Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
9
, No
.
5
,
Octo
ber
201
9
, pp.
3695
~
37
00
IS
S
N:
20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v
9
i
5
.
pp36
95
-
37
00
3695
Journ
al h
om
e
page
:
http:
//
ia
es
core
.c
om/
journa
ls
/i
ndex.
ph
p/IJECE
The ch
an
nel shortenin
g comp
aris
on betwe
en OFD
M
and
MC
-
CD
MA Syst
em over
an A
DSL
chann
el
usin
g a
wat
er
-
filli
ng
F. Boua
sri
a,
R.
Ber
ber,
S. Yousfi
La
bora
tor
y
of
Know
le
dge
Man
a
gement
and
Co
m
ple
x
Data
,
Uni
ver
sit
y
Dr.
Moul
a
y
T
aha
r
of
Sa
ïd
a
,
Alg
eria
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Feb
4
, 2
01
9
Re
vised A
pr
12
, 2
01
9
Accepte
d
Apr
2
5
, 201
9
In
thi
s pa
p
er
our
objecti
v
e
foc
use
s on
the
d
evelop
m
ent
of
digita
l
tr
ansm
ission
te
chn
ique
s
in
or
der
to
solve
the
proble
m
of
Inter
-
S
y
m
bol
Inte
rf
e
ren
ce
(ISI)
due
to
m
ultiple
pat
hs
and
the
i
m
prove
m
ent
of
Multi
-
Carr
i
er
C
ode
Division
Multi
ple
Acc
ess
(MC
-
CDM
A)
s
y
stem
per
form
anc
e
.
Th
e
both
al
gori
thm:
MERRY
al
gor
it
hm
(Multi
c
ar
r
ie
r
Equa
l
izatio
n
b
y
Resto
rat
ion
of
Redunda
nc
y
)
c
alled
TE
Q
(Ti
m
e
dom
ai
n
EQua
li
z
er)
and
the
power
al
loc
a
ti
o
n
m
et
hod
to
al
lo
c
at
e
appr
opr
ia
t
e
power
to
each
u
ser
(W
at
er
-
F
i
ll
in
g
techniqu
e
(W
F))
[1]
is
use
d.
Finall
y
,
the
n
um
eri
ca
l
result
s
show
tha
t
combi
ning
W
at
e
r
Fill
ing
wi
th
TEQ
te
chn
ique a
l
lo
ws
to
red
u
ce
the
ISI for
MC
-
CDM
A sy
st
ems
.
Ke
yw
or
d
s
:
ADSL c
ha
nn
el
MC
-
CDMA
MER
RY alg
or
i
thm
OFDM
TEQ
e
qual
iz
er
Water
-
Fil
li
ng
Copyright
©
201
9
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
F.
Bo
uas
ria
,
Lab
or
at
ory
of
Kno
wled
ge
M
anag
em
ent and
Com
plex
Data
,
Dep
a
rtm
ent o
f El
ect
ro
nics
,
Un
i
ver
sit
y D
r.
Mou
la
y Ta
har
of Saïda
,
Alge
ria.
Em
a
il
:
valencfa
ty
@yahoo.fr
1.
INTROD
U
CTION
W
i
reless
c
omm
un
ic
at
ion
is
consi
der
e
d
to
be
th
e
fa
ste
st
grow
i
ng
se
gme
nt
of
t
he
c
om
m
un
ic
at
ion
s
industry.
Ce
ll
ular
syst
em
s
hav
e
gro
w
n
f
ast
er
in
the
la
st
decad
e
an
d
cu
rrentl
y
ha
ve
tw
o
bill
ion
us
e
r
s
world
wide
.
Be
cause
of
the
be
nef
it
s
of
wire
le
ss
com
m
un
icati
on
s,
var
i
ou
s
m
ajo
r
te
le
phone
ser
vice
pro
vide
r
s
arou
nd
the
w
orl
d
ha
ve
la
unc
hed
the
ra
dio
t
el
ephon
e
.
