tothe characteristics in macrocells. For example,in Third Generation Partnership Project(3GPP) Long Term Evolution (LTE) systems,the principal physical layer parameters are thecyclic prefix (CP) length and the subcarrierspacing Df. Those two parameters are deter-mined to support mobility conditions (up to 350or even 500 km/h) and cell sizes of tens of kilo-meters [5]. In the indoor or hotspot scenario,the CP length can be decreased by severaltimes, and subcarrier spacing Df can increaseseveral times. This optimization can not onlyimprove the performance, but also reduce thecost of RF and baseband integrated circuits(ICs) of terminals and small cells by relaxingthe RF requirements.

Design for Multiple-Input Multiple-Output— MIMO technology is one of the most impor-tant technologies to significantly improve sys-tem performance in coverage, capacity, anduser data rates. The performance improvement

of MIMO depends on propagation characteris-tics of each deployment scenario. One factorimpacting MIMO performance and design isthe distance between antenna elements. Thisdistance actually impacts the mutual correlationbetween the radio-channel fading on signals ondifferent antennas. But the relationshipbetween the element distance and the mutualcorrelation in the indoor or hotspot scenario issignificantly different from that in the macroscenario. In the indoor or hotspot scenario, theenvironment as seen from the base station ismore similar to the environment as seen fromthe terminal [6]. In this case, smaller antennadistance in a base station, such as half a wave-length (0.5l ), not the 10 times wavelength(10l) usually in macro scenarios, is typicallysufficient to ensure relatively low mutual corre-lation between antennas [7]. Examples areshown in Fig. 2. This will bring a large benefitto small cells in that a small cell can supportmore antennas (e.g., 8 or even 16 antennas)based on its small size in Fig. 2a. The systemcapacity can thus be significantly improved.Moreover, the ability to supply both MIMOand beamforming technologies simultaneouslyto different UEs by one small cell can be pro-vided according to different usage in Fig. 2b. Inaddition, in the indoor or hotspot scenario,there is a richer scatter environment than thatin the macro scenario. Excessive scatteringenvironments are a challenge. Multi-userMIMO (MU-MIMO) is a perfect solution ofthe challenge in downlink and uplink, as shownin Figs. 2c and 2d. MU-MIMO can improvesystem capacity; it is especially valuable to ter-minals by combining space multiplexing usingantennas from different UEs since each UEalways has few antennas, limited by size andcost. A key factor having an impact on the per-formance of MU-MIMO is the fast and accu-rate feedback of channel information. Thefocus of the macro scenario is on supportinghigh UE mobility, so it is very difficult to quick-ly get accurate channel information when theUE is in high mobility in a macrocell, and thusimprovements of performance is l imited.Indoors or at a hotspot, low mobility isrequired. It is therefore possible to obtain time-ly and accurate feedback of channel informa-tion so that a more dedicated method can bedesigned. In time-division duplex (TDD), chan-nel reciprocity in the uplink and downlink willbe one of the advantages, offering availabilityof fast and accurate feedback of channel infor-mation. In general, local IMT small cells willpotentially have a large role in improvingMIMO design.

Energy Saving Design — As shown in Fig. 3a,power efficiency is low in current macro basestations due to the site support system consump-tion and transmission loss [8]. But small sells oflocal IMT can have different system architectureand product types. Local IMT small cells do notneed cooling systems. Excessive transmission losscan be avoided thanks to the shorter transmis-sion range. An intuitive view might be excessivepower consumed by hundreds of small cellsdeployed to replace one macrocell. However,

MIMO technology is

one of the most

important technolo-

gies to significantly

improve system

performance in

coverage, capacity

and user data rates.

The performance

improvement of

MIMO depends on

propagation charac-

teristics of each

deployment scenario.

ZHAO_LAYOUT_Layout 5/7/14 11:58 AM 8


inactive small cells will actually sleep most of thetime and consume much less active power. Forexample, as shown in Fig. 3b, small cells inoffices may turn off, and small cells in hotspotsmay turn on at night, while small cells in officesmay turn on and small cells in hotspots may turnoff during the day. The power consumption of alocal IMT small cell can be reduced by severalorders of magnitude from that in a current basestation. In 5G, the mobile broadband traffic canbe balanced and offloaded by high-volume localIMT small cells. Thus, the energy efficiency ofthe whole wireless access system will be dramati-cally increased.

