the road to 5g wireless systems: resource allocation for nomawingn/papers/slides_jan_2017.pdf ·...

The Road to 5G Wireless Systems:Resource Allocation for NOMA

The Road to 5G Wireless Systems:Resource Allocation for NOMA

Derrick Wing Kwan NgThe University of New South Wales

School of Electrical and Telecommunications Engineering

Tsinghua University, Beijing Jan. 2017

OutlineOutline1 Introduction

Overview - 5G Communication SystemsOverview - Massive MIMOOverview - Energy HarvestingOverview - 5G NOMA

2 System Model for MC-NOMA3 Resource Allocation Problem Formulation

Performance MeasureProblem Formulation

4 Optimization SolutionOptimal SolutionSuboptimal Solution

5 Simulation Results6 Conclusions7 Future Work

Derrick Wing Kwan Ng – The Road to 5G Wireless Systems: Resource Allocation for NOMA1 /451 /45

1. Introduction1. Introduction Overview - 5G Communication SystemsOverview - 5G Communication Systems

Overview - 5G Communication SystemsOverview - 5G Communication Systems

Scalable across an extreme variation of requirements

4

Massive Internet of Things Mission-critical

control

Enhanced mobile broadband

ees MissMiss

controlcontrol

Deep coverageTo reach challenging locations

Ultra-low energy10+ years of battery life

Ultra-low complexity10s of bits per second

Ultra-high density1 million nodes per Km2

Extreme capacity10 Tbps per Km2

Extreme data ratesMulti-Gigabits per second

Deep awarenessDiscovery and optimization

Extreme user mobilityOr no mobility at all

Ultra-low latencyAs low as 1 millisecond

Ultra-high reliability


Key Technologies for 5GKey Technologies for 5G

There is no single technology which can fulfill all the goals....(goodand bad)

Table: Key Technologies for 5G.

Massive MIMO [2, 3] NOMA [4, 5, 6, 7]mmWave Communications Full Duplex Communications [8, 9]Base Station Caching [10] Mobile Edge Computing

Cloud-based Radio Access Networks Visible Light CommunicationEnergy Harvesting [11]–[19] D2D Communication

Derrick Wing Kwan Ng – The Road to 5G Wireless Systems: Resource Allocation for NOMA3/453 /45


AdvertisementAdvertisement

Key Technologies for 5G Wireless Systems, CambridgeUniversity Press, Apr. 30 2017

Figure: By Vincent W. S. Wong (Editor), Robert Schober (Editor),Derrick Wing Kwan Ng (Editor), Li-Chun Wang (Editor)



5G Emerging Technologies5G Emerging Technologies

Core technologies/methods for fulfilling the strengthenquality of service (QoS) requirements:

Multiple input multiple output (MIMO) [20, 21]Extra degrees of freedom in resource allocation(diversity and multiplexing Ñ high data rate)Artificial noise generation for degrading the channel ofeavesdroppers (communication security) [22]Information signal beamforming, zero forcing etc.(communication security)




Massive MIMO: 5G technology candidateThe use of a very large number of service antennas(e.g., hundreds or thousands) to serve multiple users(each equipped with small number of antennas or evensingle-antenna) simultaneously

spatial-division multiplexing such that the different datastreams occupy the same frequencies and times. A keyelement to performing wireless spatial multiplexing is anarray of independently-controlled antennas. Under line-of-sight propagation conditions, the data streams are carriedon focused beams of data. In a cluttered propagation envi-ronment, the data streams can arrive from many directionssimultaneously. The streams tend to reinforce each otherconstructively where desired, and interfere destructivelywhere they are unwanted. To carry out the multiplexing,the array needs to know the frequency response of thepropagation channels between each of its elements andeach of the users. This channel state information (CSI) isutilized in the precoding block (shown in the figure) andit is here that the data streams are mapped into the signalsthat drive each of the antennas. By increasing the numberof antennas, the beams can be focused more selectively tothe users.

...............................................

Demand for wireless throughput willalways grow, but the quantity ofavailable electromagnetic spectrum willnever increase.

The uplink operation of the Massive MIMO system,shown in Figure 2, is substantially the reverse of down-link operation. The users transmit data streams at thesame time and over the same frequencies. The antenna ar-ray receives the sum of the data streams as modified bytheir respective propagation channels, and the decodingoperation, again utilizing CSI, untangles the received sig-nals to produce the individual data streams.

