T
E
L
KO
M
N
I
KA
T
e
lec
om
m
u
n
icat
ion
,
Com
p
u
t
i
n
g,
E
lec
t
r
on
ics
an
d
Cont
r
ol
Vol.
18
,
No.
3
,
J
une
2020
,
pp
.
1158
~
1168
I
S
S
N:
1693
-
6930,
a
c
c
r
e
dit
e
d
F
ir
s
t
G
r
a
de
b
y
Ke
me
nr
is
tekdikti
,
De
c
r
e
e
No:
21/E
/KP
T
/2018
DO
I
:
10.
12928/
T
E
L
KO
M
NI
KA
.
v18i3.
14172
1158
Jou
r
n
al
h
omepage
:
ht
tp:
//
jour
nal.
uad
.
ac
.
id/
index
.
php/T
E
L
K
OM
N
I
K
A
Pe
r
f
o
r
m
a
n
c
e
an
al
ysi
s of
i
m
ag
e
t
r
a
n
sm
is
si
on
w
it
h
var
io
u
s c
h
an
n
e
l
c
o
n
d
i
t
io
n
s/
m
od
u
la
t
io
n
t
e
c
h
n
iq
u
e
s
M
ar
wa
Jalee
l
M
oh
s
in
1
,
Was
an
Kad
h
im
S
aad
2
,
B
as
h
ar
J.
Ham
z
a
3
,
Wah
e
b
A
.
Jab
b
ar
4
1,
2,
3
E
n
g
i
n
eeri
n
g
T
ec
h
n
i
cal
C
o
l
l
eg
e
-
N
aj
af
,
Al
-
Fu
ra
t
Al
-
A
w
s
a
t
T
ech
n
i
ca
l
U
n
i
v
ers
i
t
y
,
Iraq
4
Facu
l
t
y
o
f
E
l
ec
t
ri
ca
l
&
E
l
ec
t
ro
n
i
c
E
n
g
i
n
eeri
n
g
T
ec
h
n
o
l
o
g
y
,
U
n
i
v
er
s
i
t
i
Mal
a
y
s
i
a
Pah
a
n
g
,
Mal
a
y
s
i
a
4
IBM
Cen
t
re
o
f
E
x
cel
l
en
ce,
U
n
i
v
ers
i
t
y
Mal
a
y
s
i
a
Pah
a
n
g
,
Mal
ay
s
i
a
Ar
t
icle
I
n
f
o
AB
S
T
RA
CT
A
r
ti
c
le
h
is
tor
y
:
R
e
c
e
ived
S
e
p
21,
2019
R
e
vis
e
d
J
a
n
18
,
2020
A
c
c
e
pted
F
e
b
7
,
2020
T
h
i
s
p
ap
er
i
n
v
e
s
t
i
g
a
t
es
t
h
e
i
m
p
act
o
f
d
i
ffere
n
t
m
o
d
u
l
at
i
o
n
t
ech
n
i
q
u
e
s
f
o
r
d
i
g
i
t
al
co
mm
u
n
i
cat
i
o
n
s
y
s
t
ems
t
h
at
em
p
l
o
y
q
u
a
d
rat
u
r
e
p
h
as
e
s
h
i
ft
k
e
y
i
n
g
(Q
PSK
)
a
n
d
q
u
a
d
rat
u
re
amp
l
i
t
u
d
e
mo
d
u
l
at
i
o
n
(1
6
-
Q
A
M
an
d
6
4
-
Q
A
M)
t
o
t
ran
s
mi
t
i
mag
e
s
o
v
er
A
W
G
N
an
d
Ra
y
l
e
i
g
h
fad
i
n
g
ch
a
n
n
e
l
s
fo
r
t
h
e
ce
l
l
u
l
ar
mo
b
i
l
e
n
e
t
w
o
rk
s
.
In
t
h
e
fu
rt
h
er
s
t
ep
s
,
w
i
e
n
er
an
d
me
d
i
a
n
fi
l
t
ers
h
as
b
ee
n
ad
o
p
t
e
d
t
o
t
h
e
s
i
m
u
l
a
t
i
o
n
are
u
s
ed
a
t
t
h
e
recei
v
er
s
i
d
e
t
o
remo
v
e
t
h
e
i
m
p
u
l
s
i
v
e
n
o
i
s
e
p
re
s
en
t
i
n
t
h
e
recei
v
e
d
i
ma
g
e.
T
h
i
s
w
o
rk
i
s
p
er
fo
rmed
t
o
ev
al
u
at
e
t
h
e
t
ra
n
s
m
i
s
s
i
o
n
o
f
t
w
o
d
i
men
s
i
o
n
a
l
(2
D
)
g
ray
-
s
cal
e
a
n
d
co
l
o
r
-
s
cal
e
(RG
B)
i
mag
e
s
w
i
t
h
d
i
fferen
t
v
al
u
es
fr
o
m
s
i
g
n
a
l
t
o
n
o
i
s
e
rat
i
o
s
(S
N
R),
s
u
ch
a
s
;
(5
,
1
0
an
d
1
5
)
d
B
o
v
er
d
i
fferen
t
ch
an
n
el
s
.
T
h
e
co
rrect
co
n
c
l
u
s
i
o
n
s
are
mad
e
b
y
co
m
p
ar
i
n
g
man
y
o
f
t
h
e
o
b
s
er
v
e
d
Ma
t
l
a
b
s
i
mu
l
at
i
o
n
res
u
l
t
s
.
T
h
i
s
w
as
carri
ed
o
u
t
t
h
r
o
u
g
h
t
h
e
r
es
u
l
t
s
t
h
a
t
meas
u
re
t
h
e
q
u
a
l
i
t
y
o
f
rec
ei
v
ed
i
mag
e,
w
h
i
ch
i
s
an
al
y
zes
i
n
t
erm
s
o
f
SN
Ri
ma
g
e
p
ea
k
s
i
g
n
a
l
t
o
n
o
i
s
e
rat
i
o
(PSN
R)
an
d
mean
s
q
u
are
erro
r
(MSE
).
K
e
y
w
o
r
d
s
:
AW
GN
Gr
a
y
-
s
c
a
l
e
a
nd
c
olor
-
s
c
a
le
(
R
GB
)
im
a
ge
s
M
e
a
n
s
qua
r
e
e
r
r
or
(
M
S
E
)
P
e
a
k
s
ignal
to
nois
e
r
a
ti
o
(
P
S
NR
)
R
a
yleigh
f
a
ding
c
ha
nne
l
Th
i
s
i
s
a
n
o
p
en
a
c
ces
s
a
r
t
i
c
l
e
u
n
d
e
r
t
h
e
CC
B
Y
-
SA
l
i
ce
n
s
e
.
C
or
r
e
s
pon
din
g
A
u
th
or
:
W
a
he
b
A.
J
a
bba
r
,
F
a
c
ult
y
of
E
lec
tr
ica
l
&
E
lec
tr
onic
E
ng
inee
r
ing
T
e
c
hnology,
Unive
r
s
it
i
M
a
lays
ia
P
a
ha
ng
(
UM
P
)
,
26600,
P
e
ka
n,
P
a
ha
ng,
M
a
lays
ia
.
E
mail:
1
mar
wa
.
jale
e
l@atu.
e
du.
iq
,
2
wa
s
-
s
a
a
d@a
tu.
e
du.
iq,
3
c
oj
.
ba
s
h@a
tu.
e
du.
iq
,
4
wa
he
b@ie
e
e
.
or
g
1.
