Ultrareliable and Low-Latency Communication Techniques for …dcl.yonsei.ac.kr › wordpress › wp-content › uploads › publi › ... · 2019-01-28 · Kim et al.: Ultrareliable

Ultrareliable andLow-Latency CommunicationTechniques for TactileInternet ServicesThis article introduces novel physical layer solutions for spectrally efficient ultrareliableand low-latency communication techniques.

By KWANG SOON KIM , Senior Member IEEE, DONG KU KIM , Senior Member IEEE,CHAN-BYOUNG CHAE , Senior Member IEEE, SUNGHYUN CHOI , Fellow, IEEE,YOUNG-CHAI KO , Senior Member IEEE, JONGHYUN KIM , Student Member IEEE,YEON-GEUN LIM, Student Member IEEE, MINHO YANG, Student Member IEEE,SUNDO KIM, Student Member IEEE, BYUNGJU LIM, KWANGHOON LEE, AND KYUNG LIN RYU

ABSTRACT | This paper presents novel ultrareliable and

low-latency communication (URLLC) techniques for URLLC ser-

vices, such as Tactile Internet services. Among typical use

cases of URLLC services are teleoperation, immersive virtual

reality, cooperative automated driving, and so on. In such

URLLC services, new kinds of traffic such as haptic information

including kinesthetic information and tactile information need

to be delivered in addition to high-quality video and audio traf-

fic in traditional multimedia services. Furthermore, such a vari-

ety of traffic has various characteristics in terms of packet sizes

and data rates with a variety of requirements of latency and

reliability. Furthermore, some traffic may occur in a sporadic

manner but requires reliable delivery of packets of medium

to large sizes within a low latency, which is not supported by

Manuscript received January 5, 2018; revised June 19, 2018; accepted August24, 2018. Date of publication September 27, 2018; date of current versionJanuary 22, 2019. This work was supported by the Institute for Information &communications Technology Promotion (IITP) Grant funded by the KoreaGovernment (MSIT) (2015-0-00300, Multiple Access Technique with Ultra-LowLatency and High Efficiency for Tactile Internet Services in IoT Environments).(Corresponding author: Kwang Soon Kim.)

K. S. Kim, D. K. Kim, J. Kim, M. Yang, K. Lee, and K. L. Ryu are with theSchool of Electrical and Electronic Engineering, Yonsei University, Seoul 03277,South Korea (e-mail: [email protected]).

C.-B. Chae and Y.-G. Lim are with the School of Integrated Technology, YonseiUniversity, Seoul 03277, South Korea.

S. Choi and S. Kim are with the Department of Electrical and ComputerEngineering, Seoul National University, Seoul, South Korea.

Y.-C. Ko and B. Lim are with the School of Electrical Engineering, KoreaUniversity, Seoul 02841, South Korea.

Digital Object Identifier 10.1109/JPROC.2018.2868995

current state-of-the-art wireless communication systems and

is very challenging for future wireless communication systems.

Thus, to meet such a variety of tight traffic requirements in a

wireless communication system, novel technologies from the

physical layer to the network layer need to be devised. In

this paper, some novel physical layer technologies such as

waveform multiplexing, multiple-access scheme, channel code

design, synchronization, and full-duplex transmission for spec-

trally efficient URLLC are introduced. In addition, a novel per-

formance evaluation approach, which combines a ray-tracing

tool and system-level simulation, is suggested for evaluating

the performance of the proposed schemes. Simulation results

show the feasibility of the proposed schemes providing realistic

URLLC services in realistic geographical environments, which

encourages further efforts to substantiate the proposed work.

KEYWORDS | Full-duplex communications; multiple access

schemes; tactile internet; URLLC; waveform multiplexing.

I. I N T R O D U C T I O N

Owing to continuously increasing demands for new ver-tical services, academia and industry have been placinga huge emphasis on developing the fifth-generation (5G)enabling technologies.1 According to the InternationalMobile Telecommunication (IMT) vision for 2020 [1],

1Readers are invited to visit https://www.dropbox.com/s/8qnjzla9g2lbr4e/Tactile-2017-SLS-Implementation.mp4?dl=0 for a video clipintroducing the proposed work.

376 PROCEEDINGS OF THE IEEE | Vol. 107, No. 2, February 2019

0018-9219 © 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

https://orcid.org/0000-0002-5706-174X

https://orcid.org/0000-0001-9432-2972

https://orcid.org/0000-0002-5226-6463

https://orcid.org/0000-0001-9561-3341

https://orcid.org/0000-0003-0279-445X

https://orcid.org/0000-0003-1043-9028

Kim et al.: Ultrareliable and Low-Latency Communication Techniques for Tactile Internet Services

emerging 5G services include enhanced mobile broadbandservices, massive Internet of Things (IoT), and ultrareliableand low-latency communication (URLLC) services, such asTactile Internet services.2 Among them, URLLC services areconsidered as the most challenging applications in 5G orfuture cellular systems, and their typical use cases includecollaborative automated cars, teleoperations, interpersonalcommunications (ICs), and immersive virtual reality (IVR)services [2]–[8].

Unlike the classical high-quality multimedia streamingservice, in which high-rate information flows from a sourceto a sink, sensing information, control and command infor-mation, and feedback information that occurred by theactuation according to the control and command form aloop for an information flow in typical URLLC services. Ina typical telesurgery example [9], [10], a real-time high-quality video sensing information of the affected area of apatient needs to be delivered to a surgeon and the surgeoncontrols a remote surgical robot, wherein elaborate kines-thetic information of the surgeon’s hands and fingers needsto be delivered to the robot; the force and tactile sensinginformation that occurred from the interaction betweenthe robot and the affected area needs to be fed back to helpthe surgeon along with the video sensing information. Notethat such a typical information flow requires a delivery ofhigh-rate information up to several hundred megabits persecond (Mb/s) with an end-to-end latency as low as 1 msand a reliability as high as 99.9999999% (i.e., packet errorrate of 10−9), which reveals some challenging aspects ofdeveloping URLLC enabling technologies in cellular com-munication systems.

Studies on typical URLLC use cases have been carriedout, in which typical traffic characteristics and quality ofservice (QoS) requirements of some URLLC use cases arereported. In [2], a telesurgery use case is described inwhich audio, video, and haptic information needs to bedelivered within an end-to-end latency as low as 1 ms andwith an extremely high reliability [block error rate (BLER)down to 10−9]. In [3], intelligent transportation exam-ples are described, such as cooperative collision avoidanceand high-density platooning, in which sensing informationneeds to be exchanged within an end-to-end latency as lowas 5 ms and with high reliability [frame error rate (FER)down to 10−6]. Further, in [4]–[6], industry automa-tion examples are described, such as time-critical processoptimization inside factory and remote control, in whichvideo, audio, and haptic information needs to be deliveredwithin a submillisecond end-to-end latency and with anextremely high reliability (BLER down to 10−9). Recently,the IEEE standardization activity on Tactile Internet (IEEEP1918.1) was launched, in which Tactile Internet architec-ture, functional entities, and various use cases have beeninvestigated. In [7] and [8], detailed traffic characteristics

2In this paper, URLLC service and Tactile Internet service are usedinterchangeably because both share most of interesting use cases inliterature and pursue the same service vision and requirements.

of video, audio, and haptic information such as packetsize, arrival rate, and arrival model with QoS such aslatency and reliability requirements are described. Theseexamples and scenarios show that traffic characteristicsof typical URLLC services can be quite various in termsof their packet sizes and arrival models and their QoSrequirements can be quite extreme [11], [12]; therefore,these aspects should be taken into account when develop-ing URLLC techniques.

To support such low-latency requirements of URLLCservices, studies in the 3rd Generation Partnership Project(3GPP) on the current long-term evolution (LTE) systemshave been performed [13], in which typical downlink(DL)/uplink (UL) radio access and handover latenciesare reported as 17/7.5 ms and 50 ms, respectively,and the transmit-time-interval (TTI) reduction, processingtime reduction, semi-persistent scheduling, and grant-freeaccess are enumerated as possible remedies. In addition,many technical aspects in the cellular network, includingwaveform numerology such as symbol length and subcar-rier spacing, frame structure, multiple-access scheme, pilotdesign, link adaptation strategy, and scheduling policyneed to be designed carefully for URLLC [14]–[16]. 3GPPis standardizing a new radio interface for 5G as the newradio (NR), aiming to reduce DL/UL radio access latency to0.5 ms [17]. In [18] and [19], scalable subcarrier spacingparameters for shorter orthogonal frequency-division mul-tiplexing (OFDM) symbol length and mini-slots comprisinga various number of OFDM symbols (1–13) are adopted forimplementing short TTIs. Further, various ideas on URLLCand enhanced mobile broadband (eMBB) multiplexing forefficient resource utilization [20]–[22] and various ideason two-way grant-based and grant-free multiple-accessproposals [23]–[25] to reduce the uplink protocol latencyhave been discussed. However, such simple suggestionson providing low-latency protocols and frame structuresshould be just the beginning, as practical URLLC ser-vices need a simultaneous provision of low-latency andultrareliability with high spectral efficiency, which is verychallenging.