Design for Flexible Spectrum Usage — Itis envisioned that a great number of small cellswill be deployed in 5G networks. However, therequirements on spectrum resources cannot befulfi l led by simply increasing spectrum. Apromising technology to deal with this issue iscognitive radio (CR). Although CR offers flexi-bility in spectrum usage, the flexibility seemsdifficult to use in a macro base station. Since asmall cell has low transmission power andsmall size, the coverage zone providing veryhigh data rate services can be very isolated,and in turn CR can be more easily used insmall cells than in macrocells. Moreover, newsharing of spectrum bands might be allocatedfor flexible spectrum usage, especially to smallcells. The FCC has considered implementingshared use of the 3550–3650 MHz band[9].Because of using the same single channelfor uplink and downlink operation, only onesignal band need be sensed, TDD has advan-tages over frequency-division duplex (FDD)

since FDD requires a traditional design ofpaired spectrum using two fixed channels withfixed spacing between them [9]. Also, there is aTDD drawback, which is the need for extraguard bands when small cells are deployed fordifferent operators in TDD mode.

In addition, local IMT small cells have moreopportunities for design optimization for wire-less backhaul and self-organized networking(SON) by which more flexibility and cost reduc-tion can be achieved. Furthermore, ultra-widebandwidth in very high frequency (e.g., 6 ~ 60GHz) might be a good choice to provide hugecapacity improvement to small cells in thefuture.

HETEROGONOUS LAYER COORDINATIONThe future 5G network will be a heterogonouslayer network consisting of macrocells, tradition-al micro/picocells, new Local IMT small cells,and relay and other low-power nodes. Accordingto whether the same or different frequency isused in different layers, two deployment modelsare defined as co-channel frequency deploymentand dedicated frequency deployment. Co-chan-nel frequency deployment uses the same fre-quency in different layers to deal with radiocoverage, while dedicated frequency deploymentuses different frequencies in different layers toprovide radio capacity. Since the capacity of highrate traffic is the main focus in 5G, dedicatedfrequency deployment will be the main methodin 5G.

In 5G, the two-path design for macro andlocal cells will gain many benefits. It also bringsobvious technical challenges in the coordina-tion between macrocell and local cells, espe-

Figure 2. More antenna elements and more advanced antenna technologies for local IMT small cells.

MIMO

MIMO

BF

2+2 streams

(a) (b)

(c) (d)

2+2 streams

In 5G, the two-path

design for macro

and local cells will

gain many benefits.

It also brings obvious

technical challenges

in the coordination

between macrocell

and local cells, espe-

cially when one

macro eNB may

manage tens or even

hundreds of small

cells in the future.



cially when one macro eNB may manage tensor even hundreds of small cells in the future. Itis predicted that multiple coordination meth-ods, including coordination in the core net-work, radio resource and mobility management,layer 2, and the physical layer and RF, will beintroduced and/or enhanced. Thus, more bene-fits from the heterogonous layer coordinationcan be achieved, not only in system perfor-mance, but also in energy saving and serviceprovisioning.

FLEXIBLE SEPARATION ANDCOORDINATION OF C/U PLANES

To support coordination between macrocellsand local cells, a feature of flexible separationand coordination of C/U planes will be intro-duced and enhanced. In 5G, the control plane(C plane) and user plane (U plane) for oneUE can be connected to multiple eNBs orsmall cells by separation and combination ofC/U. A simple example is given in Fig. 4. Inthis case, signaling radio bearers (SRBs) in theC plane and low-rate data service data radiobearer 0 (DRB0) are allocated to the macroeNB in frequency band f1, while DRB1 forhigh data rate service is allocated to small cellin frequency band f2. In this architecture, themain data traffic is offloaded from the eNB tosmall cells in a dedicated frequency band. Amuch higher rate can be achieved locally, so itstransmission power can be significantlydecreased, dynamic flexible frequency usagecan be supported, equivalent signaling loadfrom the current system to the core network iskept, and better service provisioning is given tocustomers. The benefits are due to decreasingunnecessary handover when UE is moving fastamong the overlap areas covered by the macro-cell and small cells . Furthermore, moreadvanced coordination technology such asadvanced coordinated multipoint (CoMP) tech-nology can be developed based on this archi-tecture. However, there are challenges thatmust be carefully studied: dual receivers maybe needed on the UE side, UE cannot workwithout macro coverage, and legacy UE cannotaccess new 5G systems.