Portions of the preceding description of MassiveMIMO also apply to a generic Multi-User MIMO system.What distinguishes Massive MIMO is that certain activi-ties make it a scalable technology. The multiplexing anddemultiplexing operations are accomplished using directlymeasuredVrather than assumedVchannel characteristics.(In contrast, sectorization as well as open-loop angle-selective beam-forming assume line-of-sight propagationconditions, and as explained later, are ultimately not scal-able.) If measured channel responses are utilized, increas-ing the number of antennas always improves theperformance of the system irrespective of the noisiness ofthe measurements. The benefits of growing the numberof service antennas relative to the number of activeusers include greater selectivity in transmitting and re-ceiving the data streams, which leads to greater through-put, a reduction in the required radiated power, effectivepower control that provides uniformly good servicethroughout the cell, and greater simplicity in signalprocessing.

The paper is organized as follows. We begin with anoverview of Point-to-Point MIMO, and explain in particu-lar why it is not a scalable technology. Next, we considerthe theory of Multi-User MIMO, which has fundamentaladvantages over Point-to-Point MIMO, but which in itsoriginal conception also is not scalable. We then provide adetailed description of Massive MIMO: why it is scalable,how CSI is acquired, the simplest effective types of multi-plexing and demultiplexing, power control, and the ulti-mate limitations of Massive MIMO. We follow with acase study of a dense urban macrocellular Massive MIMOsystem in which each cell has a 64-antenna base stationthat serves 18 users. Despite the modest size of this sys-tem, it has amazingly good performance, particularly inthat it can provide uniformly good service throughout thecell, even at the cell edges. We conclude with a brief

FIGURE 1. Downlink operation of a Massive MIMO link. The antenna array selectively transmits a multiplicity of data streams, all occupying thesame time/frequency resources, so that each user receives only the data stream that it intended for him.

Bell Labs Technical Journal · VOLUME 20

12


1. Introduction1. Introduction Overview - Massive MIMOOverview - Massive MIMO


Massive MIMO: Transmitter equipped with hundredsantennas serve multiple (e.g. single antenna)receivers [2, 3, 23]

Shift the signal processing burden from the receiversto transmitterAllow simple design and cheap receiverAchieve asymptotically optimal performance by usingsimple precoding design

Table: List of potential research problems for massive MIMO

Hardware impairment Precoding designPilot contamination Energy efficiency

FDD vs TDD Co-located versus Distributed


1. Introduction1. Introduction Overview - Energy HarvestingOverview - Energy Harvesting

5G Energy Harvesting5G Energy Harvesting

Ubiquitous and Self-sustainable NetworksTechnologies

Conventional energy harvesting (scavenging): Collectenergy from natural renewable energy sources such assolar, wind, and geothermal heatAdvantages: Self- substantiable networkTechnical challenges (engineering problems): Timevarying availability of the energy generated fromrenewable energy sources Ñ Perpetual but intermittentenergy supply Ñ Unstable communication service[11, 12]




Hybrid powered base station networks

RRH 4

Solar pannel

RRH 3

Central Processor

Bus

Diesel generator

Power line

Backhaul link

Wind turbine

RRH 2

Backhaul link 1

Backhaul link 2

Backhaul link 3

Backhaul link 4Wind turbine

RRH 1

Point of common coupling

Bus Bus

Bus Bus

Solar pannel



Overview - Energy HarvestingOverview - Energy Harvesting

“New" energy harvesting technology: RF-basedEnergy Harvesting/ Wireless powered communications[13]–[19]

Collect energy from background radio frequency (RF)electromagnetic (EM) waves from ambient transmitters

Major Applications: RFID, body area networks, wirelesssensor networks, Machine-to-Machine (M2M)communications, Internet of things (IoT), etc [25, 26].




Transmitter

Informa

tion be

am 1Information receiver 1

Information receiver 2

Energy harvesting receiver 2

Information beam 2

Energy harvesting receiver 1

Table: List of potential research problems for massive MIMO

Resource allocation Protocol designModel design Energy efficiency optimization


1. Introduction1. Introduction Overview - 5G NOMAOverview - 5G NOMA

Overview - NOMAOverview - NOMA

In 4G communication systems, orthogonal frequency divisionmultiple access (OFDMA):

A wide band signal is divided into many narrow bandsubcarriers (e.g. 64Ï2048) and each subcarrier is assignedto at most one user

Channel equalization is simplified

Provide frequency diversity and multiuser diversity

f2 f3

f4 f5

Subcarrier spacing Inte−carrier inteferenceeliminated

f1 f

C



Motivation for NOMAMotivation for NOMA

OFDMA: High flexibility in resource allocation:

The Physical Resource Block (PRB) is the basic unit of allocation.

