adaptive and software defined 5g air interface for 5g wireless...

HUAWEI TECHNOLOGIES CO., LTD.

www.huawei.com

Adaptive and Software

Defined 5G Air Interface

for 5G Wireless Network

Dr. Jianglei Ma

Jun. 24, 2014

HUAWEI TECHNOLOGIES CO., LTD. Huawei Confidential

Why New Air Interface for 5G?

Local: 10Gbps~50Gbps Micro:

3Gbps~5Gbps Macro: 1Gbps~4Gbps

Massive Capacity (x1000)

• Throughput (kbps to Gbps)

• Duty cycle (0 to 109 )

• Connectivity (102 to 107)

• Signalling % (10-3 to 102)

Hyper Diverse Services

Massive Connectivity

Air interface needs to be changed dramatically

To enable any application and any service to connect everything and anything

Hyper Hetrogenity/All Spectrum


5G Air Interface Framework

New Spectrum Massive MIMO Advanced PHY

&MAC Full Duplex

Communications

Virtualized Radio

Access

Software Defined

Air Interface

New Enabling

Technologies

Higher spectrum

efficiency

Ubiquitous high

speed wireless

access

Massive always

connected

devices

Higher area

capacity

Lower energy

consumption Low cost devices

Ultra-low latency

and extreme

reliability

Meet 5G AI Requirements

Single air interface supports ultra diverse services(HyperService) and

cover all spectrums.


35pt

Font to be used by customers and

partners :

18pt


partners :


Primary Band Complementary Band

5

100M 1G 10G 100G

10KHz

100KHz

1MHz

10MHz

100MHz

Frequency(Hz)

CH BW

1GHz

5G Complementary 5G Primary

1G

3G 4G

2G

6G

5

5G is for all spectrum:

5G primary band -> Core services

5G complementary band -> 5G technology extension


35pt


partners :

18pt


partners :

Single Air Interface for All Applications

Unified air interface to support different waveform/multiple access scheme/TTI

Co-existence of waveforms with different time-frequency granularities

Ultra narrow bandwidth SC pulse for MTC

f

f

t

f

t

Spectrum filter

f

t

OFDM

symbol

duration

Guard

time

Sub-carrier spacing

Spectrum filtered OFDM

Spectrum filtered OFDM

http://www.google.com/url?sa=i&rct=j&q=&esrc=s&frm=1&source=images&cd=&cad=rja&uact=8&docid=Uq5uRBcJj2ClgM&tbnid=QZdzU8bC1YAghM:&ved=0CAUQjRw&url=http://www.wirelessdesignmag.com/blogs/2012/09/how-non-line-sight-backhaul-really-works&ei=M8haU9uaKaTB2QX7roCwCw&bvm=bv.65397613,d.b2I&psig=AFQjCNFT81bYDI5JrNdXg0xZkqkWFcV78A&ust=1398544804839668

http://www.google.com/url?sa=i&rct=j&q=&esrc=s&frm=1&source=images&cd=&cad=rja&uact=8&docid=Uq5uRBcJj2ClgM&tbnid=QZdzU8bC1YAghM:&ved=0CAUQjRw&url=http://www.wirelessdesignmag.com/blogs/2012/09/how-non-line-sight-backhaul-really-works&ei=M8haU9uaKaTB2QX7roCwCw&bvm=bv.65397613,d.b2I&psig=AFQjCNFT81bYDI5JrNdXg0xZkqkWFcV78A&ust=1398544804839668


35pt


partners :

18pt


partners :

Software Defined Air Interface

Protocols

Waveforms

Frame

Structure

Multiple

Access

Scheme

Coding

Modulation

Air Interface

Configuration

One size fits all -> AI Adaptation

Traffic type

Transmitting & receiving

condition

Operating spectrum

band

Optimized

Air Interface

Candidate technologies

for all AI building blocks


35pt


partners :

18pt


partners :

New waveform Candidates

-6 -4 -2 0 2 4 6

-80

-60

-40

-20

0

20

40

Baseband Frequency (MHz)

TX

Sig

nal P

SD

(dB

)

Prototype Filter: Root Raised Cosine

OFDM

FBMC

-6 -4 -2 0 2 4 6

-80

-60

-40

-20

0

20

40

Filter: soft truncation using RC window with = 0.5

f (MHz)

OFDM, 50 RBs

Filtered OFDM, 50 RBs

SCMA

FBMC

F-

OFDM

1 1 1

Transition

Period Free

2

TN

2)12(

TLN

tails

time

Ch. 1

Ch. 2

Ch. 3

Freq.

Filtered

OFDM

UE #1

Filter #1Freq. Short FFT

processing

for UE #1

Short FFT

processing

for UE #2

Short FFT

processing

for UE #3

Filtered

OFDM

UE #2

Filtered

OFDM

UE #3

Filter #2

Filter #3


Huawei Confidential 9

What is SCMA?

