Indonesian J
ournal of Ele
c
trical Engin
eering and
Computer Sci
e
nce
Vol. 1, No. 3,
March 20
16, pp. 583 ~ 5
8
9
DOI: 10.115
9
1
/ijeecs.v1.i3.pp58
3-5
8
9
583
Re
cei
v
ed O
c
t
ober 2
2
, 201
5; Revi
se
d Ja
nuar
y 25, 20
1
6
; Acce
pted
February 8, 2
016
Performance of Coded Modulation in Multicarrier CDMA
System
Linda Me
y
l
a
n
i*, Nur Andi
ni, Is
w
a
h
y
udi Hida
y
a
t
Schoo
l of Elect
r
ical En
gin
eeri
ng, T
e
lkom Uni
v
ersit
y
Jl.
T
e
lekomu
ni
kasi no 1, T
r
s.
Buah Bat
u
, Bandu
ng 4
025
7, Indo
nesi
a
e-mail:l
i
n
dame
y
l
ani
@telk
o
mu
niversit
y.ac.i
d
A
b
st
r
a
ct
Coded
modulation scheme in
troduce robust and reliable
system
.
It combine coding and
mo
du
latio
n
i
n
one syste
m
. T
h
is p
a
p
e
r sho
w
performance
of cod
ed
mod
u
lati
on w
i
th co
din
g
rate
2/3,
usin
g
8PSK and co
n
v
oluti
ona
l codi
ng that
impl
e
m
ented i
n
MC CDMA system. MC
CDMA is one of multic
ar
rier
system that combi
ne OFDM and CDMA
so its offe
r relia
bl
e system in freque
ncy
selective fad
i
ng
.
Deco
din
g
proc
ess for c
o
d
ed
mo
du
latio
n
is d
one
by
usi
ng
l
og
like
lih
oo
d ra
tio for
least
sig
n
ifica
n
t bit (
defi
n
e
the subs
et) an
d the
entire
bit
s
(me
m
ber
of subset). Si
mu
l
a
tion r
e
sult
s
h
ow
n that cod
e
d
mod
u
lati
on
i
n
MC
CDMA system
have a good performanc
e in flat
fading and frequency selec
t
ive fading.
Ke
y
w
ords
: trel
lis cod
ed
mod
u
l
atio
n, eucli
de
a
n
distanc
e,
soft decisi
on, Viter
b
i dec
odi
ng,
multicarri
er
1. Introduc
tion
The m
o
st
co
nce
r
nin
g
i
s
su
es i
n
di
gital
comm
uni
cati
on a
r
e
reli
ab
ility, higher d
a
ta rate
and ban
dwi
d
th
efficie
n
cy. Relia
ble syst
em can
be
a
c
hieve
d
u
s
in
g e
rro
r
co
ntrol codin
g
su
ch a
s
convol
utional
codin
g
, turb
o codi
ng, ldp
c
, etc. Bu
t in the other si
de, the use
of erro
r control
codi
ng can cause data si
ze become
s
la
rge
r
and it
m
u
st be supp
o
r
ted with la
rg
er ba
ndwi
d
th.As
the reliability of system
s that c
an be obtainedby using error
contro
l coding, high data rate can be
obtaine
d byu
s
ing a
spe
c
ifi
c
mod
u
lation
type
. To mantain syste
m
p
e
rform
a
n
c
e, tran
smi
ssi
on
with
high d
a
ta rate
mu
st be
suppo
rted
wit
h
large b
and
width.
Unge
rboe
ck i
n
[1],
[2], [3], introduce
Trelli
s Code
d Modul
atio
n (T
CM) th
at combi
n
e
forwa
r
d
error
co
rrectio
n
usually u
s
ing
convol
utional
code an
d modulatio
n techniqu
es
. Thi
s
techni
que
can imp
r
ove
robu
stne
ss
and
reliability of system
witho
u
t requi
ring
addition
al
ba
ndwith o
r
re
duci
ng data
rate.Trellis
coded
modulatio
n combine
erro
r control codi
n
g
,
usually
co
nvolutional coding, with coding
rate
k/(k+1)
and mod
u
lati
on with 2
k+1
level. This tech
nique offers robu
st system
with limited band
width.