12 subcarriers in frequency (= 180 kHz)1 subframe in time (1ms = 14 OFDM symbols)Multiple resource blocks can be allocated to a user in agiven subframeTotal number of RBs depends on the operating bandwidth

time frequ

ency

OFDMA

block.pdf block.pdf

time

frequency

72subcarriers

6resourceblocks

radio frame = 10 ms

subframe (1 ms)slot (0.5 ms)

12subcarriers

Resource element

Resource block (RB) = 12 subcarriers x 0,5 msDerrick Wing Kwan Ng – The Road to 5G Wireless Systems: Resource Allocation for NOMA13 /4513 /45



However, traditional multicarrier orthogonal multipleaccess (MC-OMA) systems still underutilize the spectralresources

Subcarriers are allocated exclusively to one user toavoid multiuser interferenceSubcarriers may be assigned exclusively to a user withpoor channel quality to ensure fairnessOrthogonality cannot be always maintained

Hardware imperfectionDoppler shift




To overcome the aforementioned shortcomings,non-orthogonal multiple access (NOMA) has been recentlyproposed

Multiplexing multiple users on the same frequencyresourcePairing users enjoying good channel conditions withusers suffering from poor channel conditionsNOMA is enabled by successive interferencecancellation (SIC) at receivers



Channel capacity comparison of OMA and NOMA in an AWGN channelChannel capacity comparison of OMA and NOMA in an AWGN channel

NOMA achieves larger multi-user capacity comparedto OMA.

IEEE Communications Magazine • September 2015 75

later. The design principles, key features, advan-tages and disadvantages of existing dominantNOMA schemes are discussed and compared.More importantly, although NOMA can provideattractive advantages, some challenging prob-lems should be solved, such as advanced trans-mitter design and the trade-off betweenperformance and receiver complexity. Thus,opportunities and research trends are highlight-ed to provide some insights on the potentialfuture work for researchers in this field. In addi-tion, unlike the conventional way of designing aspecific multiple access scheme separately andindividually, we propose the concept of softwaredefined multiple access (SoDeMA), in whichseveral candidates among multiple accessschemes can be adaptively configured to satisfydifferent requirements of diverse services andapplications in future 5G networks. Finally, con-clusions are drawn.

FEATURES OF NOMAIn conventional OMA schemes, multiple usersare allocated with radio resources which areorthogonal in time, frequency, or code domain.Ideally, no interference exists among multipleusers due to the orthogonal resource allocationin OMA, so simple single-user detection can beused to separate different users’ signals. Theo-retically, it is known that OMA cannot alwaysachieve the sum-rate capacity of multiuser wire-less systems [10]. Apart from that, in convention-al OMA schemes, the maximum number ofsupported users is limited by the total amountand the scheduling granularity of orthogonalresources.

Recently, NOMA has been investigated todeal with the problems of OMA as mentionedabove. Basically, NOMA allows controllableinterferences by non-orthogonal resource alloca-tion with the tolerable increase in receiver com-plexity. Compared to OMA, the main advantagesof NOMA include the following.

Improved spectral efficiency: According tothe multi-user capacity analysis in the pioneeringwork [10], Fig. 1 shows the channel capacitycomparison of OMA and NOMA, where twousers in the additive white Gaussian noise(AWGN) channel are considered as an examplewithout loss of generality. Figure 1a shows thatthe uplink NOMA is able to achieve the capacitybound, while OMA schemes are in general sub-optimal except at point C. However, at this opti-mal point, the user throughput fairness is quitepoor when the difference of the received powersof the two users is significant, as the rate of theweak user is much lower than that of the stronguser. In the downlink, Fig. 1b shows that theboundary of rate pairs of NOMA is outside ofthe OMA rate region in general. In multi-pathfading channels with intersymbol interference(ISI), although OMA could achieve the sumcapacity in the downlink, NOMA is optimalwhile OMA is strictly suboptimal if channel stateinformation (CSI) is only known at the mobilereceiver [10].

Massive connectivity: The non-orthogonalresource allocation in NOMA indicates that thenumber of supported users or devices is not

strictly limited by the amount of availableresources and their scheduling granularity.Therefore, NOMA can accommodate signifi-cantly more users than OMA by using non-orthogonal resource allocation; for example,MUSA can still achieve reasonably good perfor-mance when the overloading is 300 percent [8].