In a cellular system, the channel impulse (or frequency)responses of wireless fading channels are not fully pre-dictable and the fluctuation on the received signal-to-noise-plus-interference ratio (SINR) is one of the mostchallenging aspects for reliable information delivery. Thecurrent 3GPP LTE employs an appropriate scheduling toutilize multiuser diversity, adaptive modulation and coding(AMC) according to channel quality information measuredat a receiver, and hybrid automatic repeat and request(HARQ) for an efficient retransmission to provide highreliability as well as high spectral efficiency [26]. However,such an approach requires delays for channel quality mea-sure and feedback, scheduling, and retransmission thatit becomes inappropriate for delivering highly latency-sensitive information, although it is the most efficient wayto deliver latency-insensitive information. Although somediversity schemes and fast HARQ schemes for better relia-

Vol. 107, No. 2, February 2019 | PROCEEDINGS OF THE IEEE 377


bility at a low-latency are considered, such as in [27]–[29],their reliability levels and the resulting spectral efficien-cies are far from what is required for practical URLLCservices. Further, the design of the physical (PHY) layerand medium-access control (MAC) layer technologies forURLLC need to consider the variety of different trafficcharacteristics and the QoS of URLLC services.

Since 2015, the authors had formed a joint URLLCresearch team and focused on developing spectrallyefficient protocol and multiple access technologies thatguarantee both tight low-latency and ultrareliabilityrequirements for URLLC. To provide ultrareliability, alarge amount of diversity obtained from large degreesof freedom is essential, especially without either instan-taneous channel quality feedback or retransmissions infading channels. Thus, considering a large-scale antennasystem (LSAS) [or massive multiple-input–multiple-output(MIMO) system] is a natural consequence [30], [31]. Inthis paper, some novel multiple-access schemes for URLLCbased on the LSAS are introduced and waveform mul-tiplexing and full-duplex communication techniques arealso introduced to further enhance the spectral efficiencyand reduce the latency. In addition, a new evaluationmethodology is introduced by combining a system-levelsimulator and a ray-tracing tool with digital maps on realenvironments and the performance evaluation results areprovided.

II. S O M E U S E C A S E S A N D T R A F F I CR E Q U I R E M E N T S F O R TA C T I L EI N T E R N E T S E RV I C E S

A. Some Use Cases

1) Immersive Virtual Reality: The immersive virtual real-ity (IVR) technology allows people to use their senses tointeract with virtual entities in remote or virtually createdenvironments such that they can perceive all five senseswhen they are in such remote or virtual environments [8].Because of its interaction capability beyond the physicallimitation, it has been drawing great interest in industriessuch as gaming, education, and healthcare [33].

Among the five senses, the vision, sound, and touchsenses represent the primary focus and their traffic typesand characteristics can be categorized according to eachsense. For vision sensing, since the motion-to-photonlatency should be within 10–20 ms, the allowed airlatency ranges from submilliseconds to a few millisec-onds [6], [34]. Further, considering the field of view,be it in three dimensions, or extremely high definition(32K) [35], a required data rate for vision informationwould be in the range between 10 Mb/s and 1 Gb/s with99.9%–99.999% reliability. For audio sensing, the audioinformation includes not only high-fidelity sound but alsoconsiderations for 3-D head rotations. For touch sensing,haptic information exchange is required, in which tactileinformation comprising several bytes for each degree offreedom (DoF) [8] times the number of DoFs (i.e., the

number of touch spots) up to thousands and the kinestheticinformation comprising several bytes per each DoF [8]times the number of DoFs (i.e., the number of joints inthe human body) up to hundreds with 99.999% reliability.

2) Teleoperation: Teleoperations, such as telesurgery,telemaintenance, and telesoccer using remote roboticavatars, allow people to control slave devices such asrobots in distant or inaccessible environments to performcomplex tasks [8], [36]. The exchange of haptic informa-tion, such as force, torque, velocity, vibration, touch, andpressure, is required between the master and slave devices,and the delivery of high-quality video and audio infor-mation is required from the slave devices to the masterdevices [8].

The required data rates and latency requirements ofthe traffic for teleoperation vary according to the requiredcontrol precisions for slave devices and the dynamics ofremote environments where the slave devices are placed.In a highly dynamic environment such as the one reportedin [37], haptic information exchange should be within afew milliseconds such that the allowed air latency is lessthan or equal to one millisecond. Further, for applicationsrequiring extremely high control precision such as in [9]and [10], the delivery of very high-rate video informationand the exchange of delicate haptic information with relia-bility higher than 99.999% is required. Further, for remoteskill training such as in [38], the number of DoFs can behundreds to thousands.

3) Automotive and Internet of Drones (IoD): Future carsneed connectivity with the infrastructure and other carsfor collaborative autonomous driving and in-car entertain-ments [3]. Therefore, a large amount of sensing informa-tion needs to be exchanged in a very low latency. Similarto teleoperation applications, the required latency dependson the dynamics of neighboring environments such thatthe allowed air latency can be less than or equal to onemillisecond with high reliability. In addition, for collabora-tive autonomous driving using artificial intelligence, high-quality video and audio information exchange with highreliability among neighboring cars may be required. Inremote driving, haptic information exchange with DoFs upto several tens to hundreds may be required [8].

Applications using unmanned aerial vehicles (UAVs),such as drones, are also emerging and among them aredrones for public safety, remote explorations, logistics,flying base stations, etc. [8], [39]. Owing to the highdynamics in such UAV environments, real-time video,audio, and haptic information should be exchanged withina low latency, i.e., the allowed air latency less than orequal to one millisecond for kinesthetic information anda few milliseconds to tens of milliseconds for high-qualityvideo/audio information and haptic information.

4) Interpersonal Communication: Interpersonal commu-nication (IC) supports the copresence of distant users forsocial development or emotional interaction, and haptic



Table 1 Typical Traffic Characteristics and QoS for Some Use Cases

IC (HIC) can deliver human touch as well, allowing forpromising applications such as social networking, gaming,education, and training [8], [40], [41].

High-quality video and audio information exchange isrequired with high reliability, similar to the IVR case, andhaptic information exchange is also required. In static dia-loguing, a low latency is required for the highly dynamicinteraction of haptic information such that the allowed airlatency can be as low as a few milliseconds [8].

B. Traffic Classification

In Table 1, typical traffic characteristics of the usecases in Section II-A are summarized, in which the trafficcharacteristics and QoS are represented by the typical airlatency, target reliability, packet size, packet arrival rate,and model. Here, the baseline of the traffic characteristicsand QoS come from [8] but it is further assumed thattypical air-latency requirements are set to 20% of thecorresponding end-to-end latency requirements and moreDoFs and larger packet sizes up to ten times are expectedin near future.

The current state-of-the-art cellular communicationtechnology can support various traffic with different char-acteristics and QoS with good reliability and high spectralefficiency if the required latency is not so tight by con-trolling the radio resource control (RRC) connectivity ofeach user according to its activity, scheduling active userswith good channel quality, AMC according to its chan-nel quality, and retransmissions using HARQ. However,extremely low-latency and high-reliability requirementsof URLLC services necessitate classifying such traffic andoperating different protocols and multiple-access strategies

according to different target latency and reliability lev-els. Furthermore, traffic characteristics such as the arrivalmodel and rate also need to be taken into account to designsuch protocols and multiple-access strategies.

First, traffic with loose latency requirements(i.e., L > 50 ms) can be easily supported using alegacy strategy: radio-resource efficient LTE-style four-wayRRC connection, scheduling, AMC, and HARQ regardlessof traffic type, data rate, arrival model, and targetreliability level. In this case, it is not difficult to satisfythe latency and reliability requirements and a higherspectral efficiency is of the most interest. When a latencyrequirement is slightly tight (i.e., L is approximately tensof milliseconds), it is better to deal with such packetsdifferently from those with loose latency requirements.Some good approaches include reducing the number ofhandshakes in the RRC connection protocol as in [23]and [24] and shortening the TTI as in [18] and [19].As the target latency is not so tight, it is possible toapply a grant-based multiple access with a radio resourcemanagement and a few retransmission can be allowed toguarantee the reliability requirement.

As the latency requirement becomes tighter (i.e., L isapproximately several milliseconds), more elaboratelydesigned techniques need to be applied and the trafficarrival model becomes important. For periodically gener-ated packets, a semipersistent scheduling can be appliedto reserve the radio resources for such packets. Further,at least one or two retransmissions may be allowed suchthat a reliability requirement can be met with an LTE-style spectrally efficient radio resource management. How-ever, in cases of bursty or sporadically generated packets,



Table 2 Traffic Classification Example

a grant-free multiple access, similar as in [42], is necessaryand it is important to guarantee a target reliability, which isvery challenging. As users can transmit without any grant,the number of users sharing the same radio resources (i.e.,a subchannel) varies and it becomes worse if traffic withdifferent characteristics and QoS (such as packet size andtarget reliability level) are allocated in the same subchan-nel of a multiple-access scheme. In addition, as the latencyrequirement becomes extremely tight (i.e., L is less than orequal to 1 ms), retransmission may not be allowed and itbecomes very difficult to satisfy a reliability requirementat a reasonable spectral efficiency. A good approach isto classify traffic classes according to the characteristicsand QoS such that packets with similar characteristics andQoS are allocated in each subchannel of a multiple-accessscheme.

In Table 2, traffic in Table 1 is classified as an example,mainly according to the latency and reliability require-ments, as discussed in paragraphs above. Here, the firstrow represents a class with loose latency requirements,the second row represents a class with medium latencyrequirements, and the third row represents a class withlow latency requirements but the packets are generatedperiodically. The other five classes represent very low-latency requirements with bursty or sporadic packet arrivalcharacteristics. As they need to be served using a grant-freemultiple access, those packets should be further classifiedaccording to latency, reliability requirements, and packetsizes so that traffic with similar characteristics are allocatedto a subchannel for a reasonable spectrally efficient radioresource management.

III. M U LT I P L E-A C C E S S S T R AT E G YF O R U R L L C

In Section II, typical URLLC use cases and traffic character-istics are introduced, which necessitate the developmentof not only a new frame structure with short TTIs andprotocol concepts such as in [17]–[19], but also elab-orately designed strategies for user RRC state control,radio resource management and optimization, and novelmultiple-access techniques each suitable for the varioustraffic characteristics and QoS of URLLC services.

