Post on 13-Jan-2016
description
PAPR Reduction Method for OFDM PAPR Reduction Method for OFDM Systems without Side Information Systems without Side Information
Pornpawit Boonsrimuang, Pisit Boonsrimuang Kazuo Mori, Tawil Paungma and Hideo Kobayashi
ContentsContents
Research TopicObjective of ResearchConventional MethodsProposal of PAPR Reduction Method Performance EvaluationConclusions
Research TopicResearch Topic
Proposal of PAPR Reduction Method for OFDM Signal without Side Information
OFDM: Orthogonal Frequency Division Multiplexing
Next Generation of Wireless LANNext Generation of Wireless LANNext Generation of Wireless LANNext Generation of Wireless LAN
Indoor
GigabitGigabitOutdoor
100Mbit100Mbit
Target of Research FieldTarget of Research Field
Advantages of OFDMAdvantages of OFDM
Enable the higher data rate transmissionRobustness to multi-path fadingEfficient usage of frequency bandwidth
The OFDM has already been adopted as the standard transmission techniques in
ADSL (Asynchronous Digital Subscriber Line)Terrestrial Digital TV
WiMAX(Worldwide Interoperability for Microwave Access)
Wireless LAN(Wireless Local Area Network)
Candidate transmission technique for the next generation of wireless LAN systems
Structure of OFDM SystemStructure of OFDM SystemStructure of OFDM SystemStructure of OFDM System
System Overview
.. . Parallel to Serial
(P/S)
ModSerial to Parallel (S/P)
s(t)Add GI
Fading Channel
H()
Noise n(t)
.. . Serial to Parallel (S/P)
FFTParallel to Serial
(P/S)
r(t) Equalization
TransmitterTransmitter
ReceiverReceiver
Channel
(a) OFDM spectrum
vs.
(b) conventional FDM spectrum
IFFT
.. .Demod
.. . AMP
Channel estimation
Remove GI
Larger peak to averaged power ratio (PAPR)
Larger PAPR will cause the degradation of BER performance and expansion of spectrum
re-growth in the non-linear channel
Disadvantage of OFDMDisadvantage of OFDM
2
max 0 12
max | |, =0, , 1
[| | ]
nn N
av n
SPPAPR n N
P E S
PAPR: Peak to Averaged Power Ratio
0 100 200 300 400 5000
0.5
1
1.5
2
Sample NumberA
mp
litu
de
of O
FD
M T
ime
Dom
ain
Sig
nal
0 100 200 300 400 5000
0.5
1
1.5
2
Sample Number
Am
pli
tud
e of
OF
DM
Tim
e D
omai
n S
ign
al
Comparison of OFDM Signal with Single Carrier Signal
Comparison of OFDM Signal with Single Carrier Signal
(a) OFDM signals in time domain (b) Single carrier signal in time domain
Amplitude of OFDM time domain signal is fluctuated Amplitude of OFDM time domain signal is fluctuated much larger than that for single carrier signalmuch larger than that for single carrier signal
Operation of OFDM Signal Operation of OFDM Signal at Lower IBO in Non-Linear Channelat Lower IBO in Non-Linear Channel
Operation of OFDM Signal Operation of OFDM Signal at Lower IBO in Non-Linear Channelat Lower IBO in Non-Linear Channel
-10 -8 -6 -4 -2 0 2 4-10
-8
-6
-4
-2
0
2
4
Input Power (dB)
Out
put
Pow
er (
dB)
r=2 r=4
r=30 r=6
Linear
where F[ ] represent the AM/AM conversion characteristics of non-linear amplifier
,arg
, ,
l kj y
l k l ks F y e
-10 -8 -6 -4 -2 0 2 4-10
-8
-6
-4
-2
0
2
4
Input Power (dB)
Out
put
Pow
er (
dB)
r=2 r=4
r=30 r=6
Linear
where F[ ] represent the AM/AM conversion characteristics of non-linear amplifier
Inter-modulation Noise
Operation of OFDM Signal Operation of OFDM Signal at Higher IBO in Non-Linear Channelat Higher IBO in Non-Linear Channel
Operation of OFDM Signal Operation of OFDM Signal at Higher IBO in Non-Linear Channelat Higher IBO in Non-Linear Channel
,arg
, ,
l kj y
l k l ks F y e