Low transmission latency and signaling cost:In conventional OMA with grant-based trans-mission, a user has to send a scheduling requestto the base station (BS) at first. Then, based onthe received request, the BS performs schedulingfor the uplink transmission and sends a grantover the downlink channel. This procedureresults in large latency and high signaling cost,which becomes worse or even unacceptable inthe scenario of massive connectivity anticipatedfor 5G. In contrast, such dynamic scheduling isnot required in some uplink schemes of NOMA,rendering a grant-free uplink transmission thatcan drastically reduce the transmission latencyand signaling overhead.

Due to the potential advantages above,NOMA has been actively investigated as apromising technology for 5G. In the next section,existing dominant NOMA schemes are discussedand compared in detail.

DOMINANT NOMA SOLUTIONSIn this section, we discuss dominant NOMAschemes by grouping them into two categories:power domain multiplexing and code domainmultiplexing. Power domain multiplexing meansthat different users are allocated different powerlevels according to their channel conditions toobtain the maximum gain in system perfor-mance. Such power allocation is also beneficialto separate different users, where successiveinterference cancellation (SIC) is often used tocancel multi-user interference. In this article,power domain multiplexing is applied only todownlink NOMA. Code domain multiplexing issimilar to CDMA or multicarrier CDMA (MC-CDMA), that is, different users are assigned dif-ferent codes, and are then multiplexed over thesame time-frequency resources. The difference

Figure 1. Channel capacity comparison of OMA and NOMA in an AWGNchannel: a) uplink AWGN channel; b) downlink AWGN channel.

NOMA

NOMA

OMA

A

C

B

Rate

of

user

1

Rate

of

user

1Rate of user 2

(a)Rate of user 2

(b)

OMA

Figure: Channel capacity comparison of OMA and NOMA in an AWGNchannel: (a) uplink AWGN channel; (b) downlink AWGN channel.



Uplink NOMAUplink NOMA

Transmission latency and signaling cost can bereduced in uplink NOMA

Allow multiple uplink users share the same radioresourceNo scheduling is requiredThe base station performs SIC

Base station

User 1

User 2

Scheduling request

Grant

Data

Delay

Scheduling request

Grant

Data

Delay

User 1

User 2

Data

Base station

OMA NOMA

Data


2. System Model for MC-NOMA2. System Model for MC-NOMA

System ModelSystem Model

Resource allocation for a multicarrier NOMA downlinksystem.

Base station

User 1

User 2

User 2User 1

Frequency

Power

User 1User 3User 4 User 3

Subcarrier 1 Subcarrier 2 Subcarrier 3

...

User 3User 4

K downlink users, NF subcarriersAll transceivers are equipped with a single-antenna.All downlink users can perform successive interferencecancellation (SIC)Each subcarrier can be allocated with at most two users.Power and subcarriers are available radio resources to bemanaged.


2. System Model for MC-NOMA2. System Model for MC-NOMA

System ModelSystem Model

Resource allocation design for MC-NOMA systems is morechallenging than for traditional MC-OMA systems

User pairing on each subcarrier is neededSIC ordering on each subcarrier is needed (which userperform SIC and which user do not)In next section, we will formulate the power andsubcarrier allocation design as a mixed-integernon-convex optimization problem


3. Resource Allocation Problem Formulation3. Resource Allocation Problem Formulation Performance MeasurePerformance Measure

Performance measurePerformance measure

Study the weighted throughput of user m and user non subcarrier iAssume |him|2 ≤ |hin|2

Weighted throughput of user m and user n on subcarrier i:

U im,n(pim, pin, sim,n)= sim,n[wm log2(1+ |him|2pim|him|2pin+σ2m)+wn log2 (1+|hin|2pinσ2n

)], (1)

where sim,n is subcarrier indicator and wm is the priority ofuser m.

