（安達 091022IWSDA'09）Wireless Challenge.ppt [ … Fumiyuki Adachi Dept ... 1G 2G Era e t imedi...

IWSDA'09, Fukuoka International Congress Center, Fukuoka, Japan, 22 October, 2009.

Wireless Wireless ChallengeChallengeFumiyuki Adachi Fumiyuki Adachi

Dept. of Electrical and Communications Engineering,Tohoku University, Japany, pE-mail: [email protected]://www.mobile.ecei.tohoku.ac.jp

OUTLINEOUTLINEp Wireless Evolutionp Technical Issues for 4Gp Frequency-domain Signal Processing

2009/10/22 FA/Tohoku University2009/10/22 FA/Tohoku University 1

p Frequency domain Signal Processingp Distributed Antenna Network

Wireless EvolutionWireless Evolution

2009/10/22 FA/Tohoku University 2

Wireless EvolutionWireless Evolutionp In early 1980’s, communications systems changed from

fixed “point to point” to wireless “anytime anywhere”fixed point-to-point to wireless anytime, anywherecommunication.

p Cellular systems have evolved from narrowband network ofd 10kb t id b d t k f d 10Mbaround 10kbps to wideband networks of around 10Mbps.

p Now on the way to broadband networks of 1Gps.

Narrowbanda Wideband Broadband

2G1G

NarrowbandEra

e tim

edia

3G 4G100M 1Gbps

point-to-

i t

WidebandEra

BroadbandEra

~64kbps~2.4kbps

vice

typ

Mul

t

~2Mbps 100M~1Gbps

~14Mbps0GVoice only

point

3G LTE

50~100Mbps

BroadbandwirelessS

erv

ce

IMT-2000

HSDPA

Voice only 3G LTE

2009/10/22 FA/Tohoku University2009/10/22 FA/Tohoku University 3We are here1980 1990 2000 Year

Voi

2010

HSDPA

p Multiple access technique used in cellular systemsh b h d fhas been changed from FDMA to DS-CDMA.

n FDMA was used in 1G, TDMA in 2G, DS-CDMA in 3Gn OFDMA and SC-FDMA are used for downlink and uplinkn OFDMA and SC FDMA are used for downlink and uplink ,

respectively, in 3.9G(LTE)

1GOFDMA SC-FDMA

Frequency-domain processingTime-domain processing

FDMAf3f2f1

3.9G(LTE)

OFDMA，SC-FDMA

cyqu

en

cy Ex. 6.25kHz

Time

2GDS-CDMA3G

Freq

uen

c

# 3

Fre

uen

cy

TDMA

1 3 1 2 3

2GSingle-carrier

2

# 2# 3

Spreading code#1

Hz


Freq

u

Time

TimeEx. 20ms

Ex.

5M

H

Ex. 25kHz

3G Uses DS CDMA3G Uses DS-CDMAp In 3G systems, DS-CDMA with rake combiner is

d dadopted.n The user data of rate W is spread over B (=5MHz) by

multiplying a user specific spreading sequencemultiplying a user specific spreading sequence.n Various rate data can be transmitted just by changing

the spreading factor and channel coding rate.n Chip rate: 3.84Mcps

Spreading bandwidth

B 5MH

Data-modulated signal spectrum

fff fData rate R=8~384kbp

B=5MHz

Filter

g pW=16~240kHz

data Datamodulation

Data demod. Decoding

R=8~384kbps

3 84Mcps

Filter

Channelcoding )(1

SF


ReceiveddataSpreading

Sequencec(t)

c*(t)

3.84McpsMS BS

3G Systems Using W CDMA3G Systems Using W-CDMAp Data transfer rates in 2G systems are too slow for

retrieving rich information distributed in the Internetretrieving rich information distributed in the Internet.p 3G cellular systems are designed to offer cellular users a

significantly higher data-rate services using wideband DS-CDMA technology (5MHz bandwidth).

n Indoor: 2Mbpsn Pedestrian: 384kbpsn Pedestrian: 384kbpsn Vehicular: 144kbps IMT2000

Network

F Adachi M Sawahashi and H Suda “WidebandIndoors~2Mbps

Mobile~144kbps


F. Adachi, M. Sawahashi and H. Suda, “WidebandDS-CDMA for next generation mobilecommunications systems,” IEEE Commun. Mag.,vol. 36, pp. 56-69, Sept. 1998.