-1.5 -1 -0.5 0 0.5 1 1.5-1.5
-1
-0.5
0
0.5
1
1.5
.. . Parallel to Serial
(P/S)Mod
Serial to Parallel (S/P)
s(t)Add GIIFFT
.. . AMP
-1.5 -1 -0.5 0 0.5 1 1.5-1.5
-1
-0.5
0
0.5
1
1.5
Transmitter structure of conventional OFDM
Inter-Modulation Noise Occurred Inter-Modulation Noise Occurred at Non-Linear Amplifier at Non-Linear Amplifier
Inter-Modulation Noise Occurred Inter-Modulation Noise Occurred at Non-Linear Amplifier at Non-Linear Amplifier
(a) Input of power amplifier (b) Output of power amplifier
AMP
Non-LinearNon-Linear
Required Low Input Back-Off
Inefficient Usage of Power Amplifier
Larger PAPR
N sub CarriersOFDM with N sub carriers has PAPR ratio of 10log N [dB]
Disadvantage of OFDM TechniqueDisadvantage of OFDM Technique
Several methods have been proposed to reduce the PAPR of OFDM signal
Data coding technique The main disadvantage of this method is to degrade transmission efficiencyClipping technique The main disadvantage of this method is to degrade BER performance Phase alignment technique The main disadvantage of this method is required to transmit side information which leads the degradation of transmission efficiency
Improvement of PAPRImprovement of PAPR
Conventional PAPR Reduction MethodsConventional PAPR Reduction MethodsBased on Phase Alignment Technique Based on Phase Alignment Technique The Selected Mapping (SLM) and Partial Transmission
Sequence (PTS) methods which control the phase of data sub-carrier
SLM and PTS are required to inform the phase information to the receiver as the side information (SI) for the correct demodulation
Although the conventional method could improve the PAPR performance relatively,
the transmission efficiency and system complexity would be degraded because of the necessity of
transmission of side information
Objective of ResearchObjective of Research
Improve PAPR and BER performances without
degradation of transmission efficiency
Proposal of Proposal of PAPR Reduction Method for OFDM Signal PAPR Reduction Method for OFDM Signal
Based on Phase Alignment Technique which Requires Based on Phase Alignment Technique which Requires No Side InformationNo Side Information
Proposal of PAPR Reduction MethodProposal of PAPR Reduction Methodfor OFDM Signal without Side Informationfor OFDM Signal without Side Information
Use the common weighting factor over one frame including the preamble symbols
The time-frequency domain swapping algorithm is employed in the determination of common weighting factor
The common weighting factor can be estimated together with the channel frequency response by using preamble symbols
The common weighting factor can be removed from the received data symbols by using the frequency domain equalization
Proposed method could improve the PAPR performance Proposed method could improve the PAPR performance without Side Informationwithout Side Information
Common weighting factor will be assigned for each sub-carrier over one frameCommon weighting factor will be assigned for each sub-carrier over one frame
0 1 2 N-1Number of sub-carriers
0je
1je 2je
( N 1 )je Weighting Factors for
Reducing PAPR
Structure of Proposed PAPR Reduction MethodStructure of Proposed PAPR Reduction Method
Common Common weighting factor weighting factor over one frameover one frame
1st Preamble Symbol
2nd Preamble Symbols
1st Data Symbol
Lth Data Symbol
On
e F
ram
eO
ne
Fra
me
0 10 20 30 40 50 600
0.5
1
1.5
2
2.5
3
3.5
Time domain
Am
plitu
de (
Tim
e
Dom
ain
)
Predetermined threshold level (S)