Whe
n
the
system
works o
n
a fre
quen
cy sele
ctive chann
els, the
received
sign
al po
we
r
level will mo
re fluctuative
andinte
r
feren
c
e inte
rsymb
o
l (ISI) be
ca
me high
er. T
hese ph
enom
enon
can
ma
ke
de
grad
ation i
n
system p
e
rfo
r
mance. O
ne t
e
ch
niqu
e that
ca
n
be
used
to mitigate
t
h
e
effect of fre
quen
cy sel
e
ctive fading
is multic
arrie
r
, inclu
de orthogon
al fre
quen
cy divisi
on
multiplexing (OFDM
)
and
multica
rrie
r
code divi
sio
n
multiplexing (MC-CDMA). OFDM
an
d MC
CDMA
will tra
n
smit data i
n
N pa
rallel
su
bca
rri
er. Carri
e
r si
gnal
on e
a
ch
su
bcarrie
r
orth
ogo
nal t
o
each oth
e
r to
avoid ICI. So
me resea
r
ch
also
have
do
ne to
analy
s
is pe
rform
a
n
c
e
TCM,
Zeh
a
vi in
[4] analyse
perform
a
nce of TCM us
ing
8 PSK in rayl
eigh fadi
ng
channel. In this paper, Zehavi
introdu
ce
d the use of bits interleaver t
o
impr
ove T
C
M pe
rforma
nce in rayleig
h
chan
nel.Ma
r
tin
Papez in [5]
simulate
TCM usi
ng
con
v
olutional
co
de sy
stem b
y
using
MAT
L
AB Simulink and
sho
w
its perf
o
rma
n
ce in AWG
N
cha
n
nels. Samree
n Amir and Muhamm
ad
Asif in [6] an
alyse
perfo
rman
ce
of TCM - MC CDMA ove
r
rayleigh fadi
n
g
cha
nel
s. [6] also de
scrib
e
the effect of
differen
c
e
co
nstrai
nts l
e
n
g
th in T
C
M.[
7
] sh
ow pe
rforma
nce T
C
M in O
F
DM
system
an
u
s
ed
Artifiacial Ne
ural
Net
w
o
r
ks (ANN)
for decodin
g
pr
oc
es
s
.
[8
] a
nalyz
e
d
the
pr
ob
le
m
o
f
jo
in
in
g
convol
utional
codi
ng-mapp
er and la
belli
ng pro
c
e
s
s.
This
pap
eral
so di
scu
s
s a
bout T
C
M e
s
spe
c
ially to
kno
w
perfo
rmance T
C
M
in MC
CDMA.T
he d
i
fferences of
this resea
r
ch with th
e p
r
eviou
s
a
r
e
model
syste
m
and
de
co
ding
methode
of T
C
M. In this
p
aper, bit
s
int
e
rleave
r
will
be u
s
ed
on t
he sy
stem of
TCM. Interl
e
a
ve
r
block pla
c
ed
betwee
n
convolutional
codi
ng and
mappe
r. De
coding p
r
o
c
e
s
s is do
ne b
y
sep
a
ratin
g
lsb and entire b
i
ts of the rece
ived symbol.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
1, No. 3, March 20
16 : 583 – 589
584
2. Model Sy
s
t
em
Model
sy
ste
m
in this pa
per are sh
o
w
n in figure
1. This model use T
r
ell
i
s Cod
ed
Modulatio
n a
s
erro
r co
ntrol codi
ng
with codi
ng
rate
2/3 and mo
dulation 8 P
SK. Trellis coded
modulatio
n in
this
system
usin
g blo
c
k i
n
terleave
r
at
the outp
u
t of
co
nvolutiona
l co
ding
befo
r
e
mapping to 8 PSK. System perfo
rmance has analyzed in flat
frequency fading
and frequency
sele
ctive fadi
ng ch
ann
el.
Figure 1. System Model
Interleave
r
is the first pro
c
ess in this sy
stem. All generated bit be
interleave
d
u
s
e blo
ck
interleave
r
wi
th block si
ze
20x3. Inteded
use of in
terl
eaver i
s
to avoid oc
cu
ren
c
e of burst error.