T
he
gro
wing
nu
m
ber
of
s
ubscri
ber
s
an
d
t
he
ne
ed
to
pro
vid
e
new
hi
gh
-
qu
al
it
y
servi
ces
hav
e
le
d
t
o
the
em
erg
e
nc
e
of
su
cce
ssiv
e
cel
l
syst
e
m
s
that
can
be
cl
a
ssifie
d
into
dif
fer
e
nt
ge
ner
at
io
ns
.
Re
centl
y,
broadb
and
C
DMA
syst
e
m
s
hav
e
be
en
d
e
velo
ped
a
s
a
3G
com
petit
or
to
increase
data
rates
for
the
us
er
.
H
ow
e
ve
r
,
the
la
rg
e
ba
ndwidt
h
of
th
ese
high
-
s
pee
d
li
nk
s
m
akes
the
m
vu
l
ner
a
ble
to
I
nter
-
Sym
bo
l
In
te
rf
e
re
nce
(IS
I)
.
C
DMA
an
d
OF
DM
are
not,
ind
e
pe
nden
tl
y,
able
to
m
e
et
this
requirem
ent.
The
MC
-
CD
MA
schem
e
(
m
ul
ti
ple
acce
s
s
div
isi
ble
by
m
ulti
-
carrier
cod
e
)
m
akes
i
t
po
ssible
to
si
m
ultaneou
sly
exp
l
oit
the
adv
a
ntage
s
of
CDMA
a
nd
OFDM,
a
m
ul
ti
-
carrier
m
od
ulati
on
sc
hem
e.
The
MC
-
C
DMA
syst
em
is
con
side
red
one of
the p
ote
ntial
cand
i
dates
co
nsi
der
e
d
to b
e
c
a
pab
le
o
f
m
eet
ing
th
e
requirem
ents
of
ne
xt
-
gen
e
ra
ti
on
(
4G)
[1
-
6]
hig
h
-
s
peed
wireless
m
ultim
edia
com
m
u
nicat
ion
syst
e
m
s
and
Lo
ng Tr
em
Evoluti
on
(LTE
).
In
tr
oduci
ng
t
he
rad
io
phone.
The
gr
ow
i
ng
num
ber
of
s
ubs
cribers
an
d
the
need
t
o
pro
vi
de
ne
w
hi
gh
qu
al
it
y
ser
v
ic
es
ha
ve
le
d
t
o
the
em
erg
ence
of
s
uccessi
ve
cel
l
syst
e
m
s
t
hat
can
be
cl
a
ssifie
d
int
o
di
f
fer
e
nt
gen
e
rati
ons.
R
ecentl
y,
broa
dban
d
C
DMA
s
yst
e
m
s
hav
e
be
en
de
velo
pe
d
as
a
3G
c
om
petit
or
to
incr
eas
e
data
rates
for
the
use
r.
H
oweve
r,
the
la
rg
e
band
width
of
the
se
hig
h
-
s
pee
d
li
nk
s
m
akes
the
m
vu
ln
era
ble
to
In
te
r
-
Sy
m
bo
l I
nter
fe
ren
ce
(ISI).
CDMA a
nd
OFDM are
not, i
ndepe
ndently
, a
ble to m
eet
this r
e
qu
irem
ent.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
5
,
Oct
ober
201
9
:
3
6
9
5
-
3
7
0
0
3696
The
la
tt
er
will
increase
the
a
m
plit
ud
e
of
t
he
sig
nal
each
tim
e
i
t
is
a
tt
e
nu
at
e
d.
T
his
docum
ent
is
orga
nized
as
f
ollows.
Sect
io
n
II
discu
sses
the
syst
e
m
m
o
del
an
d
the
de
scriptio
n
of
th
e
MER
RY
al
go
rithm
(r
e
dundancy
r
est
or
at
io
n
m
ul
ti
-
carrier
e
qu
al
iz
at
ion
)
[
7,
8]
.
Sect
ion
III
pr
ese
nts
the
optim
iz
at
ion
of
water
fill
ing
(
WF).
Sect
ion
I
V
as
sesses
a
com
par
iso
n
betwee
n
the
pro
po
se
d
so
luti
on
ap
plied
to
MC
-
CDMA
(
WF
.MC
-
CD
MA)
with
TE
Q
eq
ualiz
er
an
d
co
nv
e
ntio
nal
MC
-
CDMA
us
in
g
TE
Q
by
m
eans
of
nu
m
erical
si
m
ulati
on
s.