In this section, a new user RRC control strategy is sug-gested with new states for serving traffic with low-latencyrequirements and the corresponding RRC connection pro-tocols are suggested, in which different levels of protocolprocedures, core network connection strategies, and radioresource allocation strategies are provided according tothe traffic classes of URLLC users. In addition, DL and ULradio resources are appropriately partitioned to supportdifferent multiple-access schemes, where each multiple-access component handles traffic with similar characteris-tics and QoS for better spectral efficiency. To provide a highlevel of reliability even in cases of extremely low-latencyrequirements, an LSAS is assumed for a base station anda latency-optimal radio resource management scheme issuggested. According to each user’s RRC state, a differentlevel of radio resource allocation is provided to enhancethe spectral efficiency while guaranteeing the latency andreliability requirements.

A. RRC Connection Protocols

Recently, in the 3GPP standardization for NR, a new userRRC state, RRC_INACTIVE, has been defined [43]–[45].According to a general description in [46], RRC_INACTIVEis different from RRC_IDLE in that a user keeps theprevious configuration information when suspended fromRRC_CONNECTED, so that it can resume RRC connectivitywithout a long delay. Furthermore, from a core network’sperspective, the RRC_INACTIVE and RRC_CONNECTEDare the same because core network is connected(i.e., CN_CONNECTED) in both cases. If the RRC connec-tivity is lost, a user needs to perform the RRC connectionsetup similarly from RRC_IDLE.

In this paper, such a new state is further classi-fied into two different states according to the requirednumber of handshakes between base stations andusers and their levels of allocated radio resources.As shown in Fig. 1, two new states, RRC_INACTIVE andRRC_INACTIVE _CONNECTED, are introduced. Here, froma core network’s perspective, both states are the sameto the RRC_CONNECTED. To a user in RRC_INACTIVEor RRC_INACTIVE_CONNECTED, preambles are allo-cated as dedicated radio resources in addition to the



Fig. 1. RRC state transition diagram.

Fig. 2. Three different procedures for RRC connection.

RRC configuration information and they uniquely indi-cate each user’s identity and intended traffic classes.Furthermore, to each traffic class of each user inRRC_INACTIVE_CONNECTED, the subchannel for possibleUL transmissions is allocated as a shared resource. Here,traffic classes for each service are assumed to be registeredat the initial service negotiation and RRC connection stage(i.e., admission). If some traffic classes of a user requiremedium to low latency (for example, Class 2 in Table 2),then the user can utilize the RRC_INACTIVE state, in whichthe allocated preambles indicating the user identity andeach of the traffic classes with such latency requirementsenable a fast RRC connection setup once a packet of suchclasses arrives. In addition, if some traffic classes of a userrequire very low latency (for example, Class 6 in Table 2),the user can utilize the RRC_INACTIVE_CONNECTEDstate, where the allocated subchannel and user-specificpreamble for each of such traffic classes enable immediate

RRC connection resuming as soon as a packet of such aclass arrives.

The discussion above on the proposed RRC state tran-sition can be redrawn in a protocol perspective in Fig. 2.Here, three protocols for the RRC connection are pre-sented, in which the first one represents an LTE-stylefour-way handshaking RRC connection procedure, andthe second one represents a two-way handshaking RRCconnection procedure for providing a fast-grant multipleaccess (FGMA), and the last one represents an immediateRRC connection for providing a grant-free multiple access(GFMA).

The first protocol is for traffic classes with loose latencyrequirements, such as in the first row in Table 2. TheLTE-style four-way handshaking using a cell-specific com-mon set of preambles is spectrally efficient and the LTE-style granted access is performed in the UL. However,as the latency becomes slightly tight, the delay caused



Fig. 3. Multiplexing of different multiple accesses for DL and UL.

by a grant procedure needs to be reduced and sendinga scheduling request after a (sporadic or bursty) packetarrival is enough to obtain a grant for intended packetssince a unique preamble indicating the user identity andthe intended traffic class identity is already allocated andused in the scheduling request so that a base stationperforms a scheduling for the reliable delivery of suchpackets immediately and sends a grant with the allocatedsubchannel information. This protocol and FGMA is suit-able for traffic with medium latency requirements as inthe second row in Table 2 and traffic with low-latencyrequirements but periodic arrival characteristics as in thethird row in Table 2 so that a semipersistent schedulingand subchannel allocation can be used.

For traffic with low-latency requirements, the protocolsabove may not be used and an immediate packet transmis-sion is required in the UL as soon as a packet of such a classarrives. In this case, the third protocol for GFMA is sug-gested, in which a subchannel as a shared resource and auser-specific preamble as a dedicated resource are alreadyallocated for each traffic class with a low-latency require-ment and used for an immediate packet transmission assoon as a packet arrives. The GFMA with such a protocol issuitable for traffic with low-latency requirements as in thefourth to eighth rows in Table 2.

To employ such different multiple-access schemes in asingle carrier, the DL and UL radio resources are parti-tioned as shown in Fig. 3. For each service of each user,traffic is classified according to traffic characteristics andQoS as described in Section II and traffic of multipleusers with similar characteristics and QoS are grouped andserved together in each multiple access. Although different

procedures for the RRC connection and the correspondingmultiple-access concepts are proposed to support variouslatency requirements required for URLLC services, provid-ing reliability at a reasonably high spectral efficiency is stillquite challenging. One good approach is to make the trafficrequirements and QoS of multiple users in each FGMA orGFMA component as similar as possible and it can facilitatedesigning a radio resource management for reliability andhigh spectral efficiency [47].

B. Multiple Access With Latency-Optimal RadioResource Management

Reliable information delivery in a cellular communica-tion environment has been challenging because of channelquality fluctuation caused by the wireless fading channeland mobility. Although a low level of reliability couldbe provided by exploiting a limited order of diversity intime, frequency, and space in legacy cellular communi-cations (the second generation or the earlier stage thirdgeneration), the LTE has been successful in providing ahigh level of reliability primarily by using AMC based onchannel quality information measurement and feedbackand retransmissions using HARQ [26].

However for URLLC services, a low-latency requirementmay restrict the use of AMC and HARQ or at least allowthem only in a very limited manner so that most of thereliability part needs to be resorted on diversity again.Although classical repetition approaches can be adoptedin time, frequency, or even multiple communication inter-faces can be used [48], spectral efficiency may be sig-nificantly degraded, especially as more URLLC services



Fig. 4. Optimal radio resource management strategy for FGMA satisfying latency and reliability requirements. (a) Preamble indicating user

and traffic class. (b) Latency-optimal radio resource management and optimal frame configuration.

are served. Thus, better approaches without significantspectral efficiency degradation are preferred and the mostpromising solution is to employ a large number of antennasat a base station, i.e., LSAS.

Once an LSAS is assumed, the channel fluctuationcaused by the wireless fading channel and mobility canbe overcome (or significantly reduced at least) becauseof the channel hardening effect [49]. However, the chal-lenge is on the radio resource optimization in whichpreamble overhead, channel estimation quality, and usergrouping are jointly considered and optimized. In [47],Choi and Kim proposed a latency-optimal semipersistentscheduling algorithm for an LSAS, which can be utilizedfor guaranteed reliability in FGMA or GFMA.

For FGMA, a unique preamble indicating the user iden-tity and traffic class identity, as shown in Fig. 4(a), isallocated to each traffic class for each user during theadmission control process. As the traffic characteristics andQoS of an arrived packet of each user, such as packetsize, arrival model and rate, and latency and reliabilityrequirements, can be detected at a base station from thepreamble sent during its scheduling request, the base sta-tion can group users with similar traffic characteristics andQoS and apply the latency-optimal scheduling algorithmin [47]. As shown in Fig. 4(b), the transmit power ofeach user is first optimized based on the long-term channelstate information and energy information of each user andthen optimal user grouping is performed in which eachuser group shares a subchannel. From the optimizationresults, the pilot overhead for each subchannel is dynam-ically optimized as shown in Fig. 4(b) and the amountof resource to guarantee traffic delivery with reliabilityand latency requirements is determined. Subsequently,the subchannel construction and allocation informationis delivered in a resource grant. Therefore, the proposed

FGMA with a latency-optimal radio resource managementcan maximize the spectral efficiency while guaranteeinglatency and reliability requirements for URLLC services.

In GFMA, the challenge for guaranteed reliability is evenmore difficult because each user transmits its packet assoon as it arrives without any grant. However, employingan LSAS can also reduce such an uncertainty in additionto the channel hardening effect reducing the uncertaintyin channel quality and it is possible to modify the algo-rithm in [47] by considering such uncertainties togetheras indicated in Fig. 5. As traffic with similar characteristicsand QoS is already grouped at the admission control stage,a base station is aware of the arrival model and rate ofeach user so that it can be aware of the statistics for theactual transmitting users in each user group candidate.Thus, the base station can determine the optimal schedul-ing by considering such statistics of the user groupingcandidates [42], [50]. At the base station receiver, userdetection needs to be first performed using preamblesand the user detection capability should provide a successprobability higher than the required reliability level, whichis also considered in the optimization process [42], [50].Simulation results in [42] and [50] showed that therequired radio resource for GFMA does not increase sig-nificantly compared with the case of a granted multipleaccess such that the proposed protocol and GFMA witha latency-optimal radio resource management can maxi-mize the spectral efficiency while guaranteeing latency andreliability requirements for URLLC services.