Input Code
d
Modulatio
n in
this
syste
m
i
s
2
pa
ralle
l
bi
ts, so
outp
u
t
of blo
c
k interl
eaver
be
divided
to 2 pa
rallel
data. The fi
rst pa
rallel
da
ta will inte
rle
a
ve agai
n u
s
ing blo
c
k int
e
rleave
r
a
n
d
the
se
con
d
be
an
input of
Con
v
olutional
co
ding
with
codi
ng
rate
½
an
d
the output of
co
nvolutio
nal
cod
e
b
e
a
n
i
nput of
blo
c
k interl
ever. F
i
gure
2
sho
w
n convolutio
n
a
l code
that
use
in
syste
m
model.
Figure 2. Con
v
olutional cod
e
(rate
1/2)
Codi
ng gen
erator that use
d
in convoluti
onal code is
g
1
(D
) = 17
1
8
= D
0
+ D
1
+ D
2
+ D
3
+ D
6
an
d
g
2
(D
) =1
33
8
= D
0
+ D
2
+ D
3
+ D
5
+ D
6
.
2.1. Trellis Coded Modul
ation
There is a differen
c
e in T
C
M mappi
ng
with convol
utional map
p
ing
.
TCM does
not use
gray code mapping but utilize natural binary
mappi
ng
as shown in figure 3 for 8PSK.
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Perform
a
n
c
e
of Code
d Mo
dulation in M
u
lticarrie
r
CDMA System
(Linda Me
yla
n
i
)
585
Figure 3. 8 PSK Cons
tellation
There a
r
e
three im
porta
nt
step
s in
T
C
M
map
p
ing.
First step
i
s
set
partition/ sel
e
ction,
in
this step, sig
nal will
devid
e
into small
subset
with
m
a
ximum e
u
cli
dean
di
stan
ce bet
wen
sy
mbols
intra sub
s
et. The second
step is la
beli
ng, to rep
r
e
s
ent the outpu
t encod
er tha
t
have maximum
eucli
dean di
stance. And t
he
last step i
s
code selection,to se
l
e
ct
the subset.In this
step, least
signifi
cant bit
form bit const
e
llation (8PSK) will used
as selector to
define t
he
subset as subset 1
( with b3 eq
u
a
l to 1) and subset 2 (with
b3 equ
al to 0) as sh
own in figure 4.
00
0
00
1
01
0
01
1
10
0
10
1
11
0
11
1
b
1
b
2
b
3
b
3
=1
b
3
=0
b
2
=0
b
2
=1
b
2
=1
b
2
=0
Figure 4. Set Partition/s
e
lec
t
ion 8PSK [9]
Re
ceived
symbol at re
cei
v
er define a
s
:
(1)
Whe
r
e r
i
s
a
received sym
bol,
s
i
s
a
tra
n
smit
symbol
and
n i
s
zero mea
n
G
a
u
ssi
an n
o
ise
with
var
i
anc
e
2
. De
codi
ng p
r
o
c
e
ss i
n
the re
ceiver
use so
ft demappin
g
scheme. T
h
i
s
sch
e
me h
a
s
2
step.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
1, No. 3, March 20
16 : 583 – 589
586
Figure 5. Defi
ne LSB bit and ith bit in subset
First
step is t
o
define le
ast
significant bi
t
form the re
ceived symb
ol
. Least si
gnifi
cant bit
will define the subset
of si
gnal.
The second step i
s
to
define i
th
bit
s
in su
bset. If we a
s
sum
e
that
all symbol
s were transmiting with the same pro
babilit
y, the receiver can
use Log likelihood ratio
(LL
R
) al
gorith
m
to define bit ith form rece
ive sign
al r. LLR d
e
fine
s as [10] [11]:
Λ
l
n
0
|
1
l
n
∑
|
∈
∑
|
∈
(2)
whe
r
e
∈
1,2,
…
,
and
define as ith bit in label symbols.
|
is define
as[10]:
|
1
2
|
|
(3)
After re
ceive
r
ha
s a
soft
bits that d
e
fine it
h bit
ent
ire
(2 bit m
s
b), it mu
st q
uantized
interleaved and be an i
n
put for trellis decoding.