- Sparse Codebook Multiple Access

•A new frequency

domain non-

orthogonal waveform •Sparsity of the

codeword to enable

overloading with

reasonable Rx

complexity

QAM Mapper FEC Encoder (qs1,qs2,qs3,qs4)

Spreading Signature (s1,s2,s3,s4)

(b1,b2) q

CDMA/LDS

Repetition

Multi-

dimensional

codeword


What is SCMA? - SCMA Codebook

• SCMA codebook based on Multi-dimensional Lattice

Constellation to exploit shaping gain and coding gain

• Each UE stores a unique codebook

UE1 UE2 UE3 UE4 UE5 UE6

(1,0) (0,0) (1,1) (0,1) (b1,b2) (1,0) (1,1)

16-point

orthogonal lattice

in 4 real

dimensions

Unitary lattice

rotation

Rotated 16-point

lattice in 4 real

dimensions

Mother constellation

Projections on first

and second

complex

dimensions

QPSK 1 QPSK 2


What is SCMA? - Scalable SCMA with Adaptive System Parameters

System requirements and

network/UE capabilities

SCMA Parameters

SCMA

Configuration SCMA Mode

Link-budget

Coverage

Connectivity

Throughput

Multiplexing gain

Processing capabilities

……

SCMA

OFDMA (fall back mode)

MC-CDMA (fall back mode)

……

Number of codewords of an SCMA codebook: M

Spreading factor: K

Max number of layers (or codebooks/signatures) : J

Number of nonzero elements of each codeword: N

Compromise among spectral efficiency, coverage,

detection complexity, connectivity, and link budget to

adapt to different application scenarios


35pt


partners :

18pt


partners :

Why SCMA?

Enable Massive

Connectivity

Overloaded signal

superposition

Low signaling

overhead and low

latency with grant-

free transmission

Low multi-user

detection

complexity

Better Spectrum

Efficiency

Shaping gain and coding

gain from Multi-

dimensional lattice

constellation based

codebook design

More reliable link

adaptation with

Interference averaging

Better Scalability

and Flexibility

Extra domain for

multi-user resource

sharing

Easy adaptation to

different application

requirements

HUAWEI TECHNOLOGIES CO., LTD. Huawei Confidential -4 -3 -2 -1 0 1 2 3 410

-3

10-2

10-1

100

SNR(dB)

BLE

R

SCMA vs. OFDMA, AWGN, SE = 0.5, 1, 1.2 bits/tone

SCMA 8-point1/3, 2 layers



OFDMA QPSK1/4

OFDMA QPSK1/2

OFDMA QPSK0.6

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 510

-3

10-2

10-1

100

SNR(dB)

BLE

R

SCMA vs. LDS, AWGN, SE = 1 bits/tone

SCMA 16-point1/2, 2 Layers

LDS 16-point1/2, 2 Layers

LDS 4-point1/3, 6 Layers

-3 -2 -1 0 1 2 3 4 5 610

-3

10-2

10-1

100

Single user SNR (dB)

BLER

AWGN - LDS QSPK0.75 - SCMA 4-point0.75

LDS 1 or 2 layers

LDS 4 layers

LDS 6 layers

SCMA 1 or 2 layers

SCMA 4 layers

SCMA 6 layers

Gain of SCMA

SCMA outperforms LDS:

› Shaping gain of multi-

dimensional 16 point SCMA

codebook

› Better co-pared layer

interference cancellation

capability due to dimensional

power variation

Inter-layer interference cancellation gain

Shaping gain

Compare SCMA and LDS

Compare SCMA and OFDM


35pt


partners :

18pt


partners :

Gain of SCMA:

- OL Multi-user Multiplexing

OL Multi-user multiplexing without need of short

term CSI information

More flexible and robust link-adaptation mechanism

0

10

20

30

40

SCMA MU-SCMA

Throughput Gain [%]

Coverage Gain [%]


35pt


partners :

18pt


partners :

15

Gain of SCMA:

- UL Contention Based Multiple Access

Grant-free UL multiple access based on SCMA blind detection:

~3X connected devices

No dynamic scheduling singling overhead

Low latency

0.0

200.0

400.0

600.0

800.0

1000.0

1200.0

latency<=4ms Latency <= 20ms

121.9

611.7

328.1

1082.8

Number of Supported Device (1 MHz, Average Packet Drop Rate 1%)