IV. M O R E P H Y T E C H N O L O G I E S

In the previous section, a set of multiple-access tech-niques have been introduced, in which 1) data packetsfor URLLC services are classified according to their trafficcharacteristics, including packet size and arrival statistics,



Fig. 5. Radio resource management concept and receiver structure for GFMA [50].

and their latency and reliability requirements; 2) radioresources are partitioned to multiplex different multiple-access components simultaneously; 3) each user or basestation is equipped with as many queues as its numberof different packet classes; and 4) each multiple accesssupports its own packet class of multiple users. By virtueof the large number of antennas and the latency-optimalscheduling, the latency and reliability requirements of eachpacket class can be simultaneously satisfied.

However, to realize such a concept, the radio resourceneeds to be well partitioned in a waveform level withgood synchronization strategy. Moreover, to maximize thespectral efficiency, it is desired to use waveforms not onlymatched to user environment (i.e., delay spread and mobil-ity) similar to the numerology multiplexing concept [51],but also appropriate for the latency requirements of usersbecause the latency caused by the filters in a transceivercan be critical for packets with extremely low-latencyrequirements. Thus, to devise a waveform multiplexing isa natural consequence, in which different types of wave-forms [i.e., filtered-OFDM, generalized frequency-divisionmultiplexing (GFDM), etc.] each with different numerolo-gies (cyclic prefix length, subcarrier spacing, filter length,etc.) are multiplexed.

In a latency budget such as in [13], one importantcomponent is the processing delay and most of the process-ing delay comes from the channel encoding/decodinglatency. Thus, it is very important to devise channel codeswith low encoding/decoding latency. In addition, such

channel codes should have good performance in a high-reliability regime [i.e., frame error rate (FER) in therange from 10−3 to 10−7] such that both the water-fallperformance and the error floor performance need to beconsidered.

To further improve the spectral efficiency and reduce thedelay between the DL and UL, the best method is to adoptfull-duplex communication and the corresponding framestructure. Since an LSAS is assumed, a channel reciprocitysuch as in time-division duplexing (TDD) is required forefficient channel estimations. However, in TDD, the delaybetween UL and DL subframes [or (mini)slots] may causea latency problem. Thus, a practically feasible full-duplexcellular communication technique can provide not onlyalmost double the spectral efficiency but also a reduceddelay between UL and DL as in frequency-division duplex-ing (FDD). Although the feasibility for self-interferencecancellation (SIC) at a (low-power) base station has beenconfirmed [53], [54], the interference at DL users causedfrom UL users needs to be avoided. Although a full-duplexcellular communication can work if an appropriate paringof DL/UL users is assumed [55], a better interferenceavoidance scheme needs to be devised for incorporationwith various scheduling strategies without such pairings.

A. Waveform Multiplexing

To multiplex different classes of services in a commoncarrier, one approach is to multiplex URLLC packets on



Fig. 6. The proposed waveform multiplexing concept [52].

eMBB resources, such as in [21], and the other is anumerology multiplexing by resource partitioning, such asin [20], where different numerologies for OFDM parame-ters and frame structure can coexist. Owing to the capa-bility of selecting the appropriate numerologies accordingto the users’ environments and service requirements, anumerology multiplexing is considered as a promisingsolution. However, internumerology interference needs tobe taken into account and an appropriate filter design forlow out-of-band-emission (OOBE) is required [56].

As an elegant solution to implement such a numerol-ogy multiplexing concept, a waveform multiplexing sys-tem is proposed in [52] as illustrated in Fig. 6, and itemploys scalable subcarrier spacings and dynamic cyclicprefic (CP) managements. The proposed waveform mul-tiplexing selects not only appropriate subcarrier spacingsand CP lengths according to users’ channel environmentand mobility but also waveform filters for minimum guardbands according to service (latency) requirements andOOBE levels.

In the proposed waveform multiplexing, users withsimilar mobility (channel coherence time), delay spread,and latency requirements are grouped and the appropriatesubcarrier spacing and CP length are selected for eachgroup to minimize the CP overhead. For each group, anappropriate waveform filter is determined to minimizethe guard band while satisfying latency requirements. Incases where a low OOBE level is desired and latencyrequirement is loose, waveforms with very low OOBE, suchas in [57]–[59], can be used to enhance the frequency-domain SINR. However, in cases where an extremely lowlatency is required, waveforms with short filter delays at areasonable OOBE, such as in [60], [61], may be preferred.Such a waveform multiplexing concept can be considered

a generalization of numerology multiplexing in a singlewaveform, such as in [57] and [62].

B. Synchronization Issue

In the DL, each user can employ a legacy time and fre-quency synchronization, such as in [67] and [68], on thesubbands where it belongs to, even in cases of employinga waveform multiplexing with different numerologies andwaveform filters. Thus, synchronization for DL does notraise a new critical issue and can be done similarly as inthe LTE.

In the UL, it is reasonable to assume a similar closed-loop procedure for a strict time synchronization as in theLTE for eMBB and URLLC services. However, since highermobility and higher frequency bands need to be sup-ported, especially for URLLC services, time synchronizationerrors and frequency offsets due to Doppler shifts maycause nonnegligible performance degradation, especiallyfor URLLC services in which high reliability is required andsporadic access needs to be supported. As a remedy, aninterference cancellation approach is adopted at the base-station receiver [63], [64] as illustrated in Fig. 7(a). Here,different time and frequency offsets of multiple users areassumed to be estimated similarly as in [69] and [70]along with a closed-loop time synchronization as in theLTE. In order to reduce multiple user interference causedfrom time and frequency offsets of multiple users whichmay be caused from high mobility, high frequency, orsporadic access, an elaborately designed receiver filter isapplied to maximize the signal-to-interference ratio (SIR)of users by using the estimated time and frequency off-sets, and it is shown in [63] and [64] that the proposedapproach can provide better performance compared tothose in [71] and [72], as shown in Fig. 7(b).



Fig. 7. UL receiver structure for handling the synchronization issue [63], [64]. (a) The proposed receiver filter structure at a base station.

(b) Performance comparison.

C. Channel Codes for URLLC

Recently, low-density parity-check (LDPC) codes havebeen adopted for eMBB services in the NR standard [73].Such LDPC codes can be considered as raptor-like quasi-cyclic LDPC (QC-LDPC) codes and they can provide near-optimal water-fall performance as well as efficient encod-ing and decoding implementation methods such that theyare quite appropriate for eMBB applications. However,their error floor performance may not be good, especiallyas the code rate decreases, because of the lack of linearminimum distance growth (LMDG) property and too manydegree-1 variable nodes, as expected in [74], and it maylimit the use of such protograph-based raptor-like (PBRL)QC-LDPC codes for URLLC applications, especially for thecases where the required reliability is quite high (e.g.,FER in the range from 10−3 to 10−7) and the latencyrequirement is tight such that a retransmission is notallowed.

In [65], accumulate–repeat–accumulate–check–accumulate (ARACA) codes have been recently proposedto provide high reliability by having both the LMDGproperty (i.e., no error floor) and good water-fallperformance with an efficient encoding structure. Fig. 8(a)shows the protograph structure of an ARACA code, whichcomprises the two outer code parts (o1 and o2) and thetwo inner code parts (i1 and i2) as described in [65],and it is characterized by the outer connections that canprovide an efficient low-complexity encoding similar toan accumulate–repeat–accumulate code [75] as well asthe LMDG property with a water-fall performance similarto an accumulate–repeat–jagged–accumulate code [76].Further, Fig. 8(b) shows the good performance of rate-compatible ARACA codes [66] compared with PBRLQC-LDPC codes in [77]–[79]. In addition, Jeon and Kim[80] propose a low-latency and low-complexity layeredRichardson–Urbanke encoding method and encoderstructure as well as a low-latency and low-complexitybig-layer parallel decoding method and decoder structure,

which shows that the proposed ARACA codes arepromising for URLLC services.

D. Full-Duplex Cellular Communication:Code-Division Duplexing Spatial-DivisionMultiple Access (CDD-SDMA)

In a full-duplex cellular communication, DL users sufferfrom the interference caused by UL users as illustrated inFig. 9(a). As a result, without an elaborate management ofsuch interference, the overall performance, such as the ratedistribution of users, cannot be meaningfully improved,even in cases where perfect SIC is assumed at a basestation.

As a remedy, a novel CDD-SDMA is proposed, in whichUL interferences are aligned to a null space orthogonalto the signal subspace for DL multiuser MIMO (MU-MIMO) by using orthogonal codes between DL and UL andemploying antenna reconfiguration (or different versionsof analog beamforming) to align all UL interferences into asingle dimension of the DL signal space similarly as in [81]and devising an efficient DL/UL MU-MIMO schemes on theremaining signal subspaces on DL/UL similarly as in [82]with the DoFs approaching that of the normal zero-forcingMU-MIMO in a half-duplex DL/UL.

To confirm the feasibility of the proposed CDD-SDMA,not only a practical indoor hotspot environment and aspatial channel model in [83] are used but also adjacentchannel interference and in-band blocking due to theremaining frequency offsets among UL users (<100 Hz)and the finite resolution (12 b) of analog-to-digital con-verters as well as cochannel interferences at the sameresource block are considered in the case of 40 basebandstreams and 100 physical antennas at a base station.As shown in Fig. 9(b), more than 70% improvement inspectral efficiency is expected in a practical environment,even considering such nonideal effects. Thus, the proposedCDD-SDMA can be considered as a promising solution,not only for almost doubled spectral efficiency but also to



Fig. 8. Protograph structure of an ARACA code and performance comparison [65] [66]. (a) Protograph structure. (b) Required SNRs for

target FERs.

Fig. 9. CDD-SDMA concept for a full-duplex cellular communication and performance evaluation. (a) CDD-SDMA concept. (b) Performance

evaluation.

significantly reduce the delay between UL and DL whileexploiting channel reciprocity.