2MbpsPedestrian~384kbps

On-going Shift To 3G Systems(Japan)(Japan)p 3G services (~384kbps)

started in 2001 in Japan started in 2001 in Japan. p Growth rate of 3G systems

were very slow in the first few years but now it is really few years, but now it is really taking off.

p Total no. of cellular subscribers @end of Dec subscribers @end of Dec. 2008

n 105,825,200 (penetration:90 8 %)90.8 %)

p 2G (9,758,700)n PDC： 9,411,900

n cdmaOne: 346,800

p 3G (96,066,500)n 81.4 % of total

2009/10/22 FA/Tohoku University

n W-CDMA: 65,863,100n CDMA2000 1x: 30,203,400


Growing Wireless InternetGrowing Wireless Internetp Broadband multimedia services

n In line with the increasing popularity of Internet in fixednetworks, cellular systems are evolving from simplyproviding traditional voice communication services top gproviding broadband multimedia services throughInternet access.

I t t ll h @ d f D 2008p Internet cell phones @end of Dec. 2008 (source: TCA)

n Total cellular users: 105,825,200 (penetration 90.8%*)n Users connected to Internet: 90 173 100 (85 2%)n Users connected to Internet: 90,173,100 (85.2%)

p i-mode (DoCoMo): 48,149,500p Ezweb (au): 25,913,300p Yahoo Ke-tai (SoftBank): 16,061,000p Emnet: 49,300


* Japan population estimate:127.69m @Oct. 2005


4G Vision4G Visionp Broadband services: data rates for mobile

b b d h fservices may be up to 100 Mbps and those fornomadic services may be up to 1Gbps.Gi bit i l i t t t h l f thp Gigabit wireless: an important technology for therealization of 4G systems.

Beyond IMT-2000

High(60

New

Digital broadcast system

Inter-connection

New capabilities of systems

MobilityBeyond IMT 2000

(60~250km/h)

IMT2000 EnhancedIMT2000

MobileAccess

Enhancement

connection beyond IMT-2000

Low(Pede-strian

New nomadic/LocalArea wireless access


strianspeed) 1 10 100 1000

Peak useful data rate (Mb/s)ITU-R WP8F (Ottawa, June 2002) :illustration of Capabilities of IMT2000 and Systems Beyond

Before 4G There Will Be 3 9GBefore 4G, There Will Be 3.9Gp 3G band will be used

No available bandwidth of 100MHz (a hot matter of WRC 07)n No available bandwidth of 100MHz (a hot matter of WRC-07)n Present 3G bandwidth (1.25~20MHz) will be used to provide

much faster rate data servicesn Target: 100Mbps for downlink 50Mbps for uplinkn Target: 100Mbps for downlink, 50Mbps for uplink

p Difference from 3G wireless technologyn 3~3.5G：DS-CDMA�3.9G OFDMA?n Promising transmission performance and good commonalityn Promising transmission performance and good commonality

with WLAN

OFDMA

3.5G 3.9G

DS-CDMA(~14Mbps)

OFDMA,SC-FDMA(50~100Mbps)

3G band

OFDMA, SC-FDMA

4GWRC’07

N b d f 4G


SC-CDMA?(100M~1Gbps)

New band for 4G

Evolution Into 4GEvolution Into 4Gp 4G systems are required to provide much faster packet data

services of a peak rate of 100M 1Gbpsservices of a peak rate of 100M~1Gbps.p ITU allocated the spectrum for 4G systems in Dec. 2007.

p 450~470MHz (20MHz), 790~806MHz (16MHz), 2.3~2.4GMHz( ), ( ),(100MHz), 3.4G~3.6GHz Global use (200MHz)

Narrowband Wideband Broadband

2G~64kbps

1G~2.4kbpspe m

edia 3G

~2Mbps4G

100M~1Gbps

point-to-point

50 100Mbps

Era Era Era

Broadband

pp

vice

typ

Mul

tim p

~14Mbps0GVoice only

3G LTE

50~100Mbps

Broadbandwireless

Ser

vVo

ice IMT

-2000HSDPA

2009/10/22 FA/Tohoku University2009/10/22 FA/Tohoku University 11We are here

1980 1990 2000 Year

V

2010

Technical Issues of 4GTechnical Issues of 4G


Limited bandwidth ProblemLimited bandwidth Problemp ITU allocated the spectrum for 4G systems in Dec.