( )
,il ky
0 10 20 30 40 50 600
0.5
1
1.5
2
2.5
3
3.5
( ) inji
nE e
Am
plitu
de (
Tim
e
Dom
ain
)
Frequency domain
FFT
Basic concept of Time-Frequency Swapping Algorithm
( ) ( ), ,
( ),
( ),0
i il k l k
il k
il k
y if y Se
if y S
Determination of Weighting Factor by Using Time-Determination of Weighting Factor by Using Time-Frequency Domains Swapping AlgorithmFrequency Domains Swapping Algorithm
Flowchart of Time-Frequency Domains Swapping AlgorithmFlowchart of Time-Frequency Domains Swapping Algorithm
Calculation of Time-domain Signal
Calculation of Error Signal
FFT
0 20 40 60 80 100 1200
0.5
1
1.5
2
2.5
3
3.5
S1
2
3
0 N-1N-FFT
mi (
t)0 20 40 60 80 100 120
0
0.5
1
1.5
2
2.5
0 N-1N-FFT
1
2
ei (t)
Threshold level(S)
( )( , )
il ky
( )( , )
il ky
( i )l ,nE
( ) ( ), ,
( ),
( ),0
i il k l k
il k
il k
y if y Se
if y S
( ) ( )
, ,arg( 1) ( )
, ,
i il n l n
j X Ei il n l nX X e
( 1) ( 1) ( ), , ,arg argi i i
l n l n l nX X Determined Phase (weighting factor)Determined Phase (weighting factor)
Basic Concept of Proposed AlgorithmBasic Concept of Proposed Algorithm
0 10 20 30 40 50 60-1
0
1
2
Sampling time
Am
plit
ud
e
0 N-1
Frequency DomainN-10 1 2 3 4 5 6
IFFT
Time Domain
N-2
( ) ( ), ,
( ),
( ),0
i il k l k
il k
il k
y if y Se
if y S
S Threshold LevelThreshold Level
Basic Concept of Proposed AlgorithmBasic Concept of Proposed Algorithm
0 10 20 30 40 50 60-1
0
1
2
Sampling time
Am
plit
ud
e
0 N-1
Frequency DomainN-10 1 2 3 4 5 6
Time Domain
N-2
FFT
2 knN 1 j( i ) ( i ) Nl ,n l ,k
k 0
1E e e
N
Basic Concept of Proposed AlgorithmBasic Concept of Proposed Algorithm
0 10 20 30 40 50 60-1
0
1
2
Sampling time
Am
plit
ud
e
0 N-1
Frequency DomainN-10 1 2 3 4 5 6
Time Domain
N-2
FFT
( ) ( )
, ,arg( 1) ( )
, ,
i il n l n
j X Ei il n l nX X e
Basic Concept of Proposed AlgorithmBasic Concept of Proposed Algorithm
0 10 20 30 40 50 60-1
0
1
2
Sampling time
Am
plit
ud
e
0 N-1
Frequency DomainN-10 1 2 3 4 5 6
Time Domain
N-2
FFT
Basic Concept of Proposed AlgorithmBasic Concept of Proposed Algorithm
0 10 20 30 40 50 60-1
0
1
2
Sampling time
Am
plit
ud
e
0 N-1
Frequency DomainN-10 1 2 3 4 5 6
Time Domain
N-2
IFFT
Basic Concept of Proposed AlgorithmBasic Concept of Proposed Algorithm
0 10 20 30 40 50 60-1
0
1
2
Sampling time
Am
plit
ud
e
0 N-1
Frequency DomainN-10 1 2 3 4 5 6
Time Domain
N-2
( ) ( ), ,
( 1),
( ),0
i il k l k
il k
il k
y if y Se
if y S
S
FFT
Amplitude of OFDM signal in the time domain Amplitude of OFDM signal in the time domain
0
1
0.5
1.5
0 12080604020 100Time Sample Number
(b) (b) After T-F Swapping AlgorithmAfter T-F Swapping Algorithm
Am
pli
tud
e
0 20 40 60 80 100 1200
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0
1
3
2
0 12080604020 100Time Sample Number
(a) (a) Before T-F Swapping AlgorithmBefore T-F Swapping Algorithm
0 20 40 60 80 100 1200
0.5
1
1.5
2
2.5
3
3.5
Am
pli
tud
e
2 knN 1 j( i ) ( i ) Nl ,n l ,k
k 0
1E e e
N
( ) ( )
, ,arg( 1) ( )
, ,
i il n l n
j X Ei il n l nX X e
( 1) ( 1) ( ), , ,arg argi i i
l n l n l nX X
( 1)
,
1 1
argil n
Lj
i ln
e
L
Average for the weighting factors obtained for all symbols
Determination of Weighting FactorDetermination of Weighting Factor
Error signal in frequency Error signal in frequency domaindomain
OFDM signal in frequency domainOFDM signal in frequency domain
Determined weighting factorDetermined weighting factor
Common weighting factor over Common weighting factor over one frameone frame
(a) (a) Original OFDMOriginal OFDM Signal Signal (b) (b) Phase Rotation due Phase Rotation due to weighting factor to weighting factor