I
NT
RODU
C
T
I
ON
W
ir
e
les
s
c
ha
nne
l
s
a
r
e
e
xtr
e
mely
r
a
ndom
a
nd
ti
me
-
va
r
ying,
whic
h
pr
opa
ga
tes
a
c
r
os
s
mul
ti
ple
pa
ths
,
whe
r
e
it
is
unli
ke
f
ixed
a
nd
p
r
e
dicta
ble
wir
e
c
ha
nne
ls
.
How
e
ve
r
,
the
pr
oli
f
e
r
a
ti
on
of
mul
ti
ple
pa
ths
in
the
wir
e
les
s
s
ys
tem
is
a
c
h
a
ll
e
nge
[1
-
3]
.
T
he
r
e
s
ult
ing
phe
nomenon
via
the
a
r
r
ival
of
mul
ti
ple
c
opies
of
the
s
ignal
s
e
nt
to
the
r
e
c
e
iver
,
c
a
us
e
s
the
in
ter
s
ymbol
int
e
r
f
e
r
e
nc
e
(
I
S
I
)
to
be
dis
tor
ted,
thi
s
pr
oduc
e
d
highl
y
dis
tor
ted
the
s
ignal
s
e
nt
a
t
the
r
e
c
e
iver
.
I
n
the
s
a
m
e
ti
me,
thi
s
r
e
s
ult
s
in
ti
me
a
tt
e
nua
ti
on
,
dis
pe
r
s
ion
a
nd
pha
s
e
of
f
s
e
t
in
the
r
e
c
e
ivi
ng
s
ignal,
whic
h
is
known
a
s
f
a
ding
[4
-
6]
.
I
n
the
digi
tal
c
omm
unica
ti
on
s
ys
tem,
modul
a
ti
on
c
a
n
be
de
f
ined
a
s
the
pr
oc
e
s
s
of
s
e
nding
a
s
ignal
mes
s
a
ge
,
whe
r
e
the
inf
or
mation
a
r
e
s
hows
a
s
the
bit
s
a
nd
the
input
im
a
ge
is
uti
li
z
e
d
in
the
d
if
f
e
r
e
nt
mo
dulation
tec
hniques
[7
-
9]
.
As
a
r
e
s
ult
,
us
ing
the
64
-
QA
M
modul
a
ti
on
tec
hnique
wil
l
i
mpr
ove
the
de
gr
a
de
d
i
mage
,
thi
s
is
ne
c
e
s
s
a
r
y
f
or
im
a
ge
tr
a
ns
f
e
r
a
nd
pr
ov
ides
the
be
s
t
r
e
s
ult
s
c
ompar
e
d
with
dif
f
e
r
e
nt
modul
a
ti
on
tec
hniques
,
li
ke
QPS
K
a
nd
16
-
QA
M
be
c
a
us
e
it
ha
s
higher
d
a
ta
r
a
tes
.
T
hus
,
the
pe
r
f
or
manc
e
of
the
s
ys
tem
w
ould
be
a
c
c
e
ptable
e
ve
n
with
a
c
e
r
tain
leve
l
of
nois
e
f
r
om
the
c
r
it
ica
l
c
ha
nne
l.
I
n
a
nother
mea
ning,
the
pe
r
f
o
r
manc
e
of
the
s
ys
tem
c
a
n
not
be
tol
e
r
a
ble
whe
n
the
nois
e
lev
e
l
is
higher
than
thi
s
c
r
it
ica
l
leve
l
[
10
-
12]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
P
e
r
for
manc
e
analys
is
of
image
tr
ans
mis
s
ion
w
it
h
v
ar
ious
c
hanne
l
…
(
M
ar
w
a
J
alee
l
M
ohs
in
)
1159
F
a
ding
oc
c
ur
s
be
c
a
us
e
of
ove
r
lapping
be
twe
e
n
tw
o
or
mo
r
e
f
r
om
the
s
ignals
s
e
nt
to
the
r
e
c
e
iver
a
t
s
li
ghtl
y
dif
f
e
r
e
nt
ti
mes
.
How
e
ve
r
,
thes
e
wa
ve
s
a
r
e
mul
ti
pa
th
wa
ve
s
that
unit
e
in
the
r
e
c
e
ivi
ng
a
ntenna
a
nd
give
the
r
e
s
ult
of
a
s
ignal
that
c
a
n
va
r
y
gr
e
a
tl
y
in
a
mpl
it
ude
a
nd
p
ha
s
e
,
a
c
c
or
ding
to
the
t
r
a
ns
mi
tt
e
d
s
ignal
width,
the
r
e
lative
pr
opa
ga
ti
on
ti
me
of
the
wa
ve
s
a
nd
the
dis
tr
ibut
ion
o
f
int
e
ns
it
y
[
13
-
15]
.
Une
xpe
c
ted
de
lays
c
ha
nge
in
r
e
ga
r
de
d
wi
th
va
r
ious
s
ignal
pa
ths
in
a
mul
t
ipath
f
a
ding
c
ha
nne
l,
a
s
we
ll
a
s
c
a
n
be
s
tatis
ti
c
a
ll
y
dis
ti
n
guis
he
d
only.
How
e
ve
r
,
the
c
e
ntr
a
l
li
mi
t
theor
e
m
c
a
n
be
a
p
pli
e
d
a
s
a
r
a
ndom
pr
oc
e
s
s
of
the
c
ompl
e
x
-
va
lued
Ga
us
s
ia
n
to
model
the
ti
me
-
va
r
iant
im
puls
e
r
e
s
pons
e
of
the
c
ha
nne
l
whe
n
ther
e
a
r
e
a
lar
ge
number
o
f
pa
ths
.
I
f
th
e
im
puls
e
r
e
s
pons
e
is
de
s
igned
a
s
a
Ga
us
s
ian
pr
oc
e
s
s
with
z
e
r
o
c
ompl
e
x
va
lue,
it
is
s
a
id
that
the
c
ha
nne
l
is
R
a
yleigh
f
a
ding
c
ha
nne
l
[
16,
17
]
.
How
e
ve
r
,
R
a
yleigh’
s
models
a
s
s
um
e
that
the
s
ignal
magnitude
that
pa
s
s
e
d
thr
ough
thi
s
c
omm
unica
ti
on
c
ha
nne
l
will
c
ha
nge
r
a
ndoml
y
or
f
a
de
s
ba
s
e
d
on
R
a
yleigh’
s
di
s
tr
ibut
ion.
R
a
yleigh
f
a
ding
is
r
e
ga
r
de
d
a
s
a
r
e
a
s
ona
ble
model
f
or
the
pr
opa
g
a
ti
on
of
tr
opos
phe
r
ic
a
nd
ionos
phe
r
ic
s
ignals
a
nd
hig
hly
-
buil
t
ur
ba
n
e
nvir
onments
e
f
f
e
c
t
on
the
r
a
dio
s
ignals
wh
e
n
ther
e
a
r
e
a
lot
of
objec
ts
in
the
e
nvir
on
ment
th
a
t
s
pr
e
a
d
the
r
a
dio
s
ignal
be
f
o
r
e
it
r
e
a
c
he
s
to
the
r
e
c
e
iver
.
I
f
t
he
r
e
is
too
much
s
c
a
tt
e
r
ing
,
the
c
ha
nne
l
puls
e
r
e
s
pons
e
will
be
we
ll
de
s
igned
a
s
a
Ga
us
s
ian
pr
oc
e
s
s
a
c
c
or
ding
to
the
c
e
ntr
a
l
l
im
it
theor
e
m
r
e
ga
r
dles
s
of
the
i
ndiv
idual
c
omponents
dis
tr
ibut
ion,
whe
r
e
ther
e
a
r
e
z
e
r
o
mea
n
a
nd
pha
s
e
e
qu
a
ll
y
dis
tr
ibut
e
d
be
twe
e
n
0
a
nd
2
r
a
dians
whe
n
ther
e
is
no
pr
e
domi
na
nt
c
omponent
to
the
s
c
a
tt
e
r
.