V. E VA L U AT I O N M E T H O D O L O G Y A N DS I M U L AT I O N R E S U LT S

Two-dimensional regular cell layouts with 2-D stochasticwireless fading channel models have been widely usedto evaluate the performance of legacy cellular systems.As the uses of multiple antennas and small cells becomewidespread, beam-steering effects according to elevationangles matter and such 2-D channel models with a 2-D

regular cell layouts have evolved into 3-D channel mod-els by applying stochastic channel parameters for eleva-tion angles, which include the 3GPP 3-D spatial channelmodel [83] used for evaluating LTE technologies. How-ever, in most LTE system-level simulation (SLS) scenar-ios, 2-D regular layouts have been commonly used suchthat the channel parameters between a randomly selectedtransmitter and receiver pair are primarily dependent onscenario-dependent parameters and the locations of nodesincluding the two in the desired pair and interferingsources.



Fig. 10. Proposed evaluation methodology.

As a typical cell size shrinks for an enlarged area spectralefficiency, a cell deployment scenario should consider thegeography of a target environment including its landform,shapes, and heights of surrounding structures such asbuildings, and different attenuation factors due to differentconstituent materials of each surrounding structure. Toexploit such a real geography, map-based channel modelsutilizing ray-tracing tools have drawn much interest fromacademia and industry, such as in [84]–[90]. Reason-able agreement with hardware measurements has beenreported in [84]–[86], [88], [89], and [91] and link-level simulations (LLSs) and SLSs have been performedto evaluate their proposed work by using measurementsfrom hardware testbeds and/or software algorithms inenvironments similar to real world [52], [91]–[99].

Fig. 10 shows the proposed 3-D SLS evaluation strategyin this paper for the performance of UL multiple access andDL waveform multiplexing, where a high-resolution digitalmap is constructed for the GangNam station area (Seoul,South Korea), real base station deployment information,such as the locations as well as the antenna heights andtilting angles, are taken into account, and the reportedtypical user density for each part of the digital map isapplied as in [52] and [100]. Using such a realistic digitalmap and the locations for base stations and users, 3-Dchannel parameters are collected [52], [90], [100] usingthe ray-tracing tool called wireless system engineering

(WiSE) developed by Bell Laboratories [85]. Subsequently,based on the collected data, 3-D wireless channels aregenerated as in [83] according to either a deterministicmodel based on specific locations of the transmitter andreceiver pairs or a stochastic model with statistics matchedto this specific environment according to this digitalmap.

In Fig. 11, the UL latency and spectralefficiency distribution of typically distributed 6000RRC_INACTIVE_CONNECTED users are shown when anantenna array with 128 antenna elements is assumed ineach of 12 real base stations deployed in the GangNamstation area with the same height and tilting angle. Here,the traffic characteristics are assumed as follows: thepacket size is 8 KB (64 kb), the average arrival rate is100 packets per second, the arrival model is a sporadicPoisson random arrival, and the latency and reliabilityconstraints are 2 ms and 99.999%, respectively. Here, theCP overhead is assumed to be 25% and variable-length(minimum 0.2 ms) minislots comprising the subchannelsare assumed for a frame structure. Furthermore, usersare associated with the nearest base station and eachbase station determines the required amount of radioresources (minislot length and bandwidth) to supportgrant-free accesses with guaranteed QoS for its associatedusers, in which user grouping, portion of pilot symbols ineach subchannel allocated to each user group, and power



Fig. 11. UL performance evaluation using GFMA for a URLLC service. (a) Latency distribution of URLLC users. (b) Spectral efficiency

distribution for URLLC resource.

control for each user are optimized. Then, the spectralefficiency is evaluated as the ratio of the sum of goodputs(in bits per second) and the sum of required amountof bandwidth (in hertz) of these cells considering theCP/pilot overhead, channel estimation error, and channelcode FER performance. In addition, to evaluate the latencydistribution, the required TTI length as well as queuingdelay due to a random packet arrival, wireless propagationdelay, and decoder processing delay are also taken intoaccount, similarly as in [13], where the queuing delayis assumed to be uniformly distributed over a minimumTTI length and the throughput and latency of the channeldecoder at a base station are assumed to be 50 Gb/s andequal to the minimum TTI length, respectively.

In Fig. 11(a), the latency distribution of the proposedGFMA is evaluated and compared with the cases wherean LTE-style four-way access with a round-robin schedul-ing and equal power control (denoted as the “LTE-AExtension”) and the proposed FGMA are instead applied.Here, in addition to the processing delay for decoding,a processing delay for scheduling as long as two timesthe minimum TTI length is considered in FGMA. Further-more, in Fig. 11(b), the spectral efficiency distribution(for goodput only) of the proposed GFMA is evaluatedand compared with the two cases. From the results, it isconfirmed that the proposed GFMA is the most efficient fortraffic with tight latency requirements and sporadic arrivalcharacteristics. In the “LTE-Extension,” a large amountof latency budget (more than 80%) is wasted for thefour-way handshaking such that the latency distributionand the spectral efficiency distribution for goodput aresignificantly degraded. In FGMA, although some portion inthe latency budget is spent for the two-way handshakingand scheduling, a granted access with a latency-optimalscheduling improves the spectral efficiency during dataminislots so that these two schemes can provide simi-lar spectral efficiency performance in this specific case.In general, as the latency requirement becomes tighterand/or more antennas are equipped at a base station,

GFMA performs better than FGMA. Also in Fig. 11(b), thespectral efficiency distribution of FGMA is shown whenthe packet arrival model changes to be periodic withperfectly aligned arrival times so that a semipersistentscheduling and resource allocation can be allowed. In thiscase, FGMA can provide much higher spectral efficiencywith guaranteed latency and reliability because the fastprotocol of FGMA enables an initial access with guaranteedlatency requirement and such initial overhead for the grantbecomes ignorable.

In summary, Fig. 11 shows that 1) although equippedwith a large number of antennas and reduced TTIs,LTE-style RRC connection protocol and multiple accesscannot provide sufficiently high reliability and low latencyeven at a very low spectral efficiency; and 2) the proposedGFMA and FGMA can successfully guarantee high reliabil-ity and low latency at reasonably high spectral efficiencyaccording to traffic class and QoS.

In addition, Fig. 12 shows the performance gain ofemploying the proposed waveform multiplexing in DL.

Fig. 12. Achievable waveform multiplexing gain



Here, to clearly show the advantage of the proposedscheme, a single transmit antenna is instead assumed foreach base station. The upper bound (assuming ideallycontrolled dynamic CP lengths, optimal OFDM parametersand ideal filter characteristics by Genie) on the spectralefficiency distribution of the proposed waveform multi-plexing is shown and compared with the case of con-ventional LTE-based multiband OFDM. Here, the overallperformance gain can be as high as 1.67 times, in whichthe two gains from selecting the ideal waveform on eachsubband according to OOBE characteristics and latencyrequirements and from selecting the optimal CP lengthand OFDM parameters are both meaningful in a realisticscenario.

Although it might be too optimistic, a direct combi-nation of the results in Figs. 9 and 12 with those inFig. 11 may anticipate that further gains in spectral effi-ciency with respect to those shown in Fig. 11 (up to100%) can be obtained by combining waveform multi-plexing and CDD-SDMA with the proposed multiple-accessschemes.

VI. C O N C L U S I O N

In this paper, novel URLLC techniques were introducedfor realizing Tactile Internet services in realistic environ-ments. The traffic characteristics and required QoS oftypical URLLC (or Tactile Internet) services in literaturewere summarized and classified from the perspective ofdesigning the PHY and MAC layers of a cellular system.Investigations on typical traffic in typical use cases justified

the necessity of defining new user states and devisingprotocols for RRC connection according to latency require-ments, multiplexing of multiple-access schemes over radioresources to meet a variety of different traffic character-istics and QoS of URLLC services, and the developmentof latency-optimal radio resource management strategiesto maximize the spectral efficiency while guaranteeing thelatency and reliability requirements.

This paper proposed two additional user states aimedfor low latency and devised the corresponding protocolsand radio resource allocation strategies in detail. Further-more, a realistic map-based SLS approach was proposedbased on a refined digital map construction, a realisticnode distribution scenario, data collection via a ray-tracingtool, and the corresponding deterministic or stochastic3-D channel model. Simulation results showed that theproposed schemes are promising for supporting URLLCservices with high spectral efficiency while guaranteeinglatency and reliability requirements.

To implement the proposed protocols and multiple-access schemes in a spectrally efficient way, more PHYtechnologies on waveform multiplexing and synchroniza-tion strategy, channel codes for low processing delay andhigh reliability, and a novel DL/UL MU-MIMO conceptcombining interference alignment for a practical full-duplex cellular communication were further introduced,where each of them can provide significant performanceimprovement, even when incorporated with others, whichencourages further efforts to substantiate the proposedwork. �

R E F E R E N C E S[1] Framework and Overall Objectives of the Future

Development of IMT for 2020 and Beyond,document Rec. M.2083, ITU-R, Sep. 2015.[Online]. Available:https://www.itu.int/rec/R-REC-M.2083

[2] “5G and e-health,” 5GPPP Assoc., White Paper, Oct.2015. [Online]. Available: https://5gppp.eu/wp-content/uploads/2016/02/5G-PPP-White-Paper-on-eHealth-Vertical-Sector.pdf

[3] “5G automotive vision,” 5GPPP Assoc., WhitePaper, Oct. 2015. [Online]. Available:https://5gppp.eu/wp-content/uploads/2014/02/5G-PPP-White-Paper-on-Automotive-Vertical-Sectors.pdf

[4] “5G and the factories of the future,” 5GPPP Assoc.,White Paper, Oct. 2015. [Online]. Available:https://5g-ppp.eu/wp-content/uploads/2014/02/5G-PPP-White-Paper-on-Factories-of-the-Future-Vertical-Sector.pdf

[5] M. Luvisotto, Z. Pang, and D. Dzung, “Ultra highperformance wireless control for criticalapplications: Challenges and directions,” IEEETrans. Ind. Informat., vol. 13, no. 3,pp. 1448–1459, Jun. 2017.