Successful SIC can be performed, since

log2(1 + |hin|2pim|hin|2pin + σ2n ) ≥ log2(1 + |him|2pim|him|2pin + σ2m ). (2)Derrick Wing Kwan Ng – The Road to 5G Wireless Systems: Resource Allocation for NOMA20 /4520 /45

3. Resource Allocation Problem Formulation3. Resource Allocation Problem Formulation Problem FormulationProblem Formulation

Problem FormulationProblem Formulation

Problem: Maximization of the weighted systemthroughput

maximizep,s

NF∑i=1

K∑m=1

K∑n=1 s

im,n

[wmlog2(1+ |him|2pim|him|2pin+σ2m

)+wn log2 (1+|hin|2pinσ2n)]

s.t. C1:NF∑i=1

K∑m=1

K∑n=1s

im,n(pim + pin) ≤ Pmax,

C2: sim,n ∈ {0, 1}, ∀i,m,n,C3:

K∑m=1

K∑n=1s

im,n ≤ 1, ∀i,

C4: pim ≥ 0, ∀i,m, (3)Blue color = non-convex objective functionRed color = non-convex constraint

Even if the subcarrier allocation is given, NP-hard


4. Optimization Solution4. Optimization Solution Optimal SolutionOptimal Solution

Optimal SolutionOptimal Solution

In this section, we apply monotonic optimization toobtain the global optimal solutionMonotonic optimization can converge to the globaloptimal solution of non-convex optimization problemsby exploiting the monotonicity of the consideredproblem

The objective function is needed to be a increasingfunctionThe feasible set is needed to be a normal set (theconvex set belongs to normal sets)The global optimal point is attained on the upperboundaryApproach the global optimal solution by constructing asequence of polyblocks




Rewrite the weighted throughput in an equivalent form

U im,n(pim, pin, sim,n)=wm log2(1+ sim,nH impimsim,nH impin+1)+wn log2(1+sim,nH inpin)

=wm log2(1+ H imp̃im,n,mH imp̃im,n,n+1)+wn log2(1+H inp̃im,n,n)

= log2(uim,n)wm + log2(vim,n)wn , (4)where H im = |him|2σ2m , uim,n =1 + H imp̃im,n,mH imp̃im,n,n+1 , vim,n = 1 + H inp̃im,n,n,and p̃im,n,m = sim,npim.




Then, we define

fd(p̃) = {1 + H im(p̃im,n,m + p̃im,n,n), d = ∆,1 + H inp̃im,n,n, d = D/2 + ∆, (5)gd(p̃) = {1 + H imp̃im,n,n, d = ∆,1, d = D/2 + ∆, (6)

where ∆ = (i − 1)K2 + (m − 1)K + n and D = 2NFK2.We further define

z=[z1,. . .,zD]T=[u11,1,. . .,uNFK,K,v11,1,. . .,vNFK,K]T . (7)




Now, the original problem can be written as astandard monotonic optimization problem as:

Problem: Maximization of the weighted systemthroughput

maximizez

D∑d=1 log2(zd)µd

s.t. z ∈ Z, (8)

where the feasible set Z is given by

Z= {z | 1 ≤ zd ≤ fd(p̃)gd(p̃) , p̃ ∈ P, s ∈ S, ∀d},where P and S are the feasible sets spanned byconstraints C1, C3, and C4.




The equivalent monotonic optimization problem in (8)can be solved optimally via outer polyblockapproximation algorithm.

*

**

*ZZ

Z Z

(1)z(1)z(1)1z

(1)2z

(2)z(2)1z

(2)2z

(1)( )φ z

(1)2( )φ z

(1)1( )φ z

2z

1z

2z

1z

(3)z

2z2z

1z 1z

(2)2( )φ z

(2)1( )φ z

Figure: Illustration of the outer polyblock approximationalgorithm for D = 2. The red star is the optimal point on theboundary of the feasible set Z.




The proposed monotonic optimization based resourceallocation algorithm achieves the globally optimalsolution. However, the computational complexitygrows exponentially with D = 2NFK2.In next section, we propose a suboptimal algorithm toreduce complexity while achieving a close-to-optimalperformance.


4. Optimization Solution4. Optimization Solution Suboptimal SolutionSuboptimal Solution

Suboptimal SolutionSuboptimal Solution

The product term p̃im,n,m = sim,npim is an obstacle forefficient algorithm design.We adopt the big-M formulation to decompose theproduct terms. In particular, we impose the followingadditional constraints:

C5: p̃im,n,m ≤ Pmaxsim,n, ∀m,n, i,C6: p̃im,n,m ≤ pim, ∀m,n, i,C7: p̃im,n,m ≥ pim−(1− sim,n)Pmax, ∀m,n, i, andC8: p̃im,n,m ≥ 0, ∀m,n, i.