TECHNOLOGIES TOHANDLE ASYMMETRIC TRAFFIC

Another big challenge is handling highly asym-metric traffic in mobile communication systems;thus, technologies to handle asymmetric trafficneed to be enhanced in 5G.

In general, because of the traditional andequally paired frequency usage in downlink anduplink, FDD has more difficulties dealing withasymmetry issues. There are some technicalsolutions, and those solutions will be improvedfurther in 5G.

For FDD, asymmetrical FDD carrier aggre-gation (CA) and supplemental downlink only(SDL) have been taken into consideration. Theprinciple is to aggregate more frequency bandsin downlink than in uplink to cater for asymmet-ric traffic in downlink and uplink. Limited by thecurrent UE capability and spectrum resource,for example, FDD supports the combination offour blocks, one of three with the band size of 5up to 20 MHz in the downlink, and one withband size of 5 up to 20 MHz in the uplink. In5G, it is predicted that more blocks and largerblock sizes will be supported. The first challengeis the limitation of frequency allocation. Forasymmetrical FDD CA, the total frequencybandwidth equality allocated for downlink anduplink cannot be changed. For SDL, the ITUregulation needs to be changed by introducing anew SDL allocation mode. The second challengeis that two receivers will be needed, which willunavoidably increase terminal cost and powerconsumption.

In contrast, TDD has some natural advan-tages in supporting asymmetry traffic due to itsunpaired usage of the downlink and uplink.Thanks to the possibility of semi-dynamic slotallocation in uplink and downlink, TDD systemsare able to decide an asymmetric slot configura-tion in uplink and downlink in a wide areaaccording to the estimated traffic. Similar toFDD, TDD can also use carrier aggregation,especially high-low frequency TDD CA to sup-port asymmetrical traffic in uplink and downlink.Furthermore, for TDD, dynamic slot allocationin downlink and uplink can be used in a smallcell. It can provide dynamic and flexible adapta-

Figure 3. Energy savings from local IMT small cells.

Largely energy lostSite supportsystem

Rectifier

Backupsystem

InsideSmall cell

(a) (b)

Macrocell

Hotspot (on/off)

Office (off/on)

Night

Day

SoC

In general, because

of the traditional and

equally paired fre-

quency usage in

downlink and uplink,

FDD has more

difficulties dealing

with asymmetry

issues. There are

some technical

solutions, and those

solutions will be

improved further in

5G.


tion to traffic. TDD solutions can use the natureof TDD to meet asymmetrical traffic in bothwide areas and local areas/hotspots.

HYBRID TOPOLOGY NETWORKINGThe basic topology of current mobile systems isthe star structure with control points. One advan-tage is simple, and the central control providesguaranteed quality of service (QoS).

As for 5G, this topology will be difficult touse. First, the robustness of its access network isinsufficient. Efficient offload is limited since alluser data needs to go through the core network.The performance and energy efficiency of UEon the cell edge will not yet be efficient. Thistopology is also not efficient to meet the require-ment for social networking services. It will alsobe difficult to meet the tremendous accesses ofmachine-to-machine (M2M) users with thistopology. Thus, the cellular topology needs to beoptimized, and relay and device-to-device (D2D)communications will play important roles.

In 5G, as shown in Fig. 5, technologiesinclude small cells, D2D, different kinds ofrelays, and wireless backhaul. Thanks to theintroduction of these technologies into the cur-rent central controlled star network, differentkinds of connections (UE to UE, small cell tosmall cell, small cell to eNB, relay node to eNB,relay node to relay node, etc.) can therefore beestablished. Multihop connections and mobilerelay are also possible. Thus, the hybrid topologyradio network will naturally be the future mobileaccess network, which can help to efficientlyovercome the difficulties and challenges listedabove.