The output of trel
lis decodi
ng
will combine
with
r
e
co
ns
tr
uc
te
d ls
b
b
i
ts
.
2.2 Multi Carrier CDMA
Multi ca
rri
er code
division
multiple a
c
ce
ss is
a
syste
m
that
combi
ne ofdm
tech
nique
and
CDMA
syste
m
. There a
r
e two gro
u
p
s
of MC CDM
A,the first one, sp
rea
d
s the original
bit
messag
e
with
sp
re
ading
co
de
seq
uen
ce,
and
the
othe
r g
r
ou
p
sp
rea
d
seri
al to
parallel
conve
r
te
r
bit messa
ge
with spreadi
n
g
co
de
se
qu
ence [12].
T
h
is
system
u
s
ed fi
rst type
of MC
CDM
A
.
Figure 5 sh
o
w
n MC
CDM
A
system
that used in mo
d
e
l system.
Figure 6. MC
CDMA transmitter
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Perform
a
n
c
e
of Code
d Mo
dulation in M
u
lticarrie
r
CDMA System
(Linda Me
yla
n
i
)
587
Table 1. Simulation Para
meter
Parameter
Value
Modulation t
y
pe
8 PSK
Coding Rate
Cod
ed Modulation
2/3
Coding rate conv
olutional code
½
Gene
rator convol
utional code
g
1
(D
)=171
8
= D
0
+
D
1
+ D
2
+ D
3
+
D
6
g
2
(D
) =133
8
= D
0
+
D
2
+ D
3
+ D
5
+ D
6
Number subcar
rier
128; 256
Cy
clic pr
efix
¼
Channel fading
Flat Ra
y
l
eigh fad
i
ng;
frequenc
y selective Ra
y
l
eigh fadin
g
3. Results a
nd Analy
s
is
To see the
perfo
rman
ce
cod
ed m
o
d
u
lation o
n
cre
a
ted mo
del
system,the
system is
tested with
se
veral scen
ari
o
s:
a. Compare
performan
ceof CM-M
CCDM
A system wit
h
4PSK and 8PSK MC CDMA system
(with/ witho
u
t convolutio
nal
codin
g
) at AWG
N
ch
ann
e
l
b
.
C
o
mpa
r
e
per
fo
r
m
an
ce
CM-
M
CCD
MA s
y
s
t
em
at
Raylei
gh f
ading
chan
n
e
l: freq
uen
cy
sele
ctive fadi
ng and flat fading
c.
Comp
are perf
o
rma
n
ce CM
-MC CDMA at
difference nu
mber
sub
c
a
rri
er.
Figure 7. Co
mpare Performance
of cod
ed modul
atio
n in AWG
N
Samree
n Ami
r
in [6]
sho
w
t
hat co
nst
r
ant
s len
g
th give
effect to the
p
e
rform
a
n
c
e
o
f
TCM.
High
er co
nst
r
aints l
ength
have b
e
tter
p
e
rform
a
n
c
e
ra
ther than
lo
wer
con
s
trai
n
t
s len
g
th. In t
h
is
resea
r
ch, TCM is de
sig
n
e
d
by usi
ng 7
as
con
s
trai
nts len
g
th. Figu
re 7
sho
w
th
e perfo
rma
n
ce of
cod
ed mo
dul
ation sche
me
using 8 PS
K in MC CDMA in AWG
N
ch
ann
el wi
th 128 sub
c
a
rrie
r
compared to
MC
CDMA using modulat
on
with 4
PSK without
usi
ng
convol
utional
codi
ng, and 8
PSK without using convolutional c
oding,
8 PSK using
convol
utional
codi
ng (with
coding
rate 2/3)
and 8PSK
wi
th half data
using convol
utional
codi
ng
(with
coding rate ½)
. T
h
is figure shown
coded m
odul
ation
scheme usi
ng
8PSK have good
perform
a
nce t
han
others
but still below
the
performanc
e
of 4PSK. This is
c
a
used
by the us
ed of
soft demapping to es
timate
bit i
th
value fro
m
the sub
s
et. I
n
soft dem
ap
ping
and
de
coding
p
r
o
c
ess,
symbol
re
ceived
will
e
s
timate
2 m
s
b in
subset like 4PSK (figure 5) and quanti
z
ed the estima
te bit. Quantized bit be an input for trellis
decoding. Symbol received will
esti
mat
e
lsb of the subset. After de
interlever process, estimated
lsb bit and o
u
t
put of trellis decodin
g
(2 b
i
ts) will combi
ned to get re
contru
cted bits.