Fi
nally
, s
ect
ion
V pr
ese
nts s
om
e con
cl
usi
on
s.
2.
SY
STE
M MO
DEL A
N
D DESC
RIPTIO
N OF
MER
RY
ALGO
RITH
M
This
sect
io
n
de
rives
the
basi
c
MER
RY.
Wh
at
com
e
la
te
r
will
discuss
va
rio
us
ge
neral
iz
at
ion
s
a
nd
perform
ance
-
e
nh
a
ncin
g
e
xtensions.
At
th
e
beg
i
nn
i
ng,
consi
der
a
MC
-
CDMA
s
yst
e
m
in
Fig
ur
e
1.
Af
te
r
the C
P in
serti
on, th
e
last
ν
sam
ples are
identic
al
to firs
t ν
sam
ples in
t
he kth
sym
bo
l [8
-
10]
, i.e
.:
x (Mn+ i
)
=
x (
Mn + i +
N)
i
∈
{1
,2,…,
v}
a
nd n
=
0,1,…
(1)
The
sym
bo
l
pe
rio
d
is
M
=
N
+
v,
an
d
n
is
t
he
sy
m
bo
l
in
de
x.
To
sim
plify
t
he
no
ta
ti
on,
he
ncefort
h
we
assum
e n
=0 (
w
it
ho
ut l
os
s
of
ge
ner
al
it
y). T
he recei
ve
d data r
is o
btained
from
x
b
y
r
k
(i)
=
∑
h
(
l
)
.
x
k
(
i
−
l
)
+
n
(
i
)
L
h
−
1
l
=
0
(2)
a
nd the e
qual
iz
er
data y
is si
m
il
arly
o
btaine
d from
r
b
y
y
k
(i)
=
∑
w
(
j
)
.
r
k
(
i
−
j
)
T
−
1
l
=
0
(3)
Wh
e
re
T
is
t
he
le
ngth
of
the
eq
ualiz
er
w.
The
c
om
bin
ed
cha
nnel
is
de
no
te
d
by:c
=
h*w
.
Fi
gure
2
sh
ows
a
n
e
xa
m
ple
with
N
=
8,
v=
2,
a
nd
M
=
N
+
v
=
10
[
11]
.
Co
nsi
der
this
e
xam
ple.
T
he
broa
dc
ast
ed
sam
ples
2
and
10
are
the
sam
e.
But,
at
the
receiver,
the
int
erf
e
rin
g
sam
pl
es
earli
er
than
sam
ple
2
are
no
t
the
whole
ide
ntica
l
to
their
c
orres
pondence
s
be
f
or
e
sam
ple
10
.
If
h(2
),
h(3
)
a
nd
h(4
)
wer
e
z
ero,
the
n
r(2)
=
r(1
0).
If
we
t
ry
to
for
ce
r(2
)
=
r(1
0)
,
we
m
ay
fo
rce
h(2)
=
h(
3)
=
h(
4)
=
0,
f
orci
ng
the
cha
nnel
sel
ect
ed
the
sam
e
li
ke
the
CP.
The
l
oc
at
ion
of
the
window
of
v
non
-
zer
o
ta
ps
c
an
be
var
ie
d
by
com
par
ing
r
(
3)
to
r(1
1)
,
or
r(4)
to
r(12),
et
c.
In
ge
ner
al
,
if
the
channel
le
ng
t
h
L
h
+1
≤v,
in
tha
t
case
the
pr
eviou
s
sam
ple
inside
the
CP
m
us
t
loo
k
li
ke
the
pr
e
vious sam
ple insi
de
the
sym
bo
l.
Un
it
c
os
t f
unct
ion
t
hat r
e
pro
duce:
J
MERRY
(
δ
)=E
[
|
y
(
v
+
δ
)
−
y
(
v
+
N
+
δ
)
|
2
]
δ
∈
{
0
,
…
,
M
−
1
}
(4)
Wh
e
re
δ
is
t
he
sym
bo
l
sy
nchr
on
iz
at
io
n
par
am
et
er,
w
hi
ch
re
presents
the
desire
d
delay
of
th
e
channel
–
TEQ
com
bin
at
ion
.