OFDMA SCMA


Radio Access Virtualization - key Technologies

“Interference

-free “

Wireless

Access Hyper

cell

Centralized signal

processing

New PHY

channels

UE centric PHY

channel

designs

Novel UE and TP

association

mechanism

Communications

between

distributed

transmitters and

distributed receivers

UE centric TP

Optimization & UE

centric device mash

HT

UE

dedicated

connection

ID

Network

oriented

measure

ment

New UE

connection

mechanism

UL centric

measurement

scheme

Virtual

Tx &

Virtual

Rx

Cloud

group

processor


Hyper Cell and UE Dedicated Connection ID

•Hyper cell: a virtual entity covering a group of physical TPs

•A Logic entity ID is assigned to each hyper cell

•UE dedicated connection ID

•Decouple UE access with TP

Hyper Cell-2

Controller

Hyper Cell-1

Belong to Hyper Cell 1&2

Controller

Central

Controller

UE ID Logic Entity ID UE dedicated

connection ID


35pt


partners :

18pt


partners :

19

Network Oriented Measurement

To monitors and periodically measure the UL channels

TP3

TP1

TP2

S3, T3

S2, T2

S1, T1

TP4

S4, T4

UL SRS

decouples

from any TP

UE 0

UE 1

•Network maintains a TP and UE association map.

•UL centric measurement: •For UL/DL TP clustering/optimization, mobility management, link

adaptation….

•UE measurement assisted

UE and TP relation

map

UE centric TP

Optimization

UE centric Control CH

and Data CH Controller


Low Cost Deployment

•Dynamically configure

virtual transmit nodes

•UE centric transmit

node optimization and

reception node

optimization

dramatically reduce the

complexity.

Network

Central

Controller-1

Virtual Transmit Node

Virtual Transmit Node

Central

Controller-2

Regional

Controller


•UE cooperation allows long term CQI based link

adaptation

•No need for fast link adaptation

•Cell edge data rate enhancement

Robust Link Adaptation and Uniform Data

Distribution

Throughput

distribution

SU-MIMO

MU-MIMO JT

JT&JR

Macro-cell (500 m)


Autonomous Ad-hoc UE Cooperation

Strategically

deployed

terminal

d1

d1 d2

d1 d2

d2

d3

d3

d4

d4

d5

d5 ACK

d4

d5

UE0

UE1

UE2

Ad-hoc UE group based cooperation

Autonomously generated

Not rely on dedicated helping UE(s)

Enhance network reliability


35pt


partners :

18pt


partners :

Boost Performance By Adding Strategically Placed

Help UEs

-1.5 -1 -0.5 0 0.5 1 1.5-1.5

-1

-0.5

0

0.5

1

1.5

0

1 2

3

4 5

6

7 8

9

10 11

12

13 14

15

16 17

18

19 20

21

22 23

24

25 26

27

28 29

30

31 32

33

34 35

36

37 38

39

40 41

42

43 44

45

46 47

48

49 50

51

52 53

54

55 56

57

58 59

60

61 62

Topology of AllUE UEs (all 1 seed indices)

Throughput[Mbps]

0

1

2

3

4

5

6

7

8

-1.5 -1 -0.5 0 0.5 1 1.5-1.5

-1

-0.5

0

0.5

1

1.5

0

1 2

3

4 5

6

7 8

9

10 11

12

13 14

15

16 17

18

19 20

21

22 23

24

25 26

27

28 29

30

31 32

33

34 35

36

37 38

39

40 41

42

43 44

45

46 47

48

49 50

51

52 53

54

55 56

57

58 59

60

61 62

Topology of AllUE UEs (all 1 seed indices)

Throughput[Mbps]

0

1

2

3

4

5

6

7

8

-1000 -500 0 500 1000 1500

-1500

-1000

-500

0

500

1000

1500

0%

100%

200%

300%

400%

500%

600%

700%

50 100 200 300 400

100 picos, Thr

100 picos, cov

200 picos, Thr

200 picos, Cov


35pt


partners :

18pt


partners :

• Re-define physical channels to support radio access

virtualization

Cell centric -> UE centric

• Virtualization of data channel and signalling channel offers

more scheduling flexibilities.

Boost data and control capacity

Better energy saving

Better mobility management

UE Centric Virtualized Physical Channels

Virtualized PHY Channel Designs

Virtual Data

Pipes

Broadcast

common

control/virtual

dedicated

control

UE Centric

channel

sounding &

Measurement

SFN Synch


Intelligent Spectrum Utilization

Self-

Optimized

WN

Integrated Air Interface

Central Spectrum Management Controller

Unlicensed

Spectrum Licensed

Spectrum

High

Frequency

Spectrum

Access

Link

BH

Link D2D

Link

Key Technologies:

› Single AI covers all spectrum

bands

› Efficient utilization of the

spectrum by different nodes

and link

» Improve the spectral

efficiency

» Reduce the interference.

› Joint optimization


Air Interface Design Principle Change

Interference avoidance

One size-fit-all

Node computation based

UE centric

Interference free

Service-Oriented

Cell centric

Physical UE connects to

physical cell

Virtual UE connects to logic

access point

Vir

tualizati

on

UE oriented measurement Network oriented measurement

Cloud group computation

based

Thank you www.huawei.com

adaptive and software defined 5g air interface for 5g wireless...

Documents