0
2
4
6
8
10
12
10
-4
10
-3
10
-2
10
-1
10
0
Eb
/
N
o
BER
mc
c
d
ma
4
p
s
k
mc
c
d
ma
c
m
8
p
s
k
mc
c
d
ma
8
p
s
k
m
c
c
d
m
a
8p
s
k
+
c
c
(
2/
3)
m
c
c
d
m
a
8p
s
k
+
ha
l
f
da
t
a
c
c
(
1
/
2
)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
1, No. 3, March 20
16 : 583 – 589
588
Perform
a
n
c
e
of co
ded
m
odulatio
n in f
l
at
fading
an
d freq
uen
cy
sele
ctive fadi
ng have
s
h
ow
n in
figur
e
8
.
T
h
is
figu
r
e
sh
ow
th
a
t
p
e
r
f
o
r
ma
n
c
e
of
cod
ed
mo
dulation
in fla
t
fading
is bet
ter
than code
d modulatio
n
i
n
freq
uen
cy sele
ctive
fadi
ng. Thi
s
i
s
cause by m
u
ltipath fading
in
freque
ncy
sel
e
ctive fadin
g
. Multipath wit
h
differe
n
c
e
g
a
in an
d del
ay, make
sig
nal
re
ceived m
o
re
fluctuative an
d more
su
sce
p
tible with noi
se. A
nd it cau
s
e pe
rform
a
n
c
e sy
stem be
come
wo
rse.
Figure 8. Performa
nce co
d
ed modul
atio
n
Figure 9. Performa
nce co
d
ed modul
atio
n at fr
equen
cy selective fa
ding with diffe
ren
c
e nu
mbe
r
s
u
bc
ar
r
i
e
r
Figure 9, sh
own th
e pe
rf
orma
nce of
cod
ed m
odul
ation in m
c
cdma
with
di
fference
numbe
r of su
bca
rri
er at fre
quen
cy sel
e
ct
ive fadi
ng. System si
mulat
ed in 12
0km/
hour. Thi
s
fig
u
re
sho
w
syst
em
with numbe
r sub
c
arrie
r
128 have
better perf
o
rma
n
c
e than nu
m
ber of su
bcarrier
256. L
a
rg
er
numbe
r
of su
bca
rri
er ca
n
make
b
and
wi
dth on
ea
ch
sub
c
a
rri
er
m
o
re
sm
aller a
n
d
0
5
10
15
10
-4
10
-3
10
-2
10
-1
10
0
Eb
/N
o
BER
f
r
eq
uen
c
y
s
e
l
e
c
t
i
v
e
f
a
di
ng
f
l
at
f
a
d
i
ng
0
2
4
6
8
10
12
14
10
-4
10
-3
10
-2
10
-1
10
0
BE
R
Eb
/N
o
mc
c
d
ma
1
2
8
cm
-
m
c cd
m
a
1
2
8
mc
c
d
ma
2
5
6
cm
-
m
c cd
m
a
2
5
6
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Perform
a
n
c
e
of Code
d Mo
dulation in M
u
lticarrie
r
CDMA System
(Linda Me
yla
n
i
)
589
make
system
wo
rk li
ke i
n
flat cha
nnel.
But larg
er nu
mber of
sub
c
arri
er
can m
a
ke
inte
rcarrie
r
interferen
ce h
i
gher, so it ca
n make d
e
g
r
a
dation in sy
stem perfo
rma
n
ce.
4. Conclusio
n
Cod
ed mod
u
l
a
tion sche
me
in MC CDM
A
system ha
s a goo
d pe
rforma
nce co
mpare to
another
scheme 8 PSK i
n
mc
cdm
a
. In
AWGN
channel, coded mo
dulation can
get 10
-3
at 7.