Figure
1. Cha
nnel
s
hortenin
g wit
h
a
blin
d
e
qual
iz
at
ion
for M
C
-
CDMA
-
WF
syst
em
Evaluation Warning : The document was created with Spire.PDF for Python.
Int
J
Elec
&
C
om
p
En
g
IS
S
N:
20
88
-
8708
Th
e c
hannel
s
horte
ni
ng com
pariso
n betw
een
O
F
DM
&
MC
-
CDMA
S
y
ste
m over
an A
DSL
…
(
F. B
ouas
ria
)
3697
Fig
ure
2
.
Illust
rati
on of the
c
ha
ng
e
in
t
he
IS
I
at
the end o
f
th
e receive
d
sym
bo
l a
nd at t
he
r
ecei
ved
C
P
The
MER
RY
al
gorithm
[1
2
-
17
]
,
perform
s
a
stochastic
gr
adient
de
scent
of
(4),
with
a
const
raint
to
avo
i
d
the
trivia
l solutio
n w =
0.
The
MER
RY alg
ori
thm
is:
(5)
w
he
re
r(n)
=
[r(n), r(
n
-
1),
…, r(
n
-
L
w
)]
T
, (*)
denotes c
om
plex
c
onjug
at
i
on, and
µi
s
the T
EQ
a
da
ptati
on
ste
p
siz
e, it’s take
n at
0
,
75.
3.
WATER
-
FIL
LING
OPTI
MIZ
ATION
Water
-
Fil
li
ng
al
gorithm
exam
ines
per
-
to
ne
SN
R
an
d
resul
t
in
per
-
t
on
e
nu
m
ber
of
bits
and
e
ne
rg
y
le
vels that s
ho
uld
be
opti
m
a
lly use
d
as s
how
n
in
Fig
ure
3
.
To
m
axi
m
iz
e
t
he
data
rate,
R
=
b/T
,
for
a
s
et
of
pa
rall
el
s
ub
c
ha
nn
el
s
w
he
n
the
sym
bo
l
rate
1/T
is
fixe
d,
re
quires
m
axi
m
iz
ation
of
t
he
achie
va
ble
b
=
∑
b
n
n
over
b
unde
r
a
gi
ven
total
inp
ut
e
ne
rg
y
a
nd
a
ta
rg
et
prob
a
bili
ty
o
f
e
rror. S
pe
ci
fical
ly
, th
e num
ber
of
bits a
ll
oca
te
d
to
the
nth
subc
hannel
is [1
-
2]:
=
2
(
1
+
)
(6)
wh
e
re
2
2
2
n
n
n
H
g
repres
ents
the
s
ub
c
ha
nn
el
si
gn
al
-
to
-
noise
rati
o
w
hen
t
he
tra
nsm
itter
app
li
es
a
un
it
energy
to
t
hat
subcha
nnel
.
The
rati
o
g
n
is
a
fixe
d
functi
on
of
the
c
ha
nnel
,
bu
t
ε
n
,
w
hich
de
no
te
s
t
he
2
-
D
su
bsy
m
bo
l
ener
gy
al
lott
ed
to
the
n
th
s
ubcha
nnel
,
can
be
opt
i
m
iz
ed
to
m
axi
m
iz
ed
b
,
sub
j
e
ct
to
a
total
transm
i
t
energy c
on
st
raint of:
∑
≤
=
1
(7)
Since
lo
g(1
+x)
is a stric
tl
y i
nc
reasin
g functi
on
of
x
, t
he
total
en
e
rg
y c
onstra
int o
f
(
7)
will
be bin
ding,
i.e., e
qu
al
it
y i
s
m
et
. U
sin
g
La
gr
a
nge m
ulti
pli
ers,
t
he
c
os
t f
unct
ion bec
om
es
N
n
n
x
n
n
n
g
1
1
ln
2
ln
1
(8)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
5
,
Oct
ober
201
9
:
3
6
9
5
-
3
7
0
0
3698
Diff
e
re
ntiat
ing wit
h res
pect to
ε
n
pro
duces
g
g
n
n
n
2
ln
1
(9)
Th
us
, t
he
a
ggr
egate bit
rate i
n b is m
axi
m
ized
wh
e
n
t
he opt
i
m
u
m
su
bch
a
nnel
tra
ns
m
it
en
erg
ie
s
sat
isfy
ε
n
+
Г
g
n
=
K
=
con
stant
(10)
a
nd
K
is c
ho
se
n
s
uc
h
that t
he t
otal ener
gy c
onstrai
nt
give
n
by (7) i
s m
et
[
1
-
2]
Fig
ure
3.