[6] “5G for mission critical communication,” Nokia,Espoo, Finland, White Paper, 2016. [Online].Available: https://networks.nokia.com/innovation/5g

[7] V. Prasad et al., Tactile Usecase Summary, IEEEStandard P1918.1, VKKV-17, Tactile Internet,Mar. 2017.

[8] V. Prasad et al., Consolidated Usecase for the TactileInternet, IEEE Standard P1918.1, PHCR-17-4,Tactile Internet, May 2017.

[9] P. Dario, B. Hannaford, and A. Menciassi, “Smartsurgical tools and augmenting devices,” IEEETrans. Robot. Autom., vol. 19, no. 5, pp. 782–792,

Oct. 2003.[10] J. Arata et al., “Impact of network time-delay and

force feedback on tele-surgery,” Int. J. Comput.Assist. Radiol. Surg., vol. 3, nos. 3–4,pp. 371–378, Sep. 2008.

[11] G. Fettweis and S. Alamouti, “5G: Personal mobileInternet beyond what cellular did to telephony,”IEEE Commun. Mag., vol. 52, no. 2, pp. 140–145,Feb. 2014.

[12] E. Steinbach and Q. Liu, “Use case onteleoperation over the tactile Internet,” IEEEStandard P1918.1, SL-16-3-r0, Tactile Internet,Jul. 2016.

[13] Study on Latency Reduction Techniques for LTE(Release 14), document TR 36.881, 3GPP TSGRAN, May 2016.

[14] On URLLC Design Principles, documentR1-1609634, 3GPP TSG RAN WG1 Meeting#86bis, Oct. 2016.

[15] Discussion on URLLC Design Aspects, documentR1-1610366, 3GPP TSG RAN WG1 Meeting#86bis, Intel Corporation, Oct. 2016.

[16] URLLC Numerology and Frame Structure Design,document R1-1610123, 3GPP TSG RAN WG1Meeting #86bis, Qualcomm, Oct. 2016.

[17] Study on Scenarios and Requirements for NextGeneration Access Technologies (Release 14)document TR 38.913, 3GPP TSG RAN, May 2017.

[18] Study on New Radio Access Technology; PhysicalLayer Aspects (Release 14), document TR 36.802,3GPP TSG RAN, Mar. 2017.

[19] Study on New Radio Access Technology; RadioInterface Protocol Aspects (Release 14), documentTR 36.881, 3GPP TSG RAN, Mar. 2017.

[20] UL URLLC Multiplexing Considerations, documentR1-1611657, 3GPP TSG RAN WG1 Meeting #87,Huawei and Hsillicon, Nov. 2016.

[21] Scheduling and Support for Service Multiplexing,document R1-1612316, 3GPP TSG RAN WG1Meeting #87, InterDigital Communications,Nov. 2016.

[22] Multiplexing of eMBB and URLLC Frame Structures,document R1-1612316, 3GPP TSG RAN WG1Meeting #87, ZTE Corporation Nov. 2016.

[23] NR Paging Design, document R1-1611689, 3GPPTSG RAN WG1 Meeting #87, LG Electronics,Nov. 2016.

[24] Grant-Free to Grant-Based Switching for URLLC,document R1-1611253, 3GPP TSG RAN WG1Meeting #87, Nokia and Alcatel-Lucent ShanghaiBell, Nov. 2016.

[25] Cyclic Delay-Doppler Shifted M-Sequences for NewRadio PRACH, document R1-1611296, Nokia andAlcatel-Lucent Shanghai Bell, 3GPP TSG RANWG1 Meeting #87, Nov. 2016.

[26] E. Dahlman, S. Parkvall, and J. Skold, 4G:LTE/LTE-Advanced for Mobile Broadband, 1st ed.Amsterdam, The Netherlands: Academic, 2011.

[27] Control Channel Design for URLLC, documentR1-1611221, 3GPP TSG RAN WG1 Meeting #87,Huawei and Hsillicon, Nov. 2016.

[28] HARQ Design for Uplink Grant-Free Transmission,document R1-1717857, 3GPP TSG RAN WG1Meeting 90bis, Lenovo and Motorola Mobility,Oct. 2017.

[29] J. Yeo, S. Park, J. Oh, Y. Kim, and J. Lee, “Partialretransmission scheme for HARQ enhancement in5G wireless communications,” in Proc. IEEEGlobecom Workshops, Dec. 2017, pp. 1–5.

[30] T. L. Marzetta, “Noncooperative cellular wirelesswith unlimited numbers of base station antennas,”IEEE Trans. Wireless Commun., vol. 9, no. 11,pp. 3590–3600, Nov. 2010.

[31] Y.-G. Lim, C.-B. Chae, and G. Caire, “Performance



analysis of massive MIMO for cell-boundaryusers,” IEEE Trans. Wireless Commun., vol. 14,no. 12, pp. 6827–6842, Dec. 2015.

[32] G. Hasslinger and O. Hohlfeld, “The Gilbert-Elliottmodel for packet loss in real time services on theInternet,” in Proc. 14th GI/ITG Conf.-Meas.,Modelling Eval. Comput. Commun. Syst., Mar.2008, pp. 1–15.

[33] S. Michalak, “Guidelines for immersive virtualreality experiences,” White Paper, Jul. 2017.[Online]. Available:https://software.intel.com/en-us/articles/guidelines-for-immersive-virtual-reality-experiences

[34] “Making immersive virtual reality possible inmobile,” Qualcomm, San Diego, CA, USA, WhitePaper, Apr. 2016. [Online]. Available:https://www.qualcomm.com/documents/whitepaper-making-immersive-virtual-reality-possible-mobile

[35] “Virtual reality/augumented reality,” CAICT andHuawei, White Paper, Sep. 2017. [Online].Available: http://www-file.huawei.com/-/media/CORPORATE/PDF/ilab/vr-ar-en.pdf

[36] S. Hirche and M. Buss, “Human-oriented controlfor haptic teleoperation,” Proc. IEEE, vol. 100,no. 3, pp. 623–647, Mar. 2012.

[37] M. Asada, “A report on RoboCup2017[competitions],” IEEE Robot. Autom. Mag.,vol. 24, no. 4, pp. 21–23, Dec. 2017.

[38] M. Dohler, “Will the tactile Internet globalize yourskill set?” IEEE ComSoc Technol. News, Jan. 2017.

[39] D. Lee, A. Franchi, H. I. Son, C. Ha, H. H. Bulthoff,and P. R. Giordano, “Semiautonomous hapticteleoperation control architecture of multipleunmanned aerial vehicles,” IEEE/ASME Trans.Mechatronics, vol. 18, no. 4, pp. 1334–1345,Aug. 2013.

[40] S. K. A. Hossain, A. S. M. M. Rahman, andA. El Saddik, “Measurements of multimodalapproach to haptic interaction in second lifeinterpersonal communication system,” IEEE Trans.Instrum. Meas., vol. 60, no. 11, pp. 3547–3558,Nov. 2011.

[41] L.-P. Morency, “Modeling human communicationdynamics [social sciences],” IEEE Signal Process.Mag., vol. 27, no. 5, pp. 112–116, Sep. 2010.

[42] J. Kim et al., “Grant-free multiple access forultra-reliable low-latency communications in alarge-scale antenna system,” in Proc. Int. Conf. Inf.Commun. Technol. Converg. (ICTC), Oct. 2016,pp. 466–470.

[43] S. Hailu and M. Saily, “Hybrid paging and locationtracking scheme for inactive 5G UEs,” in Proc. Eur.Conf. Netw. Commun. (EuCNC), Jun. 2017,pp. 1–6.

[44] I. L. Da Silva, G. Mildh, M. Säily, and S. Hailu, “Anovel state model for 5G radio access networks,”in Proc. IEEE Int. Conf. Commun. Workshops (ICC),May 2016, pp. 632–637.

[45] Discussion of RRC States in NR, documentR2-163441, 3GPP TSG RAN WG2 Meeting #94,Nokia and Alcatel-Lucent Shanghai Bell, May2016.

[46] P. Rugeland et al., “Architectural enablers andconcepts for mm-Wave RAN integration,” WhitePaper, 5GPPP mmMAGIC, Mar. 2017. [Online].Available: https://bscw.5g-mmmagic.eu/pub/bscw.cgi/d187833/mmMAGIC_Architectural_enablers_mmWave_integration.pdf

[47] K. J. Choi and K. S. Kim, “Optimal semi-persistentuplink scheduling policy for large-scale antennasystems,” IEEE Access, vol. 5, pp. 22902–22915,Oct. 2017.

[48] J. J. Nielsen, R. Liu, and P. Popovski (Nov. 2017).“Ultra-reliable low latency communication(URLLC) using interface diversity.” [Online].Available: https://arxiv.org/abs/1711.07771

[49] B. M. Hochwald, T. L. Marzetta, and V. Tarokh,“Multiple-antenna channel hardening and itsimplications for rate feedback and scheduling,”IEEE Trans. Inf. Theory, vol. 50, no. 9,pp. 1893–1909, Sep. 2004.

[50] J. Kim and K. S. Kim, “Grant-free multiple accessusing LSAS for URLLC services,” IEEE Access, to be

published.[51] Discussion on Numerology Multiplexing for

Supporting Different Service Requirements,document R1-1609145, 3GPP TSG RAN WG1Meeting #86bis, NEC, Oct. 2016.