T
hus
,
th
e
r
e
s
pons
e
e
nve
lope
of
the
c
ha
nne
l
will
be
dis
tr
ibut
e
d
to
R
a
yleigh
[
18]
.
I
mage
f
il
ter
ing
do
not
us
e
to
im
p
r
ove
the
im
a
ge
qu
a
li
ty,
but
it
is
a
ls
o
uti
li
z
e
d
a
s
a
pr
e
-
pr
oc
e
s
s
ing
s
t
a
ge
in
man
y
a
ppli
c
a
ti
ons
,
whic
h
include
s
;
im
a
ge
c
ompr
e
s
s
ion,
pa
tt
e
r
n
r
e
c
ognit
ion,
im
a
ge
e
nc
r
ypti
on,
a
nd
tar
ge
t
tr
a
c
king.
How
e
ve
r
,
no
is
e
r
e
duc
ti
on
is
uti
li
z
e
d
t
o
r
e
s
tor
e
the
idea
l
im
a
ge
f
r
om
the
de
gr
a
de
d
ve
r
s
ion
of
th
e
im
a
ge
.
I
mage
f
il
ter
ing
a
lgor
it
hms
a
r
e
us
e
d
to
e
li
mi
na
te
t
he
va
r
ious
nois
e
types
that
e
xis
t
e
it
he
r
in
the
im
a
ge
while
be
ing
c
a
ptur
e
d
or
inj
e
c
ted
int
o
the
im
a
ge
thr
ough
t
he
tr
a
ns
mi
s
s
ion.
T
hus
,
M
e
dian
F
il
te
r
is
pe
r
f
or
med
thr
ough
taking
the
ma
gnit
ude
o
f
e
a
c
h
ve
c
tor
s
wi
thi
n
the
mas
k
a
nd
s
or
ted
ba
s
e
d
on
the
magnitude
s
,
then,
the
medium
-
magnitude
pixel
is
uti
li
z
e
d
to
r
e
plac
e
the
s
tudi
e
d
pixels
.
W
hil
e
,
the
W
iene
r
F
il
ter
pe
r
f
or
ms
a
n
opti
mal
tr
a
de
-
of
f
be
twe
e
n
the
nois
e
s
moot
hing
a
nd
the
inver
s
e
f
il
t
e
r
ing
be
c
a
us
e
it
r
e
f
lec
ts
c
a
mouf
lage
a
nd
e
li
mi
na
tes
the
a
ddit
ive
nois
e
s
im
ult
a
ne
ous
ly
[
19
-
21]
.
T
he
a
uthor
s
dis
c
us
s
in
[
18]
the
e
f
f
e
c
t
of
va
r
ious
e
qua
li
z
a
ti
on
tec
hniques
that
us
e
M
P
S
K
a
nd
M
QA
M
modul
a
ti
on
tec
hnology
f
o
r
digi
tal
c
omm
unica
ti
o
n
s
ys
tems
to
tr
a
ns
f
e
r
the
inf
o
r
mation
of
c
olor
i
mage
on
the
R
a
yle
igh
f
a
ding
c
ha
nne
l.
T
he
a
na
lys
is
of
thi
s
pe
r
f
or
manc
e
ha
s
be
e
n
made
in
ter
ms
of
bit
e
r
r
o
r
r
a
te
thr
ough
us
e
M
a
tl
a
b
s
im
ulation.
T
he
s
olut
ion
f
or
thi
s
pr
ob
lem
is
to
incr
e
a
s
e
the
S
NR
va
lue.
T
hus
,
the
va
r
ious
p
r
ope
r
ti
e
s
va
lues
,
s
uc
h
a
s
;
a
ve
r
a
ge
pa
th
ga
in,
pa
th
d
e
lay
,
dop
pler
s
pe
c
tr
um
a
nd
maximum
doppler
s
hif
ts
pa
r
a
m
e
ter
s
a
r
e
s
e
lec
ted
a
c
c
ur
a
tely
in
the
s
im
ulate
th
e
R
a
yleigh
f
a
ding
c
ha
nne
l
to
tur
ns
out
a
r
e
a
li
s
ti
c
f
a
de
c
ha
nne
l.
How
e
ve
r
,
the
s
ize
of
the
s
tep
a
nd
the
f
or
ge
tt
ing
f
a
c
tor
,
e
tc.
is
a
ls
o
c
hos
e
n
f
or
the
va
r
ious
a
da
pti
ve
a
lgor
it
hms
ut
il
ize
d
by
the
li
ne
a
r
e
qua
li
z
e
r
(
L
E
)
a
nd
de
c
is
ion
f
e
e
dba
c
k
e
q
ua
li
z
e
r
(
DF
E
)
c
o
r
r
e
c
tl
y
.
M
or
e
ove
r
,
a
c
omp
r
e
s
s
ion
wa
s
us
e
d
to
tr
a
ns
f
e
r
the
c
olor
im
a
ge
that
pr
ovides
82%
les
s
da
ta
c
ompar
e
with
the
unc
ompr
e
s
s
e
d
J
P
E
G
im
a
ge
.
E
a
c
h
a
na
lys
is
of
f
e
r
s
the
wor
s
t
e
qua
li
z
e
r
f
or
the
R
a
ylei
gh
f
a
ding
c
ha
nne
l
li
ne
a
r
e
qua
li
z
e
r
with
the
L
M
S
a
da
pti
ve
a
lgo
r
it
hm.
I
n
a
ddi
ti
on,
to
r
e
duc
e
the
s
ignal
dis
pe
r
s
ion
a
nd
I
S
I
,
the
maximum
li
ke
li
hood
s
e
que
nc
e
e
s
ti
mate
(
M
L
S
E
)
e
qua
li
z
e
r
pr
ovides
the
be
tt
e
r
e
qua
li
z
a
ti
on
.
How
e
ve
r
,
the
r
e
s
ult
s
e
xplaine
d
that
bo
th
o
f
DFE
a
nd
L
E
a
da
pti
ve
a
lgor
it
hm
e
qua
li
z
a
ti
on
gives
be
s
t
r
e
s
ul
t
t
ha
n
l
ms
.
W
hil
e
,
DFE
o
f
f
e
r
s
the
be
s
t
pe
r
f
or
manc
e
c
ompar
e
with
the
li
ne
a
r
e
qua
li
z
e
r
,
other
than
that
it
gives
t
he
poor
pe
r
f
or
manc
e
wi
th
the
lowe
r
S
NR
.
B
ut
th
e
M
L
S
E
e
qua
li
z
e
r
gives
the
be
tt
e
r
pe
r
f
or
manc
e
a
mong
a
ll
the
e
qua
li
z
e
r
s
us
e
d
in
thi
s
wor
k.
I
n
ge
ne
r
a
l,
a
ll
th
e
r
e
s
ult
s
e
xplaine
d
that
uti
l
ize
of
the
4QA
M
modul
a
ti
on
method
a
nd
the
M
L
S
E
e
qua
li
z
e
r
g
ive
a
be
s
t
r
e
s
ult
f
or
tr
a
ns
f
e
r
r
ing
the
c
olor
J
P
G
c
olor
i
mage
s
.