-1.5 -1 -0.5 0 0.5 1 1.5-1.5
-1
-0.5
0
0.5
1
1.5
-1
0
1
0 1-1
(a)
-1.5 -1 -0.5 0 0.5 1 1.5-1.5
-1
-0.5
0
0.5
1
1.5
0 1-1
-1
0
1
(b)
Scatter Diagram Scatter Diagram oof Preamblef Preamble and Data Symbols and Data Symbols
( 1)21
( 1) ( )
0
21( 1)
0
1
1
in
knN jji i Nk n
n
knN ji Nn
n
y X e eN
X eN
Structure of Transmitter for Proposal MethodStructure of Transmitter for Proposal Method
S/P
MOD
Xn
Data & Preamble
0
N-1
IFF
T
P/S
+GI SSPAD/A &
U/C
yk
Xi Number of
predetermined iteration or PAPR < Target
No Yes
Calculate error signal
FFTDetermination of
common weighting factor
njn n n nR X e H N
ˆ ( ) /
/
n
n
n
jn n n n n
jn n n
jn
H X e H N X
e H N X
e H
Structure of Receiver for Proposed MethodStructure of Receiver for Proposed Method
ˆ ˆ/
ˆ( ) /( )
ˆ/( )
n n
n
n n n
j jn n n n
jn n n
n
X R H
X e H N e H
X N e H
X
D/C &
A/D
Estimation
Preamble
X DEMOD DATA-GI S/P FFT P/ S1ˆ
nH
nR ˆnX
Received DataReceived Data
Estimated Channel Estimated Channel ResponseResponse
Demodulated DataDemodulated Data
Frequency Domain EqulizationFrequency Domain Equlization
n : Weighting Factor at n-th Sub-Carrier: Weighting Factor at n-th Sub-Carrier
List of Parameters List of Parameters in Performance Evaluationin Performance Evaluation
Modulation 64QAM
Demodulation Coherent
Allocated bandwidth 5MHz
Number of FFT points 256
Number of sub-carriers 64
Symbol duration 12.8 s
Guard interval 1.28 s
Non-linear amplifier SSPA
Non-linear parameter of SSPA r =2
Number of data symbols in one frame 12
Number of preamble symbols in one frame 2
Multi-path fading model
Power delay profile Exponential
Number of delay paths 16
Decay constant -1 dB
4 5 6 7 8 9 10 11 1210
-3
10-2
10-1
100
PAPR (dB)
CC
DF
(P
rob.
PA
PR
> A
bsci
ssa)
Conventional OFDMProposed method
Number of sub-carriers = 64 Frame length = 12 symbols
PAPR PerformancePAPR PerformanceModulation Method =64QAM
ConventionalConventional
ProposedProposed
PAPR Performance vs. Number of IterationsPAPR Performance vs. Number of Iterations
0 2 4 6 8 10 12 145.5
6
6.5
7
7.5
8
Number of iterations
Ave
rage
PA
PR
(d
B)
Conventional OFDMProposed method
Total sub-carriers = 64 Frame length = 12 symbols
Modulation Method =64QAM
ConventionalConventional
ProposedProposed
10 15 20 25 30 35 40 45 5010
-5
10-4
10-3
10-2
10-1
100
C/N (dB)
BE
RLinearIBO = -2dBIBO = -4dB
Multi path = 16 pathsPower Decay = -1dB Rayleight
BER PerformanceBER Performance
Modulation Method =16QAM
ConventionalConventional
ProposedProposed
IdealIdeal
10 20 30 40 50 6010
-5
10-4
10-3
10-2
10-1
100
C/N (dB)
BE
RLinearIBO = -6dBIBO = -8dB
Conventional OFDM
Proposed method
BER PerformanceBER Performance
Modulation Method =64QAM
ConventionalConventional
ProposedProposed
IdealIdeal
Proposed the PAPR reduction method for OFDM signal without side information
Use the common weighting factor over one frame including the preamble symbols
The time-frequency domain swapping algorithm is employed in the determination of common weighting factor
The common weighting factor can be removed from the received data symbols by using the frequency domain equalization
The common weighting factor can be estimated together with the channel frequency response by using preamble symbols
ConclusionsConclusions
The proposed technique can achieve the better PAPR performance and better BER performance than the conventional OFDM in the channel of non-linear amplifier and multi-path fading
ConclusionsConclusions