W
at
e
r
-
F
il
li
ng alg
or
i
thm
4.
SIMULATI
O
N RESULTS
We
pe
rfor
m
nu
m
erical
si
m
u
l
at
ion
s
of
I
nter
Sy
m
bo
l
In
te
r
fe
ren
ce
(
ISI)
.
Th
e
evo
luti
on
of
the
IS
I
as
a
functi
on
of
the
nu
m
ber
of
it
er
at
ion
s
al
lows
us
to
know
how
to
evaluate
an
d
com
par
e
the
perform
ances
of
the
syst
e
m
s
in
te
rm
s
of
their
spe
ed
of
c
onve
r
gen
ce
.
T
he
I
S
I
is
obta
ine
d
by
va
ryi
ng
the
nu
m
ber
of
it
e
rati
on
s
(the
nu
m
ber
of
sy
m
bo
ls)
tha
t
is
equ
al
to
1000.
Fi
gure
4
sh
ows
a
pl
ot
of
th
e
ISI
vs.
N
um
ber
of
it
era
ti
on
s.
It
sh
ows
sim
ulati
on
res
ults
usi
ng
Water
-
Fil
lin
g
a
s
ta
nda
rd
ADSL
te
st
cha
nn
el
.
T
he
FF
T
siz
e
was
512,
the
CP
le
ng
th
w
a
s
32, t
he
TE
Q had 1
6
ta
ps, a
nd sig
nal to n
oise ra
pport
was 40 dB
.
Figure
5
show
s
that
Water
-
F
il
li
ng
can
co
nver
ge
ra
pid
ly
pro
vid
e
a
pe
rfec
t
res
ult
for
MC
-
CDM
A
syst
e
m
s
wh
ic
h
com
bin
es
m
ulti
-
carr
ie
r
m
odulati
on
(OFD
M)
an
d
Mult
iple
Co
de
Di
vi
sion
M
ulti
ple
Access
(CDMA
).
I
n
th
is
fig
ur
e
,
we
noti
ce
that
the
s
yst
e
m
MC
-
CDMA.TE
Q
reac
hes
t
he
value
’
IS
I
de
(
-
4.7
5d
B)
an
d
sta
y
fixed
in
t
hi
s
value.
In
paral
le
l,
the
MC
-
CDMA.T
EQ
&
W
F
ac
hieve
a
good
pe
rform
ance
wh
i
c
h
le
ads
to
a
value o
f
’
ISI
N
en
tre
(
-
8 et
-
9dB).
Fig
ure
4
.
Per
f
orm
ance o
f OF
DM
-
TE
Q Vs
MC
-
CDMA
-
TEQ
s
yst
e
m
in
te
rm
of I
S
I
Fig
ure
5
.
Per
f
orm
ance o
f OF
DM
-
TE
Q Vs
MC
-
CDMA
-
TEQ sy
stem
in
te
rm
o
f
I
SI
,
us
ing
t
he
W
at
er
-
F
il
li
ng
Evaluation Warning : The document was created with Spire.PDF for Python.
Int
J
Elec
&
C
om
p
En
g
IS
S
N:
20
88
-
8708
Th
e c
hannel
s
horte
ni
ng com
pariso
n betw
een
O
F
DM
&
MC
-
CDMA
S
y
ste
m over
an A
DSL
…
(
F. B
ouas
ria
)
3699
Her
e
,
we
co
m
par
e
bo
t
h
s
yst
e
m
s
of
tra
ns
m
issi
on
OFDM
an
d
MC
-
CDMA
achie
ved
f
or
th
e
equ
al
iz
at
io
n
te
chn
ic
T
EQ
.