In
[
11]
the
a
uthor
s
de
mons
tr
a
ted
that
the
t
r
a
ns
mi
s
s
ion
of
im
a
ge
via
digi
tal
wi
r
e
les
s
c
omm
unica
ti
ons
by
us
e
d
dif
f
e
r
e
nt
digi
tal
modul
a
ti
on
tec
hniques
,
li
k
e
;
QPS
K,
16
-
QA
M
a
nd
64
-
QA
M
a
r
e
im
pl
e
mente
d
thr
ough
uti
li
z
e
d
the
AW
GN
c
ha
nne
l
to
de
s
ign
a
nd
im
ple
ment
the
wir
e
les
s
digi
tal
c
omm
unica
ti
on
s
ys
tem
a
s
we
ll
a
s
know
the
be
tt
e
r
modul
a
ti
on
tec
hniq
ue
s
uit
a
ble
f
or
tr
a
ns
mi
tt
ing
the
i
mage
.
How
e
ve
r
,
the
pe
r
f
o
r
manc
e
of
thi
s
s
ys
tem
will
be
a
c
c
e
ptable
li
ke
many
other
digi
tal
wir
e
les
s
c
omm
unica
ti
ons
s
ys
tems
to
r
e
a
c
h
a
a
ppoi
nted
nois
e
leve
l
f
r
om
the
c
r
it
ica
l
c
ha
nne
l.
I
n
a
nother
mea
ni
ng,
the
s
ys
tem
pe
r
f
o
r
manc
e
c
a
nnot
be
ve
r
y
quic
kly
whe
n
the
nois
e
leve
l
is
higher
than
th
e
c
r
it
ica
l
leve
l.
T
he
c
ur
r
e
nt
de
s
igned
s
ys
tem
in
thi
s
wor
k
ge
ne
r
a
tes
wor
s
e
im
a
ge
qua
li
ty
ins
tea
d
of
los
ing
the
im
a
ge
s
e
nt
c
ompl
e
t
e
ly
whe
n
the
c
ha
nne
l
is
in
high
nois
e
.
S
im
ulatio
n
r
e
s
ult
s
e
xpla
ined
that,
64
-
QA
M
modul
a
ti
on
tec
hniques
ha
s
higher
da
ta
r
a
tes
,
thi
s
is
ne
c
e
s
s
a
r
y
f
or
im
a
ge
tr
a
ns
f
e
r
,
of
f
e
r
be
s
t
r
e
s
ult
s
c
ompar
e
with
the
a
nother
modul
a
ti
on
t
e
c
hniques
(
i.
e
.
,
QPS
K
a
nd
16
-
QA
M
)
whe
n
the
S
NR
va
lue
is
10
dB
.
T
he
a
uthor
s
in
[
22
]
co
nc
e
ntr
a
ted
on
the
tr
a
ns
f
e
r
of
im
a
ge
s
by
us
e
d
M
-
QA
M
OFDM
s
ys
tem
f
or
the
f
a
ding
c
ha
nne
l.
T
he
Ge
ne
r
a
li
z
e
d
K
F
a
ding
c
ha
nne
l
is
e
va
luation
thr
ough
s
e
nding
s
ti
ll
im
a
ge
s
by
the
16
-
QAM
-
OFDM
modul
a
ted
s
ys
tem
with
va
r
ious
va
lues
f
r
om
f
a
ding
a
nd
s
ha
ding
pa
r
a
mete
r
s
.
T
he
im
a
ge
s
s
e
nt
thr
ough
c
ompos
it
e
f
a
ding
c
ha
nne
ls
will
b
e
e
xc
e
pti
ona
ll
y
s
e
ns
it
ive
to
the
inf
luenc
e
o
f
mul
ti
pa
th
pr
opa
ga
ti
on,
whic
h
may
r
e
s
ult
the
im
a
ge
qua
li
ty
de
gr
a
da
ti
on
in
the
r
e
c
e
iver
.
T
he
s
im
ulation
r
e
s
ult
s
s
howe
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
3
,
J
une
2020:
1159
-
1168
1160
th
a
t
M
-
QAM
-
OFDM
r
e
s
ult
e
d
a
good
t
r
a
de
of
f
be
twe
e
n
s
ymbol
-
e
r
r
or
dur
a
bil
i
ty,
thr
oughput
a
nd
r
e
duc
e
of
the
pe
a
k
-
to
-
a
ve
r
a
ge
r
a
ti
o
of
the
OFDM
s
ignal.
I
n
a
ddit
ion,
the
ge
ne
r
a
li
z
e
d
K
F
a
ding
c
ha
nne
l
is
ve
r
y
e
f
f
e
c
ti
ve
f
or
p
r
e
diction
the
f
a
ding
a
nd
s
ha
dowing
pe
r
f
or
ma
n
c
e
a
s
we
ll
a
s
s
howing
it
s
e
f
f
e
c
t
s
a
t
the
s
a
me
ti
me
.
M
or
e
ove
r
,
the
r
e
s
ult
s
e
xa
mi
ne
d
s
how
the
inf
luenc
e
of
f
a
de
a
n
d
s
ha
ding
f
r
om
whe
r
e
of
bit
e
r
r
or
r
a
te
a
nd
vis
ua
l
qua
li
ty
of
the
r
e
c
e
ived
im
a
ge
s
.
How
e
ve
r
,
be
c
a
us
e
of
the
mobi
le
r
a
dio
s
ignal
c
a
s
ing
us
ua
ll
y
c
o
ns
is
t
s
of
a
f
a
s
t
f
a
ding
c
omponent
ove
r
lay
on
a
s
lowly
f
a
ding
c
omponent,
thus
,
it
is
c
lea
r
that
f
r
om
the
r
e
s
ult
s
pr
e
s
e
nted
the
c
ha
nne
l
model
whic
h
c
ons
is
ts
of
mi
xe
d
dis
tr
ibut
ions
will
b
e
us
e
f
ul
in
de
s
ign
the
wir
e
les
s
s
ys
tems
a
nd
ne
twor
ks
.
F
ur
ther
mor
e
,
one
t
ype
of
thi
s
digi
tal
da
ta,
whic
h
c
or
r
e
s
ponds
to
to
a
two
-
dim
e
ns
ional
gr
a
y
-
s
ize
(
2D)
im
a
ge
is
uti
li
z
e
d
f
or
e
va
luation
the
ove
r
a
ll
of
f
un
c
ti
ona
li
on
a
nd
the
pe
r
f
or
manc
e
of
the
OFDM
s
ys
t
e
m
with
the
im
pa
c
t
of
the
AW
GN
c
ha
nne
l
in
[
22]
.
Va
r
ious
c
onf
igur
a
ti
ons
of
the
modul
a
ti
on
tec
hniques
,
li
ke
s
;
M
-
P
S
K
a
nd
M
-
QA
M
a
r
e
c
ons
ide
r
e
d
withi
n
the
OFDM
s
y
s
tem
f
or
e
va
luation
o
f
the
s
ys
tem
a
s
we
ll
a
s
the
n
e
c
e
s
s
a
r
y
c
or
r
e
c
t
c
onc
lus
ions
a
r
e
made
f
r
om
c
ompar
ing
many
of
no
ted
s
im
ulatio
n
r
e
s
ult
s
in
M
a
tl
a
b.
How
e
ve
r
,
t
he
OFDM
s
ys
tem
wa
s
im
pleme
nted
us
ing
va
r
ious
modul
a
ti
on
tec
hniques
to
tr
a
ns
mi
t
im
a
ge
s
via
the
AW
GN
c
ha
nne
l.
T
he
r
e
c
ove
r
e
d
i
mage
qua
li
ty
is
be
s
t
a
t
high
S
NR
va
lues
r
e
ga
r
dles
s
of
the
uti
l
ize
d
modul
a
ti
on
method
.
I
n
the
c
a
s
e
of
low-
S
NR
,
the
qua
li
ty
of
the
r
e
c
o
ve
r
e
d
im
a
ge
is
ve
r
y
low
be
c
a
us
e
o
f
h
igh
a
mount
of
nois
e
in
the
AW
GN
c
ha
nne
l.