[52] Y.-G. Lim et al., “Waveform multiplexing for newradio: Numerology management and 3Devaluation,” IEEE Wireless Commun., to bepublished.

[53] M. Chung, M. S. Sim, J. Kim, D. K. Kim, andC.-B. Chae, “Prototyping real-time full duplexradios,” IEEE Commun. Mag., vol. 53, no. 9,pp. 56–63, Sep. 2015.

[54] M. Chung, M. S. Sim, D. K. Kim, and C.-B. Chae,“Compact full duplex MIMO radios in D2Dunderlaid cellular networks: From system designto prototype results,” IEEE Access, vol. 5,pp. 16601–16617, 2017.

[55] R. Li, Y. Chen, G. Y. Li, and G. Liu, “Full-duplexcellular networks,” IEEE Commun. Mag., vol. 55,no. 4, pp. 184–191, Apr. 2017.

[56] A. A. Zaidi et al., “Waveform and numerology tosupport 5G services and requirements,” IEEECommun. Mag., vol. 54, no. 11, pp. 90–98,Nov. 2016.

[57] X. Zhang, M. Jia, L. Chen, J. Ma, and J. Qiu,“Filtered-OFDM—Enabler for flexible waveform inthe 5th generation cellular networks,” in Proc.IEEE Global Commun. Conf. (GLOBECOM),Dec. 2015, pp. 1–6.

[58] J. Abdoli, M. Jia, and J. Ma, “Filtered OFDM: Anew waveform for future wireless systems,” inProc. IEEE Int. Workshop SPAWC, Jun. 2015,pp. 66–70.

[59] f-OFDM Scheme and Filter Design, documentR1-165425, 3GPP TSG RAN WG1 Meeting #85,Huawei and Hsillicon, May 2016.

[60] G. Fettweis, M. Krondorf, and S. Bittner,“GFDM—Generalized frequency divisionmultiplexing,” in Proc. IEEE 69th Veh. Technol.Conf. (VTC Spring), Apr. 2009, pp. 1–4.

[61] N. Michailow et al., “Generalized frequencydivision multiplexing for 5th generation cellularnetworks,” IEEE Trans. Commun., vol. 62, no. 9,pp. 3045–3061, Sep. 2014.

[62] H. Lin, “Flexible configured OFDM for 5G airinterface,” IEEE Access, vol. 3, pp. 1861–1870,2015.

[63] B. Lim and Y.-C. Ko, “SIR analysis of OFDM andGFDM waveforms with timing offset, CFO, andphase noise,” IEEE Trans. Wireless Commun.,vol. 16, no. 10, pp. 6979–6990, Oct. 2017.

[64] B. Lim and Y.-C. Ko, “Optimal receiver filter forGFDM with timing and frequency offsets in uplinkmultiuser systems,” in Proc. IEEE WirelessCommun. Netw. Conf. (WCNC), Apr. 2018, pp. 1–6.

[65] K. J. Jeon and K. S. Kim, “Accumulate repeataccumulate check accumulate codes,” IEEE Trans.Commun., vol. 65, no. 11, pp. 4585–4599,Nov. 2017.

[66] K. J. Jeon and K. S. Kim, “Rate-compatible ARACAcodes,” Electron. Lett., vol. 54, no. 6, pp. 398–400,Mar. 2018.

[67] M. Morelli, C. C. J. Kuo, and M. O. Pun,“Synchronization techniques for orthogonalfrequency division multiple access (OFDMA): Atutorial review,” Proc. IEEE, vol. 95, no. 7,pp. 1394–1427, Jul. 2007.

[68] I. S. Gaspar, L. L. Mendes, N. Michailow, andG. Fettweis “A synchronization technique forgeneralized frequency division multiplexing,”EURASIP J. Adv. Signal Process., vol. 2014, no. 1,p. 67, Dec. 2014.

[69] M. Morelli, “Timing and frequencysynchronization for the uplink of an OFDMAsystem,” IEEE Trans. Commun., vol. 52, no. 2,pp. 296–306, Feb. 2004.

[70] J. J. V. D. Beek et al., “A time and frequencysynchronization scheme for multiuser OFDM,”IEEE J. Sel. Areas Commun., vol. 17, no. 11,pp. 1900–1914, Nov. 1999.

[71] S. Manohar, D. Sreedhar, V. Tikiya, andA. Chockalingam, “Cancellation of multiuserinterference due to carrier frequency offsets inuplink OFDMA,” IEEE Trans. Wireless Commun.,

vol. 6, no. 7, pp. 2560–2571, Jul. 2007.[72] D. Huang and K. B. Letaief, “An

interference-cancellation scheme for carrierfrequency offsets correction in OFDMA systems,”IEEE Trans. Commun., vol. 53, no. 7,pp. 1155–1165, Jul. 2005.

[73] Evolved Universal Terrestrial Radio Access(E-UTRA); Multiplexing and Channel Coding,document TS 36.212, Rev. 14.4.0, 3GPP, Sep.2017.

[74] T. Richardson, “Error floors of LDPC codes,” inProc. 41st Annu. Allerton Conf. Commun., ControlComput., Sep. 2003, pp. 1426–1435.

[75] D. Divsalar, S. Dolinar, and J. Thorpe,“Accumulate-repeat-accumulate-accumulate-codes,” in Proc. IEEE Veh. Technol. Conf. (VTC),Sep. 2004, pp. 1271–1275.

[76] D. Divsalar, C. Jones, S. Dolinar, and J. Thorpe,“Protograph based LDPC codes with minimumdistance linearly growing with block size,” in Proc.IEEE Global Commun. Conf. (GLOBECOM), Nov.2005, p. 5.

[77] Preliminary Evaluation Results on New ChannelCoding Scheme for NR, document R1-164812,3GPP TSG RAN WG1 Meeting #85, Samsung,May 2016.

[78] T. V. Nguyen and A. Nosratinia, “Rate-compatibleshort-length protograph LDPC codes,” IEEECommun. Lett., vol. 17, no. 5, pp. 948–951, May2013.

[79] T. Y. Chen, K. Vakilinia, D. Divsalar, andR. D. Wesel, “Protograph-based raptor-like LDPCcodes,” IEEE Trans. Commun., vol. 63, no. 5,pp. 1522–1532, May 2015.

[80] K. J. Jeon and K. S. Kim, “High-speedencoder/decoder structure for ARACA codes,”IEEE Trans. Commun., to be published.

[81] V. R. Cadambe and S. A. Jafar, “Interferencealignment and degrees of freedom of the K-userinterference channel,” IEEE Trans. Inf. Theory,vol. 54, no. 8, pp. 3425–3441, Aug. 2008.

[82] M. Yang, S.-W. Jeon, and D. K. Kim, “Degrees offreedom of full-duplex cellular networks withreconfigurable antennas at base station,” IEEETrans. Wireless Commun., vol. 16, no. 4,pp. 2314–2326, Apr. 2017.

[83] Study on 3D Channel Model for LTE, document TR36.873, 3GPP TSG RAN, Jun. 2015.

[84] S. Y. Seidel and T. S. Rappaport, “Site-specificpropagation prediction for wireless in-buildingpersonal communication system design,” IEEETrans. Veh. Technol., vol. 43, no. 4, pp. 879–891,Nov. 1994.

[85] V. Erceg, S. J. Fortune, J. Ling, A. J. Rustako, andR. A. Valenzuela, “Comparisons of acomputer-based propagation prediction tool withexperimental data collected in urban microcellularenvironments,” IEEE J. Sel. Areas Commun.,vol. 15, no. 4, pp. 677–684, May 1997.

[86] R. Bhagavatula, R. W. Heath, Jr., S. Vishwanath,and A. Forenza, “Sizing up MIMO arrays,” IEEEVeh. Technol. Mag., vol. 3, no. 4, pp. 31–38,Dec. 2008.

[87] Study on Channel Model for Frequency SpectrumAbove 6 GHz, document TR 38.900, 3GPP TSGRAN, Jun. 2018.

[88] METIS Channel Models, document ICT-317669METIS Project Deliverable D1.4 v.3, METIS, Jun.2015.

[89] J. Medbo, “Radio propagation modeling for 5Gmobile and wireless communications,” IEEECommun. Mag., vol. 54, no. 6, pp. 144–151,Jun. 2016.

[90] Y.-G. Lim (Nov. 2017). “Map-basedmillimeter-wave channel models: An overview,guidelines, and data.” [Online]. Available:https://arxiv.org/abs/1711.09052

[91] J. Jang et al., “Smart small cell with hybridbeamforming for 5G: Theoretical feasibility andprototype results,” IEEE Wireless Commun.,vol. 23, no. 6, pp. 124–131,Dec. 2016.

[92] T. Oh, Y. G. Lim, C. B. Chae, and Y. Lee,“Dual-polarization slot antenna with highcross-polarization discrimination for indoor



small-cell MIMO systems,” IEEE Antennas WirelessPropag. Lett., vol. 14, pp. 374–377, 2015.

[93] T. Kwon, Y.-G. Lim, B.-W. Min, and C.-B. Chae, “RFlens-embedded massive MIMO systems:Fabrication issues and codebook design,” IEEETrans. Microw. Theory Techn., vol. 64, no. 7,pp. 2256–2271, Jul. 2016.

[94] M. S. Sim, M. Chung, D. Kim, J. Chung, D. K. Kim,and C.-B. Chae, “Nonlinear self-interferencecancellation for full-duplex radios: From link-leveland system-level performance perspectives,” IEEECommun. Mag., vol. 55, no. 9, pp. 158–167,Jun. 2017.

[95] B. Ai et al., “On indoor millimeter wave massiveMIMO channels: Measurement and simulation,”IEEE J. Sel. Areas Commun., vol. 35, no. 7,pp. 1678–1690, Jul. 2017.