We
see
that
that
the
seco
nd
syst
e
m
giv
e
bette
r
resu
lt
tha
n
the
oth
e
r
syst
em
,
t
his
i
s
du
e
to
the
ad
va
ntages
of
us
i
ng
MC
-
CDM
A
w
her
e
the
re
is
a
rem
ark
abl
e
dif
fer
e
nce
beca
use
the
OFDM
be
gin
s
to
co
nver
ge
from
an
it
erati
on
c
ount
of
30
0,
but
the
othe
r
syst
em
has
al
read
y
c
onverged
f
or
a
nu
m
ber
of
it
erati
on
s
of 50
.
No
te
that
t
he
two
c
urves
are
well
i
m
pr
oved
than
ks
to
t
he
us
e
of
t
he
te
ch
nique
of
Water
-
Fil
li
ng
.
We
al
so
note
that
t
he
MC
-
CDMA
appro
ac
hes
th
e
op
ti
m
u
m
in
t
erm
so
f
IS
I
an
d
converges
to
w
ard
s
a
reali
zat
ion
of
a syst
e
m
m
or
e
eff
ic
ie
nt t
han that o
f
t
he OF
D
M.
The
represe
ntati
ve
curves
of
t
he
f
our
m
et
hods
in
te
r
m
s
of
IS
I
a
re
com
par
ed
re
sp
ect
ively
in
Figures
6
.
We
hav
e
noti
ced tha
t t
he
curves
of the MC
-
C
D
MA syst
em
are
m
uch
bet
te
r
th
an
that
of
the
OFDM,
and
the
us
e
of
the
W
at
e
r
-
F
il
li
ng
al
gorithm
i
m
pr
ov
es
t
he
re
su
lt
s
sig
nifican
tl
y.
But
the
MC
-
CDMA
syst
e
m
is
sti
ll
the
m
os
t effici
ent.
Fig
ure
6.
Per
f
orm
ance o
f
m
ulti
carrier
t
ran
sm
i
ssion f
or
di
ff
e
r
ent sce
nar
io
5.
C
O
NC
L
US
I
O
N
In
this
wor
k,
we
disp
la
ye
d
a
W
at
er
F
il
li
ng
a
rr
a
ng
em
ent
strat
egy
for
MC
-
CDMA
fra
m
ewo
r
k.
We
hav
e
ap
pear
e
d
that
the
propose
d
strat
egy
connecte
d
to
MC
-
CDMA
fra
m
ewo
r
ks
acc
om
plishes
the
requir
e
d
ideal
vital
it
y
c
onveyance
du
e
to
the
execu
ti
on
of
this
tra
nsm
issi
on
fr
am
ewor
k
an
d
the
arr
a
ng
em
ent
util
iz
ed.
Re
creati
on
co
m
es
about
a
nd
com
par
iso
n
w
it
h
an
d
with
out
W
at
er
F
il
li
ng
cal
culat
ion
s
a
ff
irm
the
via
bili
ty
of
the
MC
-
CDM
A
cha
nn
el
stra
te
gy.
Th
e
W
at
er
F
il
li
ng
arr
a
ngem
ent
is
on
e
of
a
ki
nd
si
nce
the
m
axi
m
iz
e
d
flo
w
functi
on
is
con
cave
.
Subse
quently
,
there'
s
a
sing
le
ideal
vital
i
ty
con
ve
ya
nce
and
a
com
par
ing
stream
distrib
ution.
A
fter
al
l
these
com
par
isons,
it
has
these
we
re
po
inte
d
at
op
t
i
m
izing
MC
-
C
DMA
f
ram
ewo
r
ks
to
m
ov
e
fo
r
ward
their
exec
ution.
An
ideal
TE
Q
m
us
t
decr
ease
the
est
i
m
at
e
of
the
cha
nnel
to
that
of
the
pr
e
fi
x
and
dis
po
se
of
inter
sym
bol
ob
str
uctio
ns
.
So
,
t
he
pr
oblem
of
powe
r
al
locat
ion
ca
n
be
so
l
ved
by
the
op
ti
m
iz
ation
te
chn
i
qu
e
W
at
e
r F
il
li
ng
.
REFERE
NCE
S
[1]
F.
Bouasria
and
M.