D2D Communications — D2D is considereda promising technology for future mobile sys-tems. In 5G, cellular D2D technology can beintegrated into the cellular network as a sup-

plemental part of the system. Based on howfrequencies are used, D2D can be divided intotwo types: co-channel frequency D2D and ded-icated frequency D2D. In co-channel frequen-cy D2D, D2D works in the same frequency asthat between UE and cell. Its advantage is thereuse of the legacy receiver/transmitter with acertain impact on the current protocol archi-tecture. That is why this method is currentlydiscussed [10–12]. It will bring serious interfer-ence to the cellular system and will be difficultto control when taking UE mobil ity intoaccount. Moreover, most D2D applications orservices need large bands to support high datarates, which is hard to obtain since the band-width between UE and eNB is not wideenough.

On the other hand, dedicated frequencyD2D owns a specific frequency which is differ-ent from that between UE and cell. Since sep-arate frequencies are used in D2D and cellularservice, D2D causes less interference to thecellular system. Meanwhile, D2D can use muchwider bands at higher frequencies to supporthigh data rate D2D services. Since D2D in acellular system is always network controlled ornetwork assisted, two connections need to besupported simultaneously in UE. One connec-tion is for D2D service, and the other is forthe cellular cell. This support is very difficultto achieve in the current cellular system. How-ever, it is not difficult in 5G when new archi-tecture and technologies, as described inprevious sections, are introduced. Thus, it ispredicted that dedicated frequency D2D willbe the main type of D2D, while co-channelfrequency D2D is also kept to support D2Dbetween legacy UE.

Duplex mode is another important feature inD2D design. For FDD, two receivers and twoantennas are needed. The Tx antenna and Rx


Figure 4. Example of flexible separation and combination of the C and U planes.

SRB0

CCCH

BCH PCH

PBCH PDCCH

C-Plane C-Plane U-Plane

SRB1

DCCH

DRB0

DTCH

DRB0

DTCH

DRB1

DTCH

DL-SCH

PDSCH

e-NB

SRB0

CCCH

PRACH

RACH

SRB1

DCCH

UL-SCH

PUSCH

UE

U-Plane

DRB1

DTCH

PBCH PDCCH

[BCH][PCH] DL-SCH

PDSCH

Small cell

Small cell

TrafficSignaling

In 5G, cellular D2D

technology can be

integrated into the

cellular network as a

supplemental part of

the system. Based on

how frequencies are

used, D2D can be

divided into two

types: co-channel

frequency D2D and

dedicated

frequency D2D.



antenna should have enough isolation distance.It is always difficult to implement into UE withsmall physical size. TDD has the natural advan-tage that the uplink and downlink transmissionand receiving are in different time slots in thesame frequency band, which is why many D2Dresearchers are currently focused on TDD mode[10–12]. It is predicted that in 5G for both TDDand FDD, D2D shall be supported, but D2D inTDD mode will be the main type with the con-siderations of lower cost, lower complexity, andthe possible benefit of better frequency availabil-ity.

Relay Technologies — Relay technology isalso promising in 5G. In 5G, relay can havemany more functions in a cellular system. Asshown in Fig. 5, advanced cellular relay canimprove the topology of the cellular system,improve the robustness of a network, anddecrease power consumption. A multihop struc-ture can efficiently support tremendous access ofM2M terminals. Mobile relay can deal withmobile group access. A UE-based mode canconstruct new M2M services by letting UE be acore access node in Internet of Things (IoT) net-works.

Similar to cellular integrated D2D, relays canbe divided into in-band relay and out-of-bandrelay. In-band relay is similar to co-channel fre-quency D2D without using extra frequencyresources, and sharing similar advantages anddisadvantages; out-of-band relay is similar todedicated frequency D2D. 5G advanced cellularrelay can be supported by both FDD and TDD.TDD will have some advantage over FDD by itsnature with the uplink and downlink in differenttime slots using the same frequency band. Relayin TDD mode might be the main type in 5Gwith enough consideration of the lower cost,lower complexity, and possible benefit of betterfrequency availability.