2007.p 450~470MHz (20MHz)p 790~806MHz (16MHz)p 790 806MHz (16MHz)p 2.3~2.4GMHz (100MHz)p 3.4G~3.6GHz Global use (200MHz)

200MH band idth in the global f eq enc bandp 200MHz bandwidth in the global frequency bandmust be shared by several operators (at least 2)and reused everywhereand reused everywhere.

p Probably, frequency reuse factor can be around25% � an effective bandwidth/BS is only around/ y12.5MHz/link.

p 1Gbps/12.5MHz is equivalent to 80bps/Hz/BS!!


p q p


p MIMO multiplexing can improve the spectrumff d h defficiency or can decrease the required transmit

power.p But too many antennas 70

80NxN MIMO (w/receive diversity)

p But, too many antennasare necessary. 60

70

7

8

MIMO

40

50

(bps

/Hz)

5

4

6

5

20

30C ( 4

3

2

0

10 N=1

2009/10/22 FA/Tohoku UniversityFA/Tohoku University 14FA/Tohoku University 14

00 10 20 30

Average total received Es/N0 per receive antenna (dB) G. J. Foschini and M. J. Gans, “On limits of wireless communicationsin a fading environment when using multiple antennas,” WirelessPersonal Commun., Vol.6, No. 3, pp.311-335, Mar. 1998.2009/10/22

Transmit Power ProblemTransmit Power Problemp Peak power is in proportion to “transmission rate”

“f [ f l ]” h f hx “fc2.6 [Hata-formula]” where fc is the carrierfrequency.Th i d t it f 8kb @2GH ip The required transmit power for 8kbps@2GHz is1Watt for a communication range of 1,000m.The required peak transmission power forp The required peak transmission power [email protected] needs to be increased by1Gbps/8kbps x (3.5GHz/2GHz)2.6 =535,561Gbps/8 bps (3 5G / G ) 6 535,56times, that is, 536kWatt.

p Obviously, this cannot be allowed！y,


# M. Hata, “Empirical formula for propagation loss in land mobile radio services”, IEEE Trans. Veh. Technol., VT-29, pp. 317-325, 1980.

2009/10/22

Severe Channel ProblemSevere Channel Problemp In terrestrial wireless communications, the transmitted

signal is reflected or diffracted by large buildings between signal is reflected or diffracted by large buildings between transmitter and receiver, creating propagation paths having different time delays.

p For signal transmissions using 25MHz bandwidth 1 symbol p For signal transmissions using 25MHz bandwidth, 1 symbol length in time is equivalent to the distance of 12 m. So, many distinct multipaths exist, thereby extremely enhancing the channel frequency-selectivityenhancing the channel frequency selectivity.

Large obstaclesd 4d-4

Localscatterers

Transmitter

R i


ReceiverReflection/diffraction

10p Wireless channel is l

1

anne

l gai

nextremely frequency-selective.S d d

0.1Chap Some advanced

equalization technique is

10

0.010 10 20 30 40 50 60 70 80 90 100

Frequency (MHz)

technique is necessary.

1

10nn

el g

ain

0.1Cha

n

L=16 Uniform power delay profilel-th path time delay=100l + [-50,50)ns


0.010 1 2 3 4 5 6 7 8 9 10

Frequency (MHz)

Evolution of Wireless AccessEvolution of Wireless Accessp FDMA was used in 1G, TDMA in 2G, DS-CDMA in 3G

OFDMA d SC FDMA d f d li k d li kp OFDMA and SC-FDMA are used for downlink and uplink ,respectively, in 3.9G(LTE)

p What is Giga-bit wireless technology for achievingp g gy g1Gbps/25MHz/BS in 4G?