I
n
thi
s
wo
r
k,
it
wa
s
f
ound
th
a
t
OFDM
s
ys
tem
pr
ovides
f
e
we
r
e
r
r
or
s
,
higher
r
e
c
ove
r
e
d
im
a
ge
qua
li
ty
a
t
the
r
e
c
e
iver
a
nd
les
s
B
E
R
with
16
-
QA
M
modul
ati
on
tec
hnique
c
ompar
e
d
to
the
OF
DM
s
ys
tems
a
ppli
e
d
with
other
tec
hniques
.
T
his
pa
pe
r
a
na
lyze
the
pe
r
f
o
r
manc
e
of
im
a
ge
tr
a
ns
f
e
r
a
c
r
os
s
dif
f
e
r
e
nt
c
ha
nne
ls
whic
h
dis
play
va
r
ious
e
f
f
e
c
ts
on
im
a
ge
t
r
a
ns
f
e
r
pe
r
f
or
manc
e
thr
ough
us
e
dif
f
e
r
e
nt
f
il
te
r
s
.
I
n
thi
s
wo
r
k,
ther
e
a
r
e
two
types
of
digi
tal
da
ta
whic
h
c
o
r
r
e
s
ponds
to
a
2D
g
r
a
y
-
s
c
a
le
a
nd
c
olor
-
s
c
a
le
(
R
GB
)
of
J
P
E
G
f
or
mat
i
mage
s
a
r
e
us
e
d
f
or
tr
a
ns
mi
s
s
ion
da
ta
a
nd
a
s
s
e
s
s
the
c
ompr
e
he
ns
ive
pe
r
f
or
manc
e
a
nd
the
f
unc
ti
ona
li
ty
of
the
s
ys
tem
with
t
he
e
f
f
e
c
t
of
AW
GN
a
nd
R
a
yleigh
f
a
ding
c
ha
nne
ls
.
T
he
va
r
ious
c
ha
nne
ls
s
how
va
r
ious
im
pa
c
ts
on
the
e
xe
c
uti
on
of
the
tr
a
ns
f
e
r
of
im
a
ge
s
uti
li
z
ing
dif
f
e
r
e
nt
modul
a
ti
o
n
tec
hniques
.
How
e
ve
r
,
wie
ne
r
a
nd
media
n
f
il
ter
s
a
r
e
us
e
d
a
t
the
r
e
c
e
iver
s
ide
to
r
e
move
the
im
puls
ive
no
is
e
pr
e
s
e
n
t
in
the
r
e
c
e
ived
im
a
ge
.
T
he
r
e
ma
inder
of
th
e
pa
pe
r
is
or
de
r
ly
a
s
f
oll
ows
:
the
p
r
opos
e
d
im
a
ge
t
r
a
ns
mi
s
s
io
n
s
ys
tem
model
is
dis
c
us
s
e
d
in
s
e
c
ti
on
2.
S
e
c
ti
on
3
pr
e
s
e
nts
a
nd
dis
c
us
s
e
s
the
pe
r
f
or
manc
e
e
va
luation
of
the
pr
opos
e
d
s
ys
tem,
whe
r
e
the
s
im
ulation
r
e
s
u
lt
s
a
nd
di
s
c
us
s
ion
f
or
the
pr
opos
e
d
s
ys
tem
is
highl
ight
e
d.
F
inally
,
s
e
c
ti
on
4
is
d
r
a
wn
the
c
onc
ludi
ng
r
e
mar
ks
of
thi
s
pa
p
e
r
.
2.
P
ROP
OS
E
D
T
RA
NSMI
S
I
ON
S
YST
E
M
M
ODE
L
F
igu
r
e
1
e
xplain
the
im
pleme
n
ti
on
of
the
s
ys
tem
model,
whic
h
de
s
c
r
ibes
the
s
e
qu
e
nc
e
of
e
ve
nts
u
ti
li
z
e
in
the
c
ur
r
e
nt
wor
k
to
s
e
nd
Gr
a
y
-
s
c
a
le
a
nd
R
GB
i
mage
s
of
s
ize
512x512
ove
r
AW
GN
a
nd
R
a
yleigh
F
a
ding
c
ha
nne
ls
,
r
e
s
pe
c
ti
ve
ly.
T
he
c
omm
unica
ti
on
c
ha
nne
l
is
de
s
ig
ne
d
by
a
dding
the
Ga
u
s
s
ian
white
nois
e
e
f
f
e
c
t
a
nd
the
c
a
r
r
ier
in
f
luenc
e
.
T
he
im
a
ge
s
a
r
e
r
e
s
ha
pe
d
ba
s
e
d
on
the
modul
a
ti
on
tec
hnique
us
e
d,
a
nd
then
s
e
nt
thr
ough
thes
e
c
ha
nne
ls
.
T
he
tes
t
im
a
ge
c
hos
e
n
f
or
e
va
luation
of
the
pr
opos
e
d
im
a
ge
t
r
a
ns
f
e
r
s
ys
tem
a
ppe
a
r
s
in
F
igur
e
s
2
a
nd
3
.
T
his
v
ir
tual
im
a
ge
is
e
xis
ti
ng
with
J
P
G
f
or
mat
,
s
ize
512
x
512
with
262144
pixels
,
wh
e
r
e
e
a
c
h
pixel
8
-
bit
s
,
a
nd
a
c
c
e
s
s
ibl
e
thr
ough
the
na
me
"
P
e
nguin
I
mage
"
.
T
o
tr
a
ns
f
e
r
thi
s
im
a
ge
a
va
il
a
ble
in
a
two
dim
e
ns
ional
(
2D)
matr
ix
or
s
ignal,
we
ne
e
d
to
d
o
s
ome
pr
e
-
pr
oc
e
s
s
ing
to
c
onve
r
t
thi
s
2D
im
a
ge
to
a
one
dim
e
ns
ional
(
1D)
s
igna
l.
I
n
the
s
tar
t,
r
e
a
d
the
im
a
ge
that
is
obtaina
ble
f
r
om
M
a
tl
a
b
da
taba
s
e.
Af
ter
that,
we
c
ha
nge
it
to
a
double
c
oor
dinate
then
r
e
s
tr
uc
tu
r
e
the
da
ta
to
c
onve
r
t
t
he
m
f
r
o
m
the
(
512x51
2)
matix
r
e
pr
e
s
e
ntation
to
the
(
1×
262144)
ve
c
to
r
r
e
pr
e
s
e
ntat
ion
,
a
nd
then
we
c
onve
r
t
it
to
the
binar
y
that
r
e
s
ult
s
a
ve
c
tor
of
binar
y
da
ta
(
512×
512×
8=
2097152
bi
ts
)
f
or
gr
a
y
-
s
c
a
l
e
im
a
ge
a
nd
(
512×
512×
8×
3)
f
or
R
GB
im
a
g
e
,
whe
r
e
thi
s
im
a
ge
be
c
omes
now
r
e
a
dy
f
or
t
r
a
ns
mi
s
s
ion.
T
his
da
ta
is
us
e
d
a
s
a
da
ta
s
our
c
e
or
input
s
ignal
f
or
im
a
ge
t
r
a
ns
mi
s
s
ion
s
ys
tem
,
whe
r
e
thi
s
s
our
c
e
da
ta
is
m
odulate
d
thr
ough
us
e
e
it
he
r
QP
S
K,
16
-
QA
M
,
or
64
-
QA
M
tec
hniques
.