Then, we rewrite the non-convex integer constraintC2: sim,n ∈ {0, 1} in its equivalent form:C2a:

NF∑i=1

K∑m=1

K∑n=1s

im,n −

NF∑i=1

K∑m=1

K∑n=1(sim,n)2 ≤ 0 and

C2b: 0 ≤ sim,n ≤ 1, ∀m,n, i.However, the constraint C2a is still a non-convexconstraint.




In order to handle the non-convex constraint C2a, weincorporate it as an additive penalty function term intothe objective function :

Suboptimal resource allocation problem:

minimizep̃,s

NF∑i=1

K∑m=1

K∑n=1 − log2(1+ H

imp̃im,n,m

H imp̃im,n,n+1 )wm − log2(1+H inp̃im,n,n)wn+η( NF∑

i=1K∑

m=1K∑

n=1sim,n −

NF∑i=1

K∑m=1

K∑n=1(sim,n)2

)s.t. C1,C2b,C3–C8, (9)

where η � 1 is a large constant which acts as a penaltyfactor to penalize the objective function for any sim,n that isnot equal to 0 or 1.

Problem (9) is still non-convex due to its objectivefunction




Now, we rewrite problem (9) asminimizep̃,s

F (p̃)− G(p̃) + η(H(s)−M(s))s.t. C1,C2b,C3–C8, (10)

where

F (p̃)= NF∑i=1

K∑m=1

K∑n=1−wm log2(1+H im(p̃im,n,m + p̃im,n,n))

−wn log2(1+H inp̃im,n,n),G(p̃)= NF∑

i=1K∑

m=1K∑

n=1−wm log2(1 + H imp̃im,n,n),H(s)= NF∑

i=1K∑

m=1K∑

n=1sim,n,and M(s)= NF∑

i=1K∑

m=1K∑

n=1(sim,n)2.Derrick Wing Kwan Ng – The Road to 5G Wireless Systems: Resource Allocation for NOMA31 /4531 /45



The F (p̃), G(p̃), H(s), and M(s) are convex functions andthe problem in (10) belongs to the class of differenceof convex (d.c.) function programming.As a result, we can apply successive convexapproximation to obtain a local optimal solution.For any feasible point p̃(k) and s(k), we have thefollowing inequalities

G(p̃) ≥ G(p̃(k)) +∇p̃G(p̃(k))T (p̃− p̃(k)) and (11)M(s) ≥ M(s(k)) +∇sM(s(k))T (s − s(k)). (12)




Therefore, for any given p̃(k) and s(k), we can obtain anupper bound for (10) by solving the following convexoptimization problem:

Suboptimal resource allocation problem:minimizep̃,s

F (p̃)− G(p̃(k))−∇p̃G(p̃(k))T (p̃− p̃(k))+η(H(s)−M(s(k))−∇sM(s(k))T (s − s(k)))

s.t. C1,C2b,C3–C8, (13)

The optimization problem in (13) is convex and can besolved efficiently by standard convex program solversuch as CVX.Then, we employ an iterative algorithm to tighten theupper bound.


5. Simulation Results5. Simulation Results

Results – Simulation ParametersResults – Simulation Parameters

Table: System parameters

Carrier center frequency and system bandwidth 2.5 GHz and 5 MHzThe number of subcarriers, NF 64The bandwidth of each subcarrier 78 kHzUser noise power, σ2m −125 dBmBS antenna gain 10 dBiBaseline scheme 1: suboptimal power and subcarrierallocation scheme in [1]Baseline scheme 2: random subcarrier allocationschemeBaseline scheme 3: traditional MC-OMA scheme



Results – Average system throughput versus maximum transmit power at the BSResults – Average system throughput versus maximum transmit power at the BS

Maximum transmit power (dBm)30 32 34 36 38 40 42 44 46

Ave

rage

sys

tem

thro

ughp

ut (

bit/s

/Hz)

2.5

3

3.5

4

4.5

5

5.5

6

Proposed optimal schemeProposed suboptimal schemeBaseline scheme 1Baseline scheme 2Baseline scheme 3

Proposed suboptimal scheme

System throughputimprovement

Baseline scheme 1

Baseline scheme 3

Proposed optimal scheme

Baseline scheme 2

Figure: Average system throughput versus the maximum transmit power atbase station.



Results – Average system throughput versus number of usersResults – Average system throughput versus number of users

Number of users2 3 4 5 6 7 8 9 10

Ave

rage

sys

tem

thro

ughp

ut (

bit/s

/Hz)

3.5

4

4.5

5

5.5

6

6.5Proposed optimal schemeProposed suboptimal schemeBaseline scheme 1Baseline scheme 2Baseline scheme 3

Proposed optimal scheme

Baseline scheme 1

Baseline scheme 3

Baseline scheme 2

Proposed suboptimalscheme

System throughputimprovement

Figure: Average system throughput versus the number of users.