Time-domain signal processing

Frequency-domainSignal processing

?4G

OFDMA，SC-FDMA

Time-domain signal processing

f3f2f1q

uen

cy 6.25kHzFDMA

1G3.9G(LTE)Time

Freq

CDMA3G eq

uen

cy

TDMA2G

Single-carrier

en

cy

3

3G

Fre

# 2# 3

Spreading code#1M

Hz25kHz

2009/10/22 FA/Tohoku University2009/10/22 FA/Tohoku University 18Time

Freq

ue

Time

1 3 1 2 32

Ex. 20ms

5M

Frequency-domain Signal ProcessingProcessing


DS with Rake vs MC with FDEDS with Rake vs. MC with FDEp As the number of resolvable

paths increases the channel SC-CDMA for downlink w/full code-multiplexingpaths increases, the channel

frequency-selectivity gets stronger.

p The achievable BER

1.E+00DS-CDMA withrake combining

16

multiplexing

SC-CDMA with rake combining

p The achievable BER performance of 3G SC(DS)-CDMA with rake combining degrades significantly due to

1.E-01

R

16

L=2degrades significantly due to strong IPI.

n For a heavily loaded channel, even with L=2, a high BER

1.E-02

Ave

rage

BER

L 2

Uniform delay profileeven with L 2, a high BER floor appears if the code-multiplexing order is high.

p On the other hand, MC-CDMA 1.E-03

AMC-CDMA with

N c=256SF=16, C =16 L=2

16

QPSK,

with MMSE-FDE provides much better performance.

n Performance improves as L 1.E-04

MC-CDMA withMMSE equalization

16


increases.


0 5 10 15 20 25 30

Average received E b /N 0 (dB)

FDE Improves SC-CDMA P fPerformancep Coherent Rake combining can be replaced by one-tap FDE

Bl k t i i f N hin Block transmission of Nc chipsn Insertion of guard interval (GI) at the transmittern FFT/IFFT at the receiver

(a) Transmitter

n on

Time-domain spreading

c(t ) Copy

(b) Receiver Inse

rtio

nof

GITransmit

dataD

ata

mod

ulat

ioDataGI

Nc chipsNg chips

R i don*(t )

Time-domaindespreading

W(0)

Receiveddata

･･･

Dat

am

odul

atio

Removalof GI IF

FTT

c*(t )W(k)

Integrate& dump


de-FF

NoiseW(Nc-1)･･･

FDE

p Downlink BER performance can be 1 E+00

Multicode DS-CDMA for downlink w/full code-multiplexing, uncoded

performance can be significantly improved compared to the coherent

k i1.E-01

1.E+00

SC-rake

Significant rake receiver.p However, there is still a

big performance gap from 1.E-02

BER Uncoded

OFDM

Significant improvement

g p g pthe theoretical lower bound.n This is due to residual ISI

1.E-03Ave

rage

B QPSK data modulationRayleigh fadingL=16, uniformSF=256, C =256

SC-,MC-FDEn This is due to residual ISI after MMSE-FDE.

n Introduction of ISI cancellation technique can

1.E-04DS-FDE (MMSE)DS-rakeMC-FDE (MMSE)

SC-FDE (MMSE)SC-rake

SC ,MC FDE

cancellation technique can reduce the performance gap.

1.E-050 5 10 15 20 25

Average received E b /N 0 (dB)

OFDMLower bound


# F. Adachi, D. Garg, S. Takaoka, and K. Takeda,“Broadband CDMA techniques,” IEEE Wireless Commun.Mag., Vol. 12, No. 2, pp. 8-18, April 2005

Frequency-domain Block Signal DetectionSignal Detectionp For the broadband single-carrier signal transmissions, the

BER performance significantly degrades due to the strongBER performance significantly degrades due to the stronginter-symbol interference (ISI).

p Frequency-domain block signal detection is very promisingthat combines frequency-domain equalization (FDE) andblock signal detection.

d(0) d(2) d(Nc1)···

Nc-symbol transmit block NdHFY s

TE2

Frequency-t tion

on

timed(0) d(2) d(Nc 1)

Frequency-

sT

y YFrequency-

domain block signal

detection

GI

Nc-

poin

tFF

T

Dat

ade

-mod

ulat Received

dataData

Dat

am

odul

ati o

d

+GI

Frequencyselective

channel H


d

T. Yamamoto, et. al., to be presented at The VTC-fall, Anchorage, Alaska, USA, 21~23 Sept. 2009.

SIC & MMSEDSIC & MMSEDMMSE

Decision resultWeight Y

Interference cancellation

detection

Replica generation

computation

cancellation generation

It er ati

on

st ag e

0th layer

I n it ia l it e r a ti o n s t a g e

0th layer no cancellation cancellation

......

1th layer

d(0) d(1) d(2) ・・・ d(Nc-1) d(0) d(1) d(2) ・・・ d(Nc-1)

......