F
inally,
a
c
c
or
ding
to
the
modul
a
ti
on
method
uti
l
ize
d,
we
mus
t
be
c
onve
r
s
ion
the
ve
c
to
r
da
ta
to
a
n
a
ppr
opr
iate
f
o
r
mat.
F
or
ins
tanc
e
,
if
we
uti
li
z
e
QPS
K
modul
a
ti
on
te
c
hnique,
we
will
c
onve
r
s
ion
the
ve
c
tor
da
ta
to
the
binar
y
da
ta
(
f
our
s
ignal
e
leme
nts
,
i
.
e
.
,
00,
01
,
1
0
a
nd
11,
whic
h
a
r
e
0
,
1,
2
,
a
nd
3
,
r
e
s
pe
c
ti
ve
ly)
.
H
owe
ve
r
,
it
will
be
a
good
c
ompar
is
on
if
we
us
e
a
nother
mo
dulation
tec
hniques
a
s
th
e
tr
a
ns
mi
tt
e
r
modul
a
ti
on,
s
uc
h
a
s
;
16-
QA
M
a
nd
64
-
QA
M
.
T
hus
,
f
o
r
16
-
QA
M
modul
a
ti
on,
whic
h
c
o
r
r
e
s
ponds
to
(
s
ixt
e
e
n
s
ignaling
e
leme
nts
,
i
.
e
.
,
0000,
0001
,
to
1111
,
whic
h
a
r
e
0
,
1
,
to
15,
r
e
s
pe
c
ti
ve
ly)
,
a
nd
f
or
64
-
QA
M
modul
a
ti
on,
w
hich
is
c
o
mpatibl
e
with
(
s
ixt
y
f
our
s
ignaling
e
leme
nts
,
i.
e
.
,
000000
,
t
o
111111,
whic
h
a
r
e
0
to
63,
r
e
s
pe
c
ti
ve
ly)
.
O
n
the
r
e
c
e
iver
s
ide,
the
oppos
it
e
ope
r
a
ti
ons
a
r
e
a
c
hieve
d
to
e
f
f
e
c
ti
ve
ly
r
e
s
tor
e
the
tr
a
ns
f
e
r
e
d
or
igi
na
l
i
mage
a
f
ter
r
e
c
ove
r
ing
the
digi
tal
bit
s
.
T
he
s
ys
tem
pa
r
a
mete
r
s
ut
il
ize
in
t
his
s
im
ulation
will
be
tabula
te
in
T
a
ble
1
,
whic
h
dis
plays
the
c
ha
r
a
c
ter
is
ti
c
of
the
or
igi
na
l
im
a
ge
us
e
d
a
nd
the
c
o
r
r
e
s
ponding
s
im
ulatio
n
pa
r
a
mete
r
s
va
lues
that
a
r
e
c
ons
ider
in
the
s
im
ulation
M
a
tl
a
b.
How
e
ve
r
,
ther
e
a
r
e
two
types
of
objec
ti
ve
met
r
ic
s
that
a
r
e
wide
ly
ut
il
ize
d
in
the
e
va
luating
im
a
ge
tr
a
ns
mi
s
s
ion
tec
hniques
,
s
uc
h
a
s
:
P
e
a
k
S
ignal
to
N
ois
e
R
a
ti
o
(
P
S
NR
)
,
whic
h
is
uti
li
z
e
d
to
mea
s
ur
e
th
e
im
a
ge
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
P
e
r
for
manc
e
analys
is
of
image
tr
ans
mis
s
ion
w
it
h
v
ar
ious
c
hanne
l
…
(
M
ar
w
a
J
alee
l
M
ohs
in
)
1161
qua
li
ty,
a
ls
o,
M
e
a
n
S
qua
r
e
E
r
r
o
r
(
M
S
E
)
,
whic
h
us
e
s
to
mea
s
ur
e
the
im
a
ge
dis
tor
ti
on
a
nd
may
be
e
xp
laine
d
a
s
the
s
qua
r
e
of
e
r
r
or
be
twe
e
n
the
s
tego
(
)
j
i
P
x
R
,
a
nd
c
ove
r
(
)
j
i
P
x
T
,
im
a
ge
s
.
T
hus
,
the
r
e
c
ons
tr
uc
ted
im
a
ge
s
qua
li
ty
with
s
ize
M
×
N
in
ter
ms
of
P
S
NR
is
given
by
[
23
-
25]
:
=
M
S
E
PSNR
2
10
2
5
5
l
o
g
10
(
1)
whe
r
e
;
(
)
(
)
=
=
−
=
M
i
N
j
R
T
j
i
P
j
i
P
N
M
MS
E
x
x
1
1
2
,
,
1
(
2)
F
igur
e
1.
S
ys
tem
model
T
a
ble
1.
F
e
a
tur
e
s
of
pe
nguin
tes
t
im
a
ge
a
nd
M
a
tl
a
b
s
im
ulation
pa
r
a
mete
r
s
P
r
ope
r
ti
e
s
C
or
r
e
s
ponding Va
lu
e
s
O
r
ig
in
a
l
I
ma
ge
S
iz
e
(
512 x 512)
pi
xe
ls
T
ot
a
l
P
ix
e
ls
262144
S
iz
e
da
ta
f
or
e
a
c
h pi
xe
l
8-
bi
ts
S
iz
e
of
t
he
s
ig
na
l
e
le
me
n
ts
or
s
our
c
e
s
ig
na
l
f
or
t
he
Q
P
S
K
tr
a
ns
mi
s
s
io
n s
iz
e
262144×
8/
2 =
1048576 (
gr
a
y
-
s
c
a
le
i
ma
ge
)
262144×
8×
3/
2 =
3145728 (
R
G
B
i
ma
ge
)
S
iz
e
of
t
he
s
ig
na
l
e
le
me
nt
s
or
s
our
c
e
s
ig
na
l
f
or
t
he
16
-
Q
A
M
tr
a
ns
mi
s
s
io
n s
iz
e
262144×
8/
4 =
524288 (
gr
a
y
-
s
c
a
le
i
ma
ge
)
262144×
8×
3/
4
=
1572864 (
R
G
B
i
ma
ge
)
S
iz
e
of
t
he
s
ig
na
l
e
le
me
nt
s
or
s
our
c
e
s
ig
na
l
f
or
t
he
64
-
Q
A
M
tr
a
ns
mi
s
s
io
n s
iz
e
262144×
8/
6 =
349525 (
gr
a
y
-
s
c
a
le
i
ma
ge
)
262144×
8×
3/
6 =
1048576 (
R
G
B
i
ma
ge
)
M
odul
a
ti
on S
c
he
me
s
Q
P
S
K
, 16
-
Q
A
M
, 64
-
QAM
C
ha
nne
l
ty
pe
A
W
G
N
a
nd R
a
yl
e
ig
h f
a
di
ng
R
a
nge
of
S
N
R
i
n dB
5, 10 a
nd 15
3.
RE
S
UL
T
S
AN
D
P
E
RF
ORM
AN
CE
AN
AL
YSI
S
I
n
thi
s
wor
k
,
two
types
of
digi
tal
da
ta
whic
h
r
e
pr
e
s
e
nts
two
-
dim
e
n
s
ional
(
2D)
gr
a
y
-
s
c
a
le
a
nd
R
GB
im
a
ge
s
a
r
e
uti
l
ize
d
to
a
s
s
e
s
s
the
ove
r
a
ll
pe
r
f
o
r
manc
e
a
nd
the
f
unc
ti
ons
of
th
e
pr
opos
e
d
s
ys
tem
with
t
he
e
f
f
e
c
t
of
typi
c
a
l
AW
GN
a
nd
R
a
yleigh
f
a
ding
c
ha
nne
ls
a
n
d
va
r
ious
f
il
ter
s
,
whic
h
us
e
s
on
the
r
e
c
e
iver
s
ide
to
r
e
move
the
im
puls
ive
nois
e
in
the
r
e
c
e
ived
im
a
ge
in
M
a
tl
a
b
s
im
ulation.