2 dB
(Eb/No level
)
difference 0.
8 dB with 8PSK - conv
ol
utional coding (with
coding rate 2/3) and
differen
c
e
10.
2 dB
with 8P
SK usin
g
con
v
olutional
co
de (co
d
ing
ra
te ½) fo
r h
a
lf data. Usin
g L
o
g
Likeli
hood
Ra
tio (LLR) in soft demappin
g
and de
co
di
ng, and ma
ke
system pe
rfo
r
man
c
e hi
ghe
r.
Referen
ces
[1]
G Ungerb
oeck.
“Chan
ne
l Co
d
i
ng
w
i
t
h
Multi
l
e
v
el / Phase S
i
gna
ls”.
IEEE Trans. Inf. Theory
. 1982; IT
-
28(1): 55
–6
7.
[2]
G Ungerb
oeck
.
“T
rellis-Cod
e
d
Modu
latio
n
w
i
t
h
Red
u
n
d
a
n
t Signa
l Sets Part I: Introduction”.
IEEE
Co
mmun
icati
o
ns Maga
z
i
ne
. 1
987: 5–
11.
[3]
G Ungerb
oeck
.
“T
rellis-code
d
modul
ation
w
i
th redun
da
nt sign
al sets Part II: State of
the Art”.
I
E
EE
Co
mmun
icati
o
ns Maga
z
i
ne
. 1
987; 25(
2), 12–
21.
[4]
E Zehavi. “8-P
SK trellis c
ode
s for a Ra
yl
ei
g
h
fadi
ng ch
an
n
e
l”.
{IEEE} Trans. Comm
un.
199
2; 40(
5):
873
–8
83.
[5]
M Papez an
d M Cico. “Data T
r
ansmission b
y
T
r
ellis
Co
de
d Modu
latio
n
u
s
ing C
onvo
l
uti
on Co
des”.
i
n
Rece
nt Advanc
es in Auto
mati
c Control, Infor
m
ati
on a
nd C
o
mmu
n
icati
ons
,
201
3: 227–
23
2.
[6]
S Amir and M Asif. “Anal
y
t
ic
a
l
Stud
y
of MC-
CDM
A-T
C
M o
v
er Multi-path
Ra
yle
i
g
h
F
adin
g
Cha
nne
ls”.
IPCSIT
. 2011; 2: 332–
33
7.
[7]
A Misra a
nd K
K
Sarma. “T
CM-code
d OF
DM assisted
b
y
ANN i
n
W
i
rel
e
ss Cha
n
n
e
ls”.
Int. J. Sm
art
Sensors Ad H
o
c Netw
orks
. 2012; 1(3): 50
–55
.
[8]
A Alvar
ado,
A
G
i Amat, F
Brannstrom,
a
nd E
Agr
e
ll. “
O
n Optimal
T
C
M Enc
oders”
.
IEEE Trans.
Co
mmun.
, 20
1
3
; 61(6): 21
78–
218
9.
[9]
S Benedetto, S Memb
er, a
nd G M
ontors
i
. “
Perfor
ma
nc
e Eva
l
uati
o
n
of Trellis-
C
o
d
e
d
Mo
dul
atio
n
Schem
e
s
”. Pro
c
. IEEE. 1994;
82(6): 83
3–
855
.
[10]
L Wang, D
X
u
, and
X
Zhan
g
.
“Recursive b
i
t me
tric gener
ation for PSK
sign
als
w
i
th
gr
a
y
lab
e
li
ng”.
IEEE Commun. Lett.
2012; 16
(2): 180–
18
2.
[11]
S R
y
oo, S Kim
,
and SP L
ee.
“Efficient Soft De
ma
ppi
ng Me
thod for H
i
gh
Order Mod
u
lati
on Sch
e
mes”
.
in
CDMA Inter
natio
nal C
onfer
ence
. 20
03, pp
. 2–6.
[12]
S Hara an
d R Prasad. “O
verv
ie
w
of Multic
arrier CDMA”.
IEEE Communic
a
tions Maga
z
i
n
e
. 1997: 1
26–
133.
Evaluation Warning : The document was created with Spire.PDF for Python.