1th layer1th layer

cancellation

d(0) d(1) d(2) ・・・ d(Nc-1) d(0) d(1) d(2) d(Nc 1)

cancellationcancellation

no cancellation

...Nc-1th layer

...Nc-1th layer

d(0) d(1) d(2) d(Nc 1) d(0) d(1) d(2) ・・・ d(Nc-1)

cancellation cancellation


......

d(0) d(1) d(2) ・・・ d(Nc-1) d(0) d(1) d(2) ・・・ d(Nc-1)


QRM MLDQRM-MLDp QR decomposition is applied to Y to detect the symbols in

decreasing order of reliability using MLDdecreasing order of reliability using MLD.

Y QR decomposition

Multiplication of QH

MLD based on M algorithm

Decision result

decomposition of QH M-algorithm

QRHF

p MLD based on the M-algorithmp gn Reliability based on the squared Euclidian distance is

calculated.M b l did t h i th hi h t li bilit


n M symbol candidates having the highest reliability areselected as surviving symbol candidates.


p As the number ofiterations increases theiterations increases, theBER performanceimproves and approachesth t f th l b d

1.0E-01

QRM-MLD

Iterative SIC&MMSEDi=1i=0

i=2

that of the lower bound.p The Eb/N0 gap from the

lower bound for the

1.0E-02

BER MMSE-FDE

QRM-MLDM=16 M=64

M=256

average BER=10-4reduces by 3.2 dB whenusing two iterations (i=2)

1.0E-03

Aver

age

B

Theoretical lower bound

using two iterations (i=2).p Much better performance

can be achieved by usingQRM MLD (th b M

1.0E-04 16QAMNc=64Ng=16QRM-MLD (the number M

of surviving symbol replicacandidates equals to 64 Average received Eb/N0 (dB)

1.0E-055 10 15 20 25

Ng=16L=16-path uniform


and 256) than iterativeSIC&MMSED.


T. Yamamoto, K. Takeda, and F. Adachi, “A Study of Frequency-Domain Signal Detection forSingle-Carrier Transmission,” Proc. IEEE VTC 2009 Fall, 20-23 Sept, 2009, Araska, US.

Overlap FDEOverlap FDEp The insertion of GI (CP) for FDE reduces the throughput by

a factor of 1/(1+Ng/Nc)a factor of 1/(1+Ng/Nc).

CopyDataGI

Nc symbolsNg symbols

p Without GI insertion, the inter-block interference (IBI) isproduced.produced.

n FDE output without GI insertion

12 1 E cN

NoiseIBI residualISI residual)()()(12)(ˆ0

tskHkW

NTEts

c

kcc

s


Average equivalent channel gain

p Impulse response of circular FDE filter

0.5circular FDE filter concentrates at the vicinity of t=0.

0.3

0.4

afte

r FD

E MMSE-FDEN c =256L =16E /N 10dB

p This property can be exploited to mitigate the IBI after FDE. 0 1

0.2

0.3

lse re

spon

se E b /N 0=10dB

0

0.1

-128 -64 0 64 128

Inpu

l

p Overlap FDE tA partial sequence of M symbols is picked up

FFT block

time


FFT blockFFT block

Application of Overlap FDE to DS CDMADS-CDMAp Overlap FDE

improves the 2 L=16-pathimproves the

throughput of SC HARQ.

1 5

puniform power delay profileQPSKTurbo coded HARQT II S P2

p Since GI (CP) insertion is not required, overlap FDE

1.5

bps/

Hz)

Type II S-P2Nc=256, SF=C=16

Overlap FDE (M=64)q , pcan be applied to the present HSDPA using SC-CDMA technology

1

ughp

ut (

b Overlap FDE (M 64)

Conventional FDEusing Ng=32-chip GISC CDMA technology.

nMuch higher throughput can be achieved

0.5

Thro

u g g p

Coherent rake combiningachieved.nThe 3G air interface does not need to be changed at all 0

Coherent rake combining


changed at all.


0-5 0 5 10 15 20 25 30

Average received Es/N0 (dB)

p Only 200MHz bandwidth is available for the globalf h b d d h buse of 4G spectrum. This bandwidth must be

shared by the up/down links. Even though thesingle frequency reuse is used an effectivesingle frequency reuse is used, an effectivebandwidth may be around 25MHz/BS.

p Therefore achievable peak data rate is whenp Therefore, achievable peak data rate is, whenusing 25MHz, 1.8bps/Hz×25MHz=45Mbps

p Single use of HARQ technique cannot achievep S g e use o Q tec que ca ot ac e e1Gbps!

p Multiple-input/multiple-output (MIMO) antennap p p p ( )technique will play an important role to realize 4Gsystems.


