年季学期研究生课程考核
（读书报告、研究报告）
考核科目凸优化理论及应用
学生所在院（系）学生所在学科学学学考生核类结生姓名号别果阅卷人
第1页（共5页）
fTheReportofco
vexoptimizatio
course
PARMi
imizatio
ofOFDMSig
alsbyCo
vexOptimizatio
1Thei
troductio
oftheproblemPeaktoAveragePower
Reductio
ofOFDMSig
als11Thereaso
oftheproblemge
erated
Highba
dwidthefficie
cya
drobust
essagai
stfreque
cyfadi
garethefeaturesofOFDMsig
alsduetomultipathpropagatio
HoweverhighpeaktoaverageratioPARofOFDMsig
alslimitsperforma
ceofpoweramplifierPABecauseofthisPARshouldbereducedi
ordertokeepthePAoperati
gwithli
earlevelwhichwillresulti
lossesofpowerefficie
cyI
awordPAPRofOFDMsig
alsshouldbereducedNamelytomi
imizethePAR12TheoverviewofthePARmi
imizatio
A
ewpowerspectrumdistributio
wasdesiredtobedevelopedbyoptimizi
gthetimedomai
peakpowersubjecttoerrorvectormag
itudeEVMco
strai
tsPowerthesig
aloveritssubcarriersshouldberedistributedi
cludi
gthepowerfreeo
esItisprocessedbyapowerredistributio
ofthesig
aloveritssubcarriersi
cludi
gthepowerfreeo
es
2PARmi
imizatio
i
mathematics
AssumethatOFDMsig
alhas48data4pilot12freesubcarriersa
d128symbolsThe52subcarriersdatapilotweremodulatedusi
gQAM
cC
0


de
otesa
idealOFDMco
stellatio

cC


de
otesa
OFDM
co
stellatio
tra
smitteda
d
xC

l
de
otesthecorrespo
di
gtimedomai
sig
al
obtai
edbyIFFTwithltimesoversampli
g
cisa
i
subsetofavailablecarriersof
OFDMe
codi
ga
dii1i2idd
where
64d52co
sta
tβshouldbe
fadjustedfora
yFFTscali
g
p
0
isistheaveragepowerofthecarrier16QAM
modulatio
SR
isadiago
alcarrierselectio
matrix
1withdata1S0others
ii
Accordi
gtothepri
cipleofOFDMsystemxca
beexpressedas
xIFFTlc
PARca
beexpressedas
2
PARmax
i1
l


EVMca
beexpressedas
2
xiSc
2
2
3
EVM

1d

idii1
cici4p
00
2
With1a
d4theEVMco
strai
tca
beexpressedo
cas
SccEVM
0
max
dp
0
5
Tosumuptheoptimizatio
problemis
mi
imizemax
i1
l


Subjectto
SccEVM
0max
2
xiSc
2
2
dp
0
3Co
vexformulatio
ofPARmi
imizatio
Becausethepracticalpowershouldbelargertha
theidealo
esothereisa
otherpowerco
strai
t
ScSc
2
2
0
6
A
ditisequivale
tto
fRcccc
e00

Scc
0
2
2
Accordi
gto5itca
expa
dthelimitsas
Rcscc
e00


2
2
Fi
allyaccordi
gtothek
owledgeofco
vexoptimizatio
wehavelear
editca
beformulatedr