6. Conclusions6. Conclusions

ConclusionsConclusions

The resource allocation algorithm for MC-NOMAsystems was formulated with the objective tomaximize the weighted system throughputThe proposed problem is solved optimally by usingmonotonic optimization methodThe low-complexity suboptimal scheme is proposed toachieve close-to-optimal performanceSimulation results unveiled a significant improvementin system performance compared to conventionalMC-OMA system.


7. Future Work7. Future Work

Future WorkFuture Work

Employ NOMA transmission scheme in full-duplex (FD)systems to further improve system throughput.Investigate MISO-NOMA system where the BS isequipped with multiple antennas.Study resource allocation design for MC-NOMA systemwhere multiplex arbitrary number of users on eachsubcarrier.Study robust resource allocation design for NOMAsystems where the channel gains are imperfectlyknown.Study resource allocation design for uplink NOMAsystems.


8. References8. References

References[1] (2015) Qualcomm’s 5G Vision. [Online]. Available:

https://www.qualcomm.com/invention/technologies/5g

[2] T. Marzetta, “Noncooperative Cellular Wireless with UnlimitedNumbers of Base Station Antennas,” IEEE Trans. WirelessCommun., vol. 9, pp. 3590–3600, Nov. 2010.

[3] D. W. K. Ng, E. Lo, and R. Schober, “Energy-Efficient ResourceAllocation in OFDMA Systems with Large Numbers of Base StationAntennas,” IEEE Trans. Wireless Commun., vol. 11, pp. 3292 –3304,Sep. 2012.

[4] Y. Sun, D. W. K. Ng, Z. Ding, and R. Schober, “Optimal Joint Powerand Subcarrier Allocation for MC-NOMA Systems,” accepted forpresentation at IEEE Global Commun. Conf. 2016. [Online].Available: http://arxiv.org/abs/1503.06021

[5] ——, “Optimal Joint Power and Subcarrier Allocation for Full-DuplexMulticarrier Non-Orthogonal Multiple Access Systems,” acceptedfor presentation, TCOM. [Online]. Available:http://arxiv.org/abs/1503.06021


https://www.qualcomm.com/invention/technologies/5ghttp://arxiv.org/abs/1503.06021http://arxiv.org/abs/1503.06021


[6] Z. Wei, D. W. K. Ng, and J. Yuan, “Power-efficient resource allocationfor MC-NOMA with statistical channel state information,” CoRR, vol.abs/1607.01116, 2016. [Online]. Available:http://arxiv.org/abs/1607.01116

[7] Z. Wei, J. Yuan, D. W. K. Ng, M. Elkashlan, and Z. Ding, “A survey ofdownlink non-orthogonal multiple access for 5g wirelesscommunication networks,” CoRR, vol. abs/1609.01856, 2016.[Online]. Available: http://arxiv.org/abs/1609.01856

[8] Y. Sun, D. W. K. Ng, J. Zhu, and R. Schober, “Multi-ObjectiveOptimization for Robust Power Efficient and Secure Full-DuplexWireless Communication Systems,” IEEE Trans. Wireless Commun.,vol. 15, no. 8, pp. 5511–5526, Aug. 2016.

[9] D. W. K. Ng, Y. Wu, and R. Schober, “Power efficient resourceallocation for full-duplex radio distributed antenna networks,” IEEETrans. Wireless Commun., vol. 15, no. 4, pp. 2896–2911, 2016.

[10] L. Xiang, D. W. K. Ng, R. Schober, and V. W. Wong, “Cache-enabledphysical-layer security for video streaming in wireless networkswith limited backhaul,” CoRR, vol. abs/1612.01189, 2016. [Online].Available: https://arxiv.org/abs/1612.01189


http://arxiv.org/abs/1607.01116http://arxiv.org/abs/1609.01856https://arxiv.org/abs/1612.01189


[11] D. Ng, E. Lo, and R. Schober, “Energy-Efficient Resource Allocationin OFDMA Systems with Hybrid Energy Harvesting Base Station,”IEEE Trans. Wireless Commun., vol. 12, pp. 3412–3427, Jul. 2013.