MIMO Multiplexing May Be A SaviorSaviorp Independent data streams are transmitted simultaneously

from transmit antennas using the same carrier frequencyfrom transmit antennas using the same carrier frequency.p SDM is to increase achievable data rate within the limited

bandwidth, i.e., the channel capacity in bps/Hz.

Nt antennas Nr antennas

Multipathchannel

Coding S/P Signal]detection Decod.

channel


SDMSDMp Next generation (4G) wireless systems are

t d t id b db d k t d texpected to provide broadband packet dataservices of up to 1Gbps. However, availablebandwidth is limitedbandwidth is limited.

n In December 2007, ITU allocated 3.4~3.6GHz band for4G services. Only 200MHz is available for global use.

n This must be shared by the up/down links. Althoughone-cell reuse of 100MHz is possible, but, effectivebandwidth which can be used at each BS is only aroundba d dt c ca be used at eac S s o y a ou d25MHz/link. 1Gbps/25MHz is equivalent to40bps/Hz/BS!!

Th th d l t f hi hl tp Thus, the development of a highly spectrumefficient wireless transmission technology of>40bps/Hz/BS is demanded


>40bps/Hz/BS is demanded


Single-carrier MIMO SDMSingle-carrier MIMO SDMp Joint MMSE frequency-domain equalization (FDE) and

parallel interference cancellation (PIC) is repeated for deparallel interference cancellation (PIC) is repeated for de-multiplexing while achieving frequency-diversity gain.

Nr antennasNt antennas

Frequency-domainiterative PIC

-GI･･･

-GI

Nr antennasReceived

data

･･･

Nt

Nt antennas

+GI･･･

+GIS/P

Data

Mod. P/S-GI+GI

Nt signals

SISO channel

MIMO channel

LLR cal･･2D MMSE

Nr signals

Nt signalsN

FFTi=0

IFFT･･

… …… …......

...

...

･･･

LLR cal.

･･･

･2D MMSE-FDE

Nt signalsrNt

TFFTPIC… …...

SIMO channel

Nt signalsFFT

i>0

Replica generation

IFFT･･･

････


･･･

Frequency-domain Time-domain

…… …...... ...

generation

1.E+00Multicode DS CDMA (4 4)MIMO multiplexingSC

1.E-01

Multicode DS-CDMA (4,4)MIMO multiplexingwith iterative FDIC, L=16, =0dB, N c=256,

N g =32, SF (=C )=1, QPSK

SC

Ng=32, QPSK

1.E-02

ER

1.E-03

Ave

rage

BE

FD ZF

FD MMSE

1.E-04FD V-BLAST

IterativeFDIC

(1 4)SIMO

1.E-05

(1,4)SIMO

Matched filter bound


-5 0 5 10 15 20

Average received Eb/N0 per antenna (dB)

4x4 MIMO SDM Throughput4x4 MIMO SDM Throughputp MIMO signal detection

using FDE and FD IC

8

using FDE and FD IC provides higher throughput than OFDM. 6

7 SC w/ FDIC OFDM

w/ FDIC

Start

MMSE FDE Weight comp.

5

bit

/s/H

z)

SC w/o

FDIC

FDICca

gen

e.

R c

omp.

MMSE FDE Weight comp.

3

4

rou

ghp

ut (

b

OFDM w/o

Rep

lic

LLR

Output 2

3

Th

rSC (4, 4)MIMO SDM with iterative FDIC

OFDM w/o FDIC

# A. Nakajima, D. Garg and F. Adachi, “Frequency-domain iterative parallel interference cancellationfor multicode DS-CDMA-MIMO multiplexing,” Proc.IEEE VTC’05 Fall, Vol.1, pp. 73-77, Dallas, U.S.A.,26-28 Sept 2005 0

1

( , )(i=4)HARQ type-II S-P2QPSK, Nc=256, Ng=32L=16-path uniform power delay profile


26 28 Sept. 2005.# A. Nakajima and F. Adachi, “Iterative FDIC using

2D-MMSE FDE for turbo-coded HARQ in SC-MIMOmultiplexing,” IEICE Trans. Commun. Vol. E90-B,No.3, pp.693-695, Mar. 2007.