Chet
ioui
,
“
W
at
te
r
-
Fil
li
ng
f
or
C
y
clic
Prefe
xed
Multi
-
c
arr
ier
Tra
nsm
ission
over
an
ADS
L
Channe
l
using
a
TE
Q
Equa
li
z
er
,
”
Inte
rnational
Journal
of
app
li
ed
Engi
n
ee
ring
Re
search
,
Vol
10,
N°
10,
pp.
26339
-
26345
,
2
015
.
[2]
Suresh,
Maha
de
van,
“
Pow
er
Co
ntrol
for
Mc
-
Cd
m
aSy
st
em
with
Predic
t
ed
Csi
U
sing
W
at
er
Fil
ling
Algorit
hm
,
”
IOSR
Journal
of
El
e
ct
roni
cs
and
Comm
unic
ati
on
Engi
ne
ering
(I
OS
R
-
JECE)
,
e
-
I
SS
N:
2278
-
2834,
p
-
ISS
N:2278
-
8735.
Volum
e
9,
Iss
ue
5,
Ver
.
1,
PP
24
-
30.
ww
w.i
osrjourna
ls.or
g
,
Sep
-
Oct.
20
14
.
[3]
R.
K.
Marti
n,
C.
R.
Johns
on,
M.
Ding
and
B.
L.
“
Eva
ns,
Infini
t
e
Le
ngth
Result
s
for
ch
anne
l
Shorteni
n
g
Equa
lizer
s,
”
V
I
EE
E
S
ignal
Proce
ss
ing
Workshop
on
Signal
Proce
ss
ing
Ad
v
ance
in
Wireles
s
Comm
unic
ati
o
n
(SP
A
W
C
2003)
June
15
-
18
,
200
3
.
[4]
J.
A.
C.
ingha
m
,
“
Multi
-
ca
rr
ier
Modulat
ion
for
Data
Tra
nsm
ission:
An
Ide
a
W
hose
Ti
m
e
H
as
Com
e,
”
IEEE
Comm
un.
Mag
,
vol.
28
,
No
.
5
,
p
p.
5
-
14
,
1990
.
[5]
N.
Yee
,
J
-
P.
Li
n
nar
tz
and
G.
Fet
twei
s,
“
Multi
-
C
a
rrie
r
CDM
A
in
I
ndoor
W
ire
le
ss
Radi
o
Networks
,
”
Proc.
o
f
IE
EE
PIM
RC
93
,
pp
.
1
09
-
113,
Yokoha
m
a,
Japa
n
,
1993
.
[6]
P.
J.
W
.
Mels
a,
R.
C.
Youn
ce
,
and
C
.
E.
Rohrs,
“
Im
puls
e
Response
Sh
orte
ning
for
D
i
scre
te
Mult
it
on
e
Tra
nsce
ive
rs,
”
I
EE
E
Tr
ans.
on
Comm
.
,
vol
.
44
,
pp.
1662
-
16
72
,
1996
.
[7]
G.
Ars
la
n,
B.
L
.
Eva
ns,
and
S.
Kiae
i,
“
Equ
al
i
z
at
ion
for
Discre
te
Multi
tone
Re
ce
iv
ers
To
Maximize
Bit
Ra
te,
”
IEE
E
Tr
ans.
on
Signal
Proce
ss
in
g
,
vol
.
49
,
no
.
12
,
pp
.
3123
-
3135
,
2001
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
5
,
Oct
ober
201
9
:
3
6
9
5
-
3
7
0
0
3700
[8]
B.
Farha
ng
,
Bor
ouje
n
y
and
M.
Ding,
“
Design
Methods
for
T
i
m
e
-
Dom
ai
n
Equa
lizer
s
in
DM
T
Tra
nsce
ive
rs,
”
I
EE
E
Tr
ans on
Comm
.
,
vol
.
49
,
no
.
3
,
p
p.
554
-
562
,
200
1
.
[9]
R.
K.
Marti
n,
J.
Bal
akr
ishnan,
W
.
A.
Setha
r
es,
an
d
C.
R.
Johns
on,
Jr.,
“
Bli
nd,
Ada
pti
ve
T
EQ
Chan
nel
Shorteni
ng
for
Multi
ca
rri
er S
y
stems
,
”
I
EEE
Signal
Proce
ss
in
g
Letters
,
vol
.