How
e
ve
r
,
va
r
ious
c
onf
igu
r
a
ti
ons
of
p
r
omi
ne
nt
modul
a
ti
on
tec
hni
que
s
,
li
ke
;
QPS
K,
16
-
QA
M
a
nd
64
-
QA
M
ha
ve
be
e
n
u
s
e
d
in
thi
s
wor
k
to
e
va
luate
the
pr
opos
e
d
s
ys
tem
a
s
we
ll
a
s
make
the
c
or
r
e
c
t
c
o
nc
lus
ions
f
or
c
ompar
ing
many
r
e
s
ult
s
o
f
s
im
ulation
M
a
tl
a
b.
I
n
the
f
ir
s
t
pa
r
t
of
thi
s
s
e
c
ti
on,
we
a
r
e
going
to
c
om
pa
r
e
the
r
e
s
ult
of
the
two
types
of
im
a
ge
s
tr
a
ns
mi
s
s
ion
with
thr
e
e
dif
f
e
r
e
nt
S
NR
ove
r
AW
GN
c
ha
nne
l.
I
n
the
ne
xt
s
tep
we
c
ompar
e
the
r
e
s
ult
s
with
the
s
im
ulation
a
nd
the
theor
e
ti
c
a
l
pe
r
f
or
manc
e
of
the
im
a
ge
tr
a
ns
mi
s
s
ion
ove
r
ove
r
the
R
a
yleigh
f
a
ding
c
ha
nne
l.
Ho
we
ve
r
,
the
gr
a
ph
will
a
na
lyze
the
c
ompar
is
on
be
twe
e
n
two
im
a
ge
s
whic
h
uti
li
z
e
s
to
dis
c
us
s
the
pe
r
f
or
manc
e
.
T
he
s
ys
tem
ha
s
be
e
n
im
pleme
nted
a
s
de
s
c
r
ibed
in
s
e
c
ti
on
2,
with
us
e
QPS
K
,
16
-
QA
M
a
nd
64
-
QA
M
a
s
a
modul
a
ti
on
tec
hnique
f
or
a
na
lys
is
the
pe
r
f
or
ma
nc
e
of
i
mage
tr
a
ns
f
e
r
s
by
us
ing
va
r
ious
c
ha
nne
l
c
ondit
ions
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
3
,
J
une
2020:
1159
-
1168
1162
with
a
nd
without
two
types
f
r
om
f
il
ter
s
th
r
ough
M
a
tl
a
b
s
im
ulation.
T
he
s
e
f
il
ter
s
a
r
e
W
iene
r
a
nd
M
e
dian,
r
e
s
pe
c
ti
ve
ly,
whic
h
us
e
s
to
r
e
move
the
nois
e
f
r
om
the
im
a
ge
.
T
hus
,
we
c
a
n
s
e
e
that
the
s
im
ulatio
n
r
e
s
ult
s
a
r
e
not
the
s
a
me
a
s
the
or
igi
na
l
im
a
ge
,
but
they
a
r
e
a
lm
os
t
identica
l.
S
im
ulation
r
e
s
ult
s
a
r
e
c
onduc
ted
to
c
ompar
e
the
im
a
ge
s
r
e
c
ove
r
e
d
f
or
thr
e
e
va
r
ious
S
NR
va
lues
a
r
e
5,
10
a
nd
15
dB
,
r
e
s
pe
c
ti
ve
ly
,
whic
h
c
onf
o
r
mi
ng
to
va
r
ious
modul
a
ti
on
tec
hniques
.
T
a
bles
2-
7
a
r
e
a
go
od
tool
to
c
ome
up
the
s
ys
tem
pe
r
f
or
manc
e
c
ompar
is
on
f
o
r
M
S
E
,
P
S
NR
a
nd
S
NR
i
mage
on
dif
f
e
r
e
nt
types
of
im
a
ge
s
,
S
NR
c
ha
nne
l,
f
il
te
r
s
,
a
nd
modi
f
ica
ti
on
te
c
hniques
f
or
AW
GN
a
nd
R
a
yleigh
f
a
de
c
ha
nne
ls
.
Obviou
s
ly
f
r
om
t
he
P
S
NR
r
e
s
ult
s
,
the
s
ys
tem
wo
r
ks
we
ll
in
the
AW
GN
c
ha
nne
l
c
ompar
e
with
the
R
a
yleigh
f
a
ding
c
ha
nne
l,
a
s
we
ll
a
s
with
a
s
li
ght
incr
e
a
s
e
in
s
ignal
s
tr
e
ngth,
the
im
a
ge
will
be
r
e
c
e
ived
idea
ll
y,
whe
r
e
the
highe
r
P
S
NR
will
im
p
r
ove
the
qua
li
ty
of
r
e
c
ove
r
e
d
im
a
ge
.
T
a
ble
2.
C
ompar
is
on
M
S
E
f
or
dif
f
e
r
e
nt
im
a
ge
s
,
m
odulation
s
c
he
mes
a
nd
f
il
ter
s
(
AW
GN
c
ha
nne
l)
64-
QAM
16-
QAM
Q
P
S
K
S
N
R
c
ha
nne
l
(
dB
)
I
ma
ge
T
ype
s
M
e
di
a
n
f
il
te
r
(
dB
)
W
ie
ne
r
f
il
te
r
(
dB
)
W
it
hou
t
f
il
te
r
(
dB
)
M
e
di
a
n
f
il
te
r
(
dB
)
W
ie
ne
r
f
il
te
r
(
dB
)
W
it
hout
f
il
te
r
(
dB
)
M
e
di
a
n
f
il
te
r
(
dB
)
W
ie
ne
r
f
il
te
r
(
dB
)
W
it
hout
f
il
te
r
(
dB
)
1.2620e
+
03
1.9147e
+
03
5.360e
+
03
280.654
352.817
1.5378e
+
03
69.3990
344.469
1.011e
+
03
5
G
r
a
y
-
s
c
a
le
I
ma
ge
442.525
1.0820e
+
03
3.817e
+
03
70.6717
128.085
540.264
5
6.9569
43.7866
30
.2378
10
156.2942
570.0549
2.605e
+
03
55.6
810
63.1253
163.711
4
6.5855
40.6287
22
.6794
15
1.2280e
+
03
1.6364e
+
03
5.2678
e
+
03
310.983
290.013
1.6574e
+
03
78.5478
341.577
958.994
5
R
G
B
I
ma
ge
461.7473
1.0582e
+
03
3.7758
e
+
03
66.0657
137.857
580.783
4
4.6936
4
9
.470
20.1701
10
101.2831
558.8409
2.2483
e
+
0
3
50.7978
57.1098
102.652
3
4.5352
4
6
.8864
0.0469
15
T
a
ble
3
.