0-5 0 5 10 15 20 25

Average received Es/N0 per receive antenna (dB)

p p y p

Spread 4x4 MIMO/SDM ThroughputThroughputp Spread MIMO SDM w/FDIC

provides the throughput

8MC-CDMA (4,4)SDM with iterative FDIC, QPSK,SF (=U )=256 N c=256 N g =32 L=16 S-P2 i=4provides the throughput

similar to non-spread SC.n Much better throughput than

OFDM in a high Es/N0 egion6

7SF ( U ) 256, N c 256, N g 32, L 16, S P2, i 4

OFDMMC, DS

OFDM in a high Es/N0 region.n Almost the same throughput as

OFDM in a low Es/N0 region.Start

5

bps/

Hz) MC, DS

w/ FDIC

FDIC

Start

2D-MMSE FDE

de ing

de ngWeight comp.

3

4

Thro

ughp

ut (b

OFDM

w/o FDIC

Single transmit antenna limit

ca g

ene.

R c

omp.

Mul

tico

dde

spre

adi

Mul

tico

dsp

read

in

2

Rep

lic

LLR

EndA. Nakajima, D. Garg and F. Adachi, “Frequency-domainiterative parallel interference cancellation for multicode DS-CDMA MIMO lti l i ” P IEEE VTC’05 F ll V l 1 73 0

1 MC DSw/ FDICw/o FDIC


CDMA-MIMO multiplexing,” Proc. IEEE VTC’05 Fall, Vol.1, pp. 73-77, Dallas, U.S.A., 26-28 Sept. 2005.A. Nakajima and F. Adachi, “Iterative FDIC using 2D-MMSE FDEfor turbo-coded HARQ in SC-MIMO multiplexing,” IEICE Trans.Commun. Vol. E90-B, No.3, pp.693-695, Mar. 2007.

0-5 0 5 10 15 20 25 30

Average received E s /N 0 per antenna (dB)

p Increasing the no. f

80N N MIMO ( /r i di r it )of antennas can

improve the spectrum efficiency 60

708

NxN MIMO (w/receive diversity)

spectrum efficiency or can decrease the required transmit 40

50

s/H

z)

5

6

5

7

qpower.

30

40

C (b

ps

4

3

10

20 2

N=1

00 10 20 30

Average total received Es/N0 per receive antenna (dB)


G. J. Foschini and M. J. Gans, “On limits of wireless communicationsin a fading environment when using multiple antennas,” WirelessPersonal Commun., Vol.6, No. 3, pp.311-335, Mar. 1998.

However, Space for MIMO Antennas Is Limited !Antennas Is Limited !p How to implement many antennas in a hand

blportable unit?p Wearable antenna on your head?

H d tHead antenna

No space in a No space in a small hand portable unitp


MIMO Cannot Solve Power ProblemProblemp Peak power is in proportion to “transmission rate” x “fc2.6

[Hata formula]” where fc is the carrier frequency[Hata-formula] where fc is the carrier frequency.p Links for broadband data services will be severely power-

limited.p Let’s consider the peak transmit power for [email protected]

at a communication range of 1,000m. We assume therequired transmit power for 8kbps@2GHz is 1Wattrequired transmit power for 8kbps@2GHz is 1Watt.

p The required peak transmission power increases by1Gbps/8kbps x (3.5GHz/2GHz)2.6 = 535,561 times, that is536kWatt. Obviously, this cannot be allowed.


M. Hata, “Empirical formula for propagation loss in land mobile radio services”, IEEE Trans. Veh. Technol., VT-29, pp. 317-325, 1980.

2009/10/22

Distributed Antenna Network

(or Distributed MIMO)(or Distributed MIMO)


Transmit Power ProblemTransmit Power Problemp Peak power is in proportion to “transmission rate” x “fc2.6

[Hata formula]” where fc is the carrier frequency[Hata-formula] where fc is the carrier frequency.n Assume that the required transmit power for 8kbps@2GHz is 1Watt for

a communication range of 1,000m. The required peak transmissionpower for 1Gbps@3 5GHz needs to be increased by 1Gbps/8kbps xpower for [email protected] needs to be increased by 1Gbps/8kbps x(3.5GHz/2GHz)2.6 = 535,561 times, that is, 536kWatt. Obviously, thiscannot be allowed.

p To keep the 1W power the communication range should bep To keep the 1W power, the communication range should bereduced by 43 times(i.e., 1,000m �23m). �Femto cellularnetwork.