SIGNAL PROCESSING ANDIC PROGRESS DRIVING TECHNOLOGIES

The progress in signal processing and IC alwaysdrive technology development. In mobile com-munication, the adoption of an advanced tech-nology always depends on the status of signalprocessing IC including CPU, digital signal pro-cessing (DSP), analog-to-digital (A/D), RF, andpassive components including antennas and oth-ers. Although a bit saturated, Moore’s law stillprovides us twice the computing performanceand scale of IC every 18 months. It will drivetechnical progress in 5G just as it has for 2G,3G, and 4G. There are several significant indica-tions.

First, the macrocell will continue to be thekey provider of capacity and coverage in 5G,although small cells will be widely deployed. Asfor macrocells, one promising technology isactive antenna array technology in which multi-ple RF components (power amplifiers andtheir passive components) will be integrateddirectly into the antenna’s radiator elementswith flexible local control. Significant benefitscan be achieved, including capacity, coverage,flexible frequency usage, and heterogonouslayer coordinating. It already seems feasible toimplement active antenna array in 4G, and itwill be easier to implement the technology ona large scale in 5G due to further RF compo-nent integration.

Second, 5G UE needs ultra high computingperformance to support data and signal process-ing for much larger bandwidth such as 40 MHz,100 MHz, or higher, so the standby timebecomes a key factor. Also, more advanced ICand passive component technologies shall beused in 5G. The technologies that have to beused are configurable low-power software-defined baseband chipsets, wideband RF receiv-er and transmitter ICs, digital RF chipsets,adaptive-powered RF power amplifiers, digitalRF power amplifiers, multi-mode integrated pas-sive components, and micro electro-mechanicalsystem (MEMS) switches with better isolationand lower insertion loss.

Finally, all those chipset modules will be inte-grated into a single chipset. To support thosetechnologies, more advanced semiconductordevice and IC technologies will be developed;for example, 14 nm or better complementarymetal oxide semiconductor (CMOS) technolo-gies may be the driver of 5G to meet the require-ments in both performance and powerconsumption. By introducing more spectrumbands, passive components will dominate thetotal cost and set up a new challenge. Its require-ments and the cost of TDD will be relativelylower thanks to the nature of TDD.

CONCLUDING REMARKS ANDFUTURE TRENDS

In summary, future traffic development bringsnew requirements and challenges to futuremobile broadband systems, including highertraffic volume, indoor or hotspot traffic, trafficasymmetry, and spectrum, energy, and cost effi-

Figure 5. Hybrid topology networking in 5G.

(g) Mobile relay

(f) Multi-hop relay

eNBeNBeNB

12

3

4

s1s1

x2

RRU

AP

(e) CoMP

Small cell

(a) UE-based relay

(d) Nodes forInternet thing

(b) D2D

(c)Connectionsto multi eNB

x2

s1

EPC

Operator IPservices Internet



ciency. Technologies are thus needed, includinglocal IMT small cells with redesign of the physi-cal layer, MIMO, energy saving strategy, andflexible spectrum usage. More technologies,including heterogeneous layer coordination,C/U plane flexible separation-coordination,handling of asymmetrical traffic, and hybridtopology networking shall be sufficiently devel-oped. The development of 5G is impacted sig-nificantly by industrial scale and smoothevolution of the current mobile system. More-over, state-of-the-art IC technologies are need-ed to support high-performance computing forsignal processing, RF transceivers, and poweramplifiers.

To conclude the article, a whole profile of 5Gis shown in Fig. 6. Besides the research in theevolution from IMT to 5G mentioned, there aresome new possible revolutionary technologies inthe pre-study phase of 5G including orbitalangular momentum encoding [13], full duplex[14], non-orthogonal waveforms, and more. Webelieve new research will benefit 5G in the longrun, and new revolutionary technologies mightbe used in some specific scenarios integratedwith IMT. Thanks to TDD’s nature and specificcharacteristics [15], it has great potential in thedevelopment of 5G.

ACKNOWLEDGMENTThis work was partly supported by the NationalScience and Technology Major Project (No.2013ZX03001025-001) of China. We would liketo thank Prof. Liu Dake from BIT for his valu-able comments.