[12] I. Ahmed, A. Ikhlef, D. Ng, and R. Schober, “Power Allocation foran Energy Harvesting Transmitter with Hybrid Energy Sources,”IEEE Trans. Wireless Commun., vol. 12, pp. 6255–6267, Dec. 2013.

[13] L. Varshney, “Transporting Information and EnergySimultaneously,” in Proc. IEEE Intern. Sympos. on Inf. Theory, Jul.2008, pp. 1612 –1616.

[14] D. W. K. Ng, E. S. Lo, and R. Schober, “Robust Beamforming forSecure Communication in Systems With Wireless Information andPower Transfer,” IEEE Trans. Wireless Commun., vol. 13, no. 8, pp.4599–4615, Aug 2014.

[15] P. Grover and A. Sahai, “Shannon Meets Tesla: WirelessInformation and Power Transfer,” in Proc. IEEE Intern. Sympos. onInf. Theory, Jun. 2010, pp. 2363 –2367.



[16] I. Krikidis, S. Timotheou, S. Nikolaou, G. Zheng, D. W. K. Ng, andR. Schober, “Simultaneous Wireless Information and PowerTransfer in Modern Communication Systems,” IEEE Commun. Mag.,vol. 52, no. 11, pp. 104–110, Nov. 2014.

[17] Z. Ding, C. Zhong, D. W. K. Ng, M. Peng, H. A. Suraweera,R. Schober, and H. V. Poor, “Application of Smart AntennaTechnologies in Simultaneous Wireless Information and PowerTransfer,” IEEE Commun. Magazine, vol. 53, no. 4, pp. 86–93, Apr.2015.

[18] R. Zhang and C. K. Ho, “MIMO Broadcasting for SimultaneousWireless Information and Power Transfer,” IEEE Trans. WirelessCommun., vol. 12, pp. 1989–2001, May 2013.

[19] X. Chen, Z. Zhang, H.-H. Chen, and H. Zhang, “Enhancing WirelessInformation and Power Transfer by Exploiting Multi-AntennaTechniques,” IEEE Commun. Magazine, no. 4, pp. 133–141, Apr.2015.

[20] D. W. K. Ng, E. S. Lo, and R. Schober, “Dynamic resource allocationin mimo-ofdma systems with full-duplex and hybrid relaying,” IEEETrans. Wireless Commun., vol. 60, no. 5, pp. 1291–1304, May 2012.



[21] J. Chen, X. Chen, W. H. Gerstacker, and D. W. K. Ng, “Resourceallocation for a massive mimo relay aided secure communication,”IEEE Transactions on Information Forensics and Security, vol. 11,no. 8, pp. 1700–1711, Aug. 2016.

[22] D. W. K. Ng, E. S. Lo, and R. Schober, “Secure Resource Allocationand Scheduling for OFDMA Decode-and-Forward Relay Networks,”IEEE Trans. Wireless Commun., vol. 10, pp. 3528–3540, Oct. 2011.

[23] Y. Wu, R. Schober, D. W. K. Ng, C. Xiao, and G. Caire, “Securemassive mimo transmission with an active eavesdropper,” IEEETrans. Inf. Theory, vol. 62, no. 7, pp. 3880–3900, Jul. 2016.

[24] E. Boshkovska, D. W. K. Ng, N. Zlatanov, and R. Schober,“Practical Non-Linear Energy Harvesting Model and ResourceAllocation for SWIPT Systems,” IEEE Commun. Lett., vol. 19, no. 12,pp. 2082–2085, Dec. 2015.

[25] F. Zhang, S. Hackworth, X. Liu, H. Chen, R. Sclabassi, and M. Sun,“Wireless Energy Transfer Platform for Medical Sensors andImplantable Devices,” in Annual Intern. Conf. of the IEEE Eng. inMed. and Biol. Soc., Sep. 2009, pp. 1045–1048.



[26] V. Chawla and D. S. Ha, “An Overview of Passive RFID,” IEEECommun. Magazine, vol. 45, pp. 11–17, Sep. 2007.



Q&A


TitleIntroductionOverview - 5G Communication Systems Overview - Massive MIMOOverview - Energy HarvestingOverview - 5G NOMA

System Model for MC-NOMAResource Allocation Problem FormulationPerformance MeasureProblem Formulation

Optimization SolutionOptimal SolutionSuboptimal Solution

Simulation ResultsConclusionsFuture Work

the road to 5g wireless systems: resource allocation for nomawingn/papers/slides_jan_2017.pdf ·...

Documents