8kb @2GH

# M Hata “Empirical formula

8kbps@[email protected]


# M. Hata, Empirical formula for propagation loss in land mobile radio services”, IEEE Trans. Veh. Technol., VT-29, pp. 317-325, 1980. 2009/10/22

1000m23m

Distributed Antenna NetworkDistributed Antenna Networkp A group of distributed antennas serves a user in a

cooperative manner to mitigate the powercooperative manner to mitigate the power.p The distributed MIMO diversity transmission can mitigate

the problems resulting from path loss, shadowing loss,and multi-path fading.

BS

Distributed MIMO

SDM/diversityDistributed MIMO relay

2009/10/22 FA/Tohoku University2009/10/22 FA/Tohoku University 42Optical fiber cable

Distributed network antenna

Distributed wireless port

1.0E+00=3.5

=7.0

Es/N0=10dB

1.0E-01

ssa]

Es/N0=10dB

L=16 uniform

tp The distributed MIMO diversity

using the maximal ratio transmit

1 0E 02rob[

C<a

bsci

s

tt

using the maximal ratio transmit (MRT) diversity increases significantly the channel

it 1.0E-02Pr t

t

capacity.p The use of only around 5

antennas near a user provides a 1.0E-03

0 1 2 3 4 5 6 7 8 9 10Ch l it C (bit/ /H )W0(k)

psufficient improvement.

p Transmit power can reduce significantly Channel capacity C (bit/s/Hz)W0(k)

ing

. . Opt

. de

-m

od. Distributed

antennas

y da

ta FFT

IFFT

+G

I

n [%

]

10(0,0) (1,0)

Receiver

y

significantly.

Cha

nnel

cod

Dat

a m

od

Opt

. m

od.

de-

d.nsm

it b

inar

y

T

WNt-1(k)

T I e po

wer

dis

tribu

tion

1

(1,1)(0,1)

Antenna

Receiver x


Opt

. d

modTra

Opt. fibercable

FFT

IFFT

+G

I

Aver

age

0.1

OFDMA 2-hop VCNOFDMA 2-hop VCNp An introduction of 2-hop relay may be a good

l h bsolution at the beginning.

relay Single-hop

Terminal

Base station

Terminal


p OFDMA is suitable for multi hop relaying

Coremulti-hop relaying.

p The same subcarriers can be reused at different links.

MTMT Network

Different groups of subcarriers are allocated to neighbor links, but the

WP MTCPg ,

same subcarriers can be reused by links which are far away each other

2-hop VCNMT

far away each other.p Adaptive subcarrier

allocation based on the SINR i i ti i f

SIN

R

SIN

R

SINR maximization is promising*.

# H Ishida E Kudoh and F Adachi “A study on

f

f

S

WP MTCP


# H. Ishida, E. Kudoh and F. Adachi, A study onsubcarrier allocation for a 2-Hop OFDMA virtualcellular network”, Proc. 4th InternationalWorkshop of Tohoku University and YeungnamUniversity, pp.7-8, 11-13 Nov. 2007.

Adaptive channel allocation

WP MT

Some Concluding RemarksSome Concluding Remarksp Next generation wireless networks will require

b l h l f b dG i g a - b i t w ireless technology of ~1Gbps and40bps/Hz/BS under severe co-channelinterferenceinterference.

p Frequency-domain signal processing plays animportant role in the signal transmissions in aimportant role in the signal transmissions in asevere frequency-selective channel.

p To meet the spectrum efficiency and transmitp o eet t e spect u e c e cy a d t a s tpower requirement, distributed antenna networkis attractive.

p Lots of interesting and important research topicsremain before the born of next generation

i l t


wireless systems.


Please Visit Our Homepage for More Info.Homepage for More Info.

htt http: //www.mobile.ecei.tohoku.ac.jp


（安達 091022IWSDA'09）Wireless Challenge.ppt [ … Fumiyuki Adachi Dept ... 1G 2G Era e t imedi...

Documents

Transcript of （安達 091022IWSDA'09）Wireless Challenge.ppt [ … Fumiyuki Adachi Dept ... 1G 2G Era e t imedi...