REFERENCES[1] Cisco Visual Networking Index: Global Mobile Data Traf-

fic Forecast Update, 2012–2017, Feb. 2013.[2] Cisco Visual Networking Index: Global Mobile Data Traf-

fic Forecast Update, 2010-2015, Cisco, February 2011.[3] http://www.itc23.com/fileadmin/ITC23_files/papers/tuto-

rial3.pdf.[4] L. Xiang et al., “Energy Efficiency Evaluation of Cellular

Networks Based on Spatial Distributions of Traffic Loadand Power Consumption,” IEEE Trans. Wireless Com-mun., vol. 12, no. 3, Mar. 2013, pp. 961–73.

[5] 3GPP Technical Specification 36.211,”Evolved UniversalTerrestrial Radio Access (E-UTRA) Physical Channels andModulation (Release 10),” www.3gpp.org.

[6] E. Dahlman et al., 4G LTE/LTE-Advanced for MobileBroadband.

[7] W. Lee, Mobile Communications Engineering, McGraw-Hill.

[8] E. Calvanese et al.,”Energy Efficient Cellular Networks,”Hermes Partnership Wksp. Future Networks, May 2011.

[9] PCAST, “Enforcement Coordinator Annual Report on Intel-lectual Property Enforcement Cover Title Here Report to thePresident Realizing the Full Potential of Government-HeldSpectrum to Spur Economic Growth,” July 2012.

[10] P. Jänis et al., “Device-to-Device CommunicationUnderlaying Cellular Communications Systems,” Int’l. J.Commun., Network and Sys. Sci., vol. 2, no. 3, June2009, pp. 169–78.

[11] K. Doppler et al., “Device-to-Device Communication asan Underlay to LTE-Advanced Networks,” IEEE Com-mun. Mag., vol. 7, no. 12, Dec. 2009, pp. 42–49.

[12] M. S. Corson et al., “Towards Proximity-Aware Inter-networking,” IEEE Wireless Commun., vol. 17, no. 6,Dec. 2010, pp. 26–33.

[13] F. Tamburini et al.,” Encoding Many Channels in theSame Frequency through Radio Vorticity: First Experi-mental Test,” July 2011

[14] M. Jainy et al.,” Practical, Real-time, Full Duplex Wire-less,” MobiCom ’11, Sept. 19–23, 2011.

[15] S. Chen et al., “Technical Innovations Promoting Stan-dard Evolution: From TD-SCDMA to TD-LTE andBeyond,” IEEE Wireless Commun., vol. 19, no. 2, Feb.2012, pp. 60–66.

BIOGRAPHIESSHANZHI CHEN [SM’04] ([email protected]) receivedhis Ph.D. degree from Beijing University of Posts andTelecommunications (BUPT), China, in 1997. He joinedDatang Telecom Technology & Industry Group in 1994, andhas been serving as CTO since 2008. He was a member ofthe steering expert group on information technology ofthe 863 Hi-Tech Research and Development Plan of Chinafrom1999 to 2011. He is director of the State Key Labora-tory of Wireless Mobile Communications, and a boardmember of Semiconductor Manufacturing InternationalCorporation (SMIC). He has made great contributions toTD-SCDMA 3G industrialization and TD-LTE-Advanced 4Gstandardization. He received the State Science and Technol-ogy Progress Award of China in 2001 and 2012. His cur-rent research interests include wireless mobilecommunication, IoT, and emergency communication.

JIAN ZHAO ([email protected]) received his M.S. degree fromBUPT in 1998. He joined Datang Telecom TechnologyIndustry Group in 2002. From 2003 to 2008, he participat-ed in 3GPP meetings as standards manager of Datang.Since 2007, he has been acting in ITU meetings as a tech-nical manager. Before joining Datang, he worked inHuawei as a software R&D manager. His research interestsinclude mobile communication technology, spectrum, andstandard activities.

Figure 6. Illustration of evolution from 3G to 5G.

LTE-A FDD

3G

4G

5G

Requ

irem

ents

Year202020102000

LTE FDD

WCDMA

TD-SCDMAOthers

Others

TD-LTE

TD-LTE-A

Macroand local

coordinatedIMT

Local IMT small cell,Heterogeneous layer coordinatingIntegrated cellular D2DAdvanced cellular relay

IMT

New revolutionary technologies


http://www.3gpp.org

mailto:[email protected]

mailto:[email protected]

http://www.itc23.com/fileadmin/ITC23_files/papers/tutorial3.pdf