9
,
no.
11
,
pp
.
341
-
343
,
2002
.
[10]
J.
W
u,
G.
Ars
l
a
n,
and
B.
L.
E
vans,
“
Eff
i
ci
en
t
Matri
x
Multi
p
l
ic
a
ti
on
Me
thods
to
Im
ple
m
ent
a
Nea
r
Optimum
Channe
l
Short
e
ning
Method
fo
r
Discre
t
e
Multi
tone
Tr
ans
-
ceiv
ers,
”
in
Proc.
I
EE
E
Asil
omar
Conf.
on
S
ignals
,
S
yste
ms
,
and
Co
mputers
,
Pacific
Grove,
CA,
vol
.
1,
pp
.
152
-
157
,
2000
.
[11]
P.
J.
W
.
Mels
a,
R.
C.
Youn
ce
,
and
C
.
E.
Rohrs,
“
Im
puls
e
Response
Sh
orte
ning
for
Di
scre
te
Mult
it
on
e
Tra
nsce
ive
rs,
”
I
EE
E
Tr
ans.
on
Com
.
,
vol
.
44
,
p
p.
1662
-
1672
,
1
996
.
[12]
Sam
ir
Abd
El
gh
afa
r,
S
alahe
ldi
n
M.
Diab,
Bassioun
y
M.
Sallam
,
Moawad
I.
Dessouk
y
,
E
l
-
Sa
y
ed
M.
El
-
Raba
i
e
,
Fathi
E.
Abd
El
-
Sam
ie
,
“
Mult
i
-
ca
rr
ie
r
Equ
al
i
z
at
ion
b
y
Restor
at
ion
of
Redun
danc
Y
(MERRY)
for
Adapti
ve
Channe
l
Shorte
ning
in
Multi
-
ca
rri
er
S
y
stems
,
”
Inte
rnational
Journal
of
Networ
k
Sec
urity
&
I
ts
Appl
ic
ations
(
IJNSA
)
,
Vol.
5,
No.6,
Novem
ber
2013.
[13]
G.
Ars
la
n,
M.
Ding,
B.
Lu
,
Z.
S
hen,
and
B.
L
.
E
vans,
“
TE
Q
design
tool
box,
The
Univer
sit
y
of
Te
xas
at
Aus
ti
n
,
”
ava
i
la
bl
e at
h
tt
p:
//
ww
w.e
ce.utexa
s.e
du/~be
v
ans/pr
oje
c
ts/a
dsl/
dm
tte
q/dmtte
q
.
html [
ac
c
essed
2009]
,
2009
.
[14]
L.
Yang
,
C
.
G.
Kang,
“
Desing
of
Novel
Ti
m
e
-
Dom
ai
n
Equa
l
izer
(T
EQ)
for
O
FD
M
Sy
stem,
”
IEI
CE
TR
ANS;
Comm
un
,
Vol.
E
89
-
B,
N°.10
,
20
06.
[15]
Olivi
er
.
Riou
l, C
orps Finis,
Versi
on
2.
à
(199
6
-
2
006),
p
27
-
34
,
2
006
.
[16]
R.
K.
Marti
n,
M.
Ding,
B.
L.
Eva
ns
and
C.
R.
Johns
on,
“
Eff
icient
Chann
el
Shorteni
ng
E
qual
i
ze
r
Design,
”
EURA
SIP
Journ
al
on
Appl
i
ed
S
i
gnal
Proc
essing
,
13,
1279
-
1290
,
20
03
.
[17]
Fati
m
a
Bouasri
a,
Ali
Dj
ebba
r
i,
Moham
m
ed
Chet
ioui,
“
A
Bl
ind
Cha
nn
e
l
Sh
ort
enin
g
for
Mult
iuser
,
Mult
ic
arr
ie
r
CDMA
Syst
e
m
ov
er
Mult
ipath
Fa
ding
Chan
nel
,
”
TEL
KOMNIKA
(
Tele
communic
at
ion
,
Computing,
El
e
c
tronic
s and
Con
t
rol)
,
Vol.
17,
No.
4
,
2019
.
Evaluation Warning : The document was created with Spire.PDF for Python.