[96] V. Degli-Esposti et al., “Ray-tracing-basedMM-wave beamforming assessment,” IEEE Access,vol. 2, pp. 1314–1325, Mar. 2014.

[97] H. Kim, Y.-G. Lim, C.-B. Chae, and D. Hong(Mar. 2017). “Multiple access for 5G new radio:Categorization, evaluation, and challenges.”[Online]. Available:https://arxiv.org/abs/1703.09042

[98] Y.-G. Lim, Y. J. Cho, T. Oh, Y. Lee, and C.-B. Chae,

“Relationship between cross-polarizationdiscrimination (XPD) and spatial correlation inindoor small-cell MIMO systems,” IEEE WirelessCommun. Lett., vol. 7, no. 4, pp. 654–657,Aug. 2018.

[99] Y. J. Cho, G.-Y. Suk, B. Kim, D. K. Kim, andC.-B. Chae, “RF lens-embedded antenna array formmWave MIMO: Design and performance,” IEEECommun. Mag., vol. 56, no. 7, pp. 42–48,Jul. 2018.

[100] K. L. Ryu et al., “MIESM based system-levelsimulator for 3D realistic MIMO channel,” in Proc.KICS Winter Conf., Jan. 2018.

A B O U T T H E A U T H O R S

Kwang Soon Kim (Senior Member, IEEE)received the B.S. (summa cum laude),M.S.E., and Ph.D. degrees in electrical engi-neering from Korea Advanced Institute ofScience and Technology (KAIST), Daejeon,South Korea, in 1994, 1996, and 1999,respectively.From March 1999 to March 2000, he was

with the Department of Electrical and Com-puter Engineering, University of California at San Diego, La Jolla,CA, USA, as a Postdoctoral Researcher. From April 2000 to Feb-ruary 2004, he was with the Mobile Telecommunication ResearchLaboratory, Electronics and Telecommunication Research Institute,Daejeon, South Korea, as a Senior Member of Research Staff.Since March 2004, he has been with the Department of Electricaland Electronic Engineering, Yonsei University, Seoul, South Korea,where he is now a Professor. His research interests are generallyin signal processing, communication theory, information theory,and stochastic geometry applied to wireless heterogeneous cellularnetworks, wireless local area networks, wireless D2D networks andwireless ad hoc networks, and are focused on the new radio accesstechnologies for 5G.Prof. Kim served as an Editor of the Journal of the Korean Institute

of Communications and Information Sciences from 2006 to 2012, asthe Editor-in-Chief of the Journal of the Korean Institute of Commu-nications and Information Sciences from 2013 to 2016, as an Editorof the Journal of Communications and Networks since 2008, andas an Editor of the IEEE Transactions on Wireless Communicationsfrom 2009 to 2014. He was a recipient of the Postdoctoral Fellow-ship from the Korea Science and Engineering Foundation (KOSEF)in 1999. He received the Outstanding Researcher Award fromthe Electronics and Telecommunication Research Institute (ETRI)in 2002, the Jack Neubauer Memorial Award (best system paperaward, the IEEE Transactions on Vehicular Technology) from theIEEE Vehicular Technology Society in 2008, and the LG R&D Award:Industry-Academic Cooperation Prize, LG Electronics, in 2013.

Dong Ku Kim (Senior Member, IEEE)received the B.S. degree from KoreaAerospace University, Goyang, Gyeonggi,South Korea, in 1983, and the M.S. andPh.D. degree from the University of SouthernCalifornia, Los Angeles, CA, USA, in 1985 and1992, respectively.He worked on CDMA systems in the

cellular infrastructure group of Motorola,Fort Worth, TX, USA, in 1992. He has been a Professor in the Schoolof Electrical and Electronic Engineering, Yonsei University, Seoul,South Korea, since 1994. He is a Chair of the Executive Committeeof the 5G Forum.Prof. Kim received the Minister Award for Distinguished Service

for ICT R&D from MSIP in 2013, the Award of Excellence in theleadership of 100 Leading Core Technologies for Korea 2020 fromNAEK in 2013, and the Dr. Irwin Jacobs Academic AchievementAward 2016 from Qualcomm and KICS.

Chan-Byoung Chae (Senior Member, IEEE)received the Ph.D. degree from the Univer-sity of Texas at Austin, Austin, TX, USA, in2008.Currently, he is the Underwood Dis-

tinguished Professor at Yonsei University,Seoul, South Korea. Before joining Yonsei,he was with Bell Laboratories and HarvardUniversity.Dr. Chae was the recipient of the IEEE INFOCOM Best Demo

Award (2015), the IEEE SPMag Best Paper Award (2013), the IEEEComSoc Outstanding Young Researcher Award (2012), and the IEEEDaniel Noble Fellowship Award (2008). He serves/has served as anEditor for the IEEE Transactions on Wireless Communications, IEEECommunications Magazine, IEEE Wireless Communications Letters,IEEE Journal on Selected Areas in Communications, and the IEEETransactions on Molecular, Biological, and Multi-Scale Communica-tions.

Sunghyun Choi (Fellow, IEEE) received theB.S. (summa cum laude) and M.S. degreesfrom Korea Advanced Institute of Scienceand Technology, Daejeon, South Korea, in1992 and 1994, respectively, and the Ph.D.degree from The University of Michigan, AnnArbor, MI, USA, in 1999.Currently, he is a Professor at the Depart-

ment of Electrical and Computer Engineer-ing, Seoul National University (SNU), Seoul, SouthKorea. He coau-thored over 230 technical papers and holds over 160 patents in theareas of wireless/mobile networks and communications.Prof. Choi served on the editorial boards of the IEEE Transactions

on Mobile Computing, the IEEE Transactions on Wireless Communi-cations, and the IEEE Wireless Communications Magazine.

Young-Chai Ko (Senior Member, IEEE)received the B.Sc. degree in electricaland telecommunication engineering fromHanyang University, Seoul, South Korea, andthe M.S.E.E. and Ph.D. degrees in electri-cal engineering from the University of Min-nesota, Minneapolis, MN, USA, in 1999 and2001, respectively.He was with Novatel Wireless as a

Research Scientist in 2001. In 2001, he joined the Wireless Center,Texas Instruments, Inc., San Diego, CA, USA, as a Senior Engineer.He is currently a Professor with the School of Electrical Engineering,Korea University, Seoul, South Korea. His current research interestsare the performance analysis and the design of wireless communi-cation systems.



Jonghyun Kim (Student Member, IEEE)received the B.S. degree in electrical andelectronic engineering from Yonsei Univer-sity, Seoul, South Korea, in 2016, where he iscurrently working toward the combined M.S.and Ph.D. degree at the School of Electricaland Electronic Engineering.He has been involved in several national

R&D projects in Korea supported by ADD,NRF, and IITP since 2015. From 2013 to 2014, he was with Qual-comm CDMA Technologies Korea. His current research interestsinclude MIMO waveform transceiver design and radio access tech-nology for 5G and B5G communications.

Yeon-Geun Lim (Student Member, IEEE)received the B.S. degree in informationand communications engineering fromSungkyunkwan University, Seoul, SouthKorea, in 2012. Currently, he is workingtoward the Ph.D. degree at the School ofIntegrated Technology, Yonsei University,Seoul, South Korea.He was the recipient of Samsung Human-

tech Paper Award (2018). He has been involved in several industrialand national projects sponsored by Samsung Electronics, IITP, andKCA. His research interest includes massive MIMO, next-generationwaveforms, full-duplex, millimeter-wave technologies, and systemlevel simulation for 5G networks.

Minho Yang (Student Member, IEEE)received the B.S. degree from the School ofElectrical and Electronic Engineering, Yon-sei University, Seoul, South Korea, in 2012,where he is currently working toward thePh.D. degree.His research interests include network

information theory and wireless communica-tion systems.

Sundo Kim (Student Member, IEEE)received the B.S. degree from KoreaAdvanced Institute of Science and Technol-ogy, Daejeon, South Korea. He is currentlyworking toward the Ph.D. degree at theDepartment of Electrical and ComputerEngineering, Seoul National University,Seoul, South Korea.His research interests include latency

reduction algorithm and low latency protocol development for fifth-generation networks.

Byungju Lim received the B.S. and M.S.degrees in electrical engineering from KoreaUniversity, Seoul, South Korea, in 2015 and2017, respectively, where he is currentlyworking toward the Ph.D. degree at theSchool of Electrical Engineering.His current research interests include syn-

chronization, multicarrier systems, and sig-nal processing.

Kwanghoon Lee received the B.S. degreein electrical and electronic engineering fromYonsei University, Seoul, South Korea, in2016, where he is currently working towardthe combined M.S. and Ph.D. degree at theSchool of Electrical and Electronic Engineer-ing.He has been involved in several national

R&D projects in Korea supported by ADD,NRF, and IITP since 2016. His current research interests includeradio resource management, millimeter-wave beamforming, andrealistic system-level simulation for V2X networks.

Kyung Lin Ryu was born in Busan, SouthKorea, in 1995. He received the B.S. degreesin electronic and radio engineering fromKyungHee University, Yongin, South Korea,in 2017, and is currently working towardthe M.S. degree at the School of Electricaland Electronic Engineering, Yonsei Univer-sity, Seoul, South Korea.He has been involved in several national

R&D projects in Korea supported by NRF and IITP since 2017. Hisresearch interests include massive MIMO and grant-free multipleaccess.


Documents

Ultrareliable and Low-Latency Communication Techniques for …dcl.yonsei.ac.kr › wordpress › wp-content › uploads › publi › ... · 2019-01-28 · Kim et al.: Ultrareliable