C
ompar
is
on
P
S
NR
f
or
di
f
f
e
r
e
nt
im
a
ge
s
,
modul
a
ti
on
s
c
he
mes
a
nd
f
il
ter
s
(
AW
GN
c
ha
nne
l)
64-
QAM
16-
QAM
Q
P
S
K
S
N
R
c
ha
nne
l
(
dB
)
I
ma
ge
T
ype
s
M
e
di
a
n
f
il
te
r
W
ie
ne
r
f
il
te
r
W
it
hout
f
il
te
r
M
e
di
a
n
f
il
te
r
W
ie
ne
r
f
il
te
r
W
it
hout
f
il
te
r
M
e
di
a
n
f
il
te
r
W
ie
ne
r
f
il
te
r
W
it
hout
f
il
te
r
17.1202
14.3343
10.8392
23.6491
20.2922
16.2617
29.5914
22.7170
18.0733
5
G
r
a
y
-
s
c
a
le
I
ma
ge
21.6714
17.4765
12.3136
30.3415
25.9956
20.8047
31.4138
30.2137
34.6317
10
28.3947
20.7669
13.9730
32.9143
30.0946
25.9900
35.4483
32.2276
65.2210
15
18.2389
15.7501
11.0145
23.9311
21.1050
15.9365
29.1795
22.7959
18.3126
5
R
G
B
I
ma
ge
22.4868
17.7479
12.3607
30.7825
27.7365
20.4907
32.
01
14
30.4705
35.0837
10
29.2476
20.6238
14.6123
32.511
7
30.7980
30.0172
32.
75
72
31.3460
61.4214
15
T
a
ble
4
.
C
ompar
is
on
P
S
NR
f
or
di
f
f
e
r
e
nt
im
a
ge
s
,
modul
a
ti
on
s
c
he
mes
a
nd
f
il
ter
s
(
R
a
yleigh
f
a
ding
c
ha
nne
l)
64-
QAM
16-
QAM
Q
P
S
K
S
N
R
c
ha
nne
l
(
dB
)
I
ma
ge
T
ype
s
M
e
di
a
n
f
il
te
r
W
ie
ne
r
f
il
te
r
W
it
hout
f
il
te
r
M
e
di
a
n
f
il
t
er
W
ie
ne
r
f
il
te
r
W
it
hout
f
il
te
r
M
e
di
a
n
f
il
te
r
W
ie
ne
r
f
il
te
r
W
it
hout
f
il
te
r
14.0395
13.4196
9.5892
16.9202
14.4251
10.7082
17.7593
14.6190
12.8825
5
G
r
a
y
-
s
c
a
le
I
ma
ge
16.1857
14.7692
10.5305
26.3719
22.2361
18.0990
31.3390
29.3158
24.7746
10
25.1048
16.4549
12.6452
32.4056
28.7198
23.4187
32.9241
30.4705
40.5959
15
16.4571
15.1511
10.6886
22.6888
20.4204
14.4099
27.3415
20.7003
16.1179
5
R
G
B
I
ma
ge
19.1690
16.1695
11.2729
29.2925
27.3148
19.1640
30.4376
28.9960
27.2346
10
28.6534
22.8007
14.1425
32.2226
30.7421
28.7499
32.7623
27.1929
49.9106
15
T
a
ble
5
.
C
ompar
is
on
S
NR
im
a
ge
f
o
r
dif
f
e
r
e
nt
im
a
ge
s
,
modul
a
ti
on
s
c
he
mes
a
nd
f
il
ter
s
(
AW
GN
c
ha
nne
l)
64-
QAM
16-
QAM
Q
P
S
K
S
N
R
c
ha
nne
l
(
dB
)
I
ma
ge
T
ype
s
M
e
di
a
n
f
il
te
r
(
dB
)
W
ie
ne
r
f
il
te
r
(
dB
)
W
it
hout
f
il
te
r
(
dB
)
M
e
di
a
n
f
il
te
r
(
dB
)
W
ie
ne
r
f
il
te
r
(
dB
)
W
it
hout
f
il
te
r
(
dB
)
M
e
di
a
n
f
il
te
r
(
dB
)
W
ie
ne
r
f
il
te
r
(
dB
)
W
it
hout
f
il
te
r
(
dB
)
6.6290
5.2633
3.4885
9.8935
8.3359
6.1998
12.8646
9.4445
7.1056
5
G
r
a
y
-
s
c
a
le
I
ma
ge
8.9046
6.9632
4.2257
12.9397
11.0853
8.4713
13.7758
12.6928
15.3848
10
12.7833
9.0038
5.0554
15.5261
13.1333
11.0639
18.7931
14.1998
28
.0195
15
6.3180
5.6945
3.1558
9.2451
8.0197
5.6668
12.2883
9.0965
6.8549
5
R
G
B
I
ma
ge
8.9441
6.6412
3.8789
13.0898
11.0668
7.9439
14.0743
13.1338
15.2404
10
12.8224
8.0275
5.0047
13.9544
13.0976
12.7071
15.0806
13.9715
28.4092
15
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
P
e
r
for
manc
e
analys
is
of
image
tr
ans
mis
s
ion
w
it
h
v
ar
ious
c
hanne
l
…
(
M
ar
w
a
J
alee
l
M
ohs
in
)
1163
T
a
ble
6
.
C
ompar
is
on
M
S
E
f
or
dif
f
e
r
e
nt
im
a
ge
s
,
m
odulation
s
c
he
mes
a
nd
f
il
ter
s
(
R
a
yleigh
f
a
ding
c
ha
nne
l)
64-
QAM
16-
QAM
Q
P
S
K
S
N
R
c
ha
nne
l
(
dB
)
I
ma
ge
T
ype
s
M
e
di
a
n
f
il
te
r
(
dB
)
W
ie
ne
r
f
il
te
r
(
dB
)
W
it
hout
f
il
te
r
(
dB
)
M
e
di
a
n
f
il
te
r
(
dB
)
W
ie
ne
r
f
il
te
r
(
dB
)
W
it
hout
f
il
te
r
(
dB
)
M
e
di
a
n
f
il
te
r
(
dB
)
W
ie
ne
r
f
il
te
r
(
dB
)
W
it
hout
f
il
te
r
(
dB
)
2.1565e
+
03
2.4281e
+
03
7.1477e
+
03
1.222+
e
03
1.964+
e
03
4,268+
e
03
1.0893
e
+
03
1.9391
e
+
03
3.3067e
+
03
5
G
r
a
y
-
s
c
a
le
I
ma
ge
1.5650e
+
03
1.8417e
+
03
5.7548e
+
03
148.758
370.498
1.0074e
+
03
47.773
3
75.524
4
280.529
10
1.3269e
+
03
1.9419e
+
03
4.1520e
+
03
59.2264
86.6346
295.945
47.773
3
75.524
4
59.8434
15
1.4720e
+
03
1.7592e
+
03
5.5491e
+
03
440.757
558.290
2.3556e
+
03
119.93
0
549.05
2
1.5896
e+
03
5
R
G
B
I
ma
ge
787.369
1.4382e
+
03
4.8505e
+
03
88.531
182.282
650.849
58.792
7
81.936
6
97.2376
10
152.663
371.024
2.5014e
+
03
50.013
58.124
111.712
54.556
7
124.10
6
0.6638
15
T
a
ble
7
.
C
ompar
is
on
S
NR
im
a
ge
f
o
r
dif
f
e
r
e
nt
im
a
ge
s
,
modul
a
ti
on
s
c
he
mes
a
nd
f
il
ter
s
(
R
a
yleigh
f
a
ding
c
ha
nne
l)
64-
QAM
16-
QAM
Q
P
S
K
S
N
R
c
ha
nne
l
(
dB
)
I
ma
ge
T
ype
s
M
e
di
a
n
f
il
te
r
(
dB
)
W
ie
ne
r
f
il
te
r
(
dB
)
W
it
hout
f
il
te
r
(
dB
)
M
e
di
a
n
f
il
te
r
(
dB
)
W
ie
ne
r
f
il
te
r
(
dB
)
W
it
hout
f
il
te
r
(
dB
)
M
e
di
a
n
f
il
te
r
(
dB
)
W
ie
ne
r
f
il
te
r
(
dB
)
W
it
hout
f
il
te
r
(
dB
)
6.1427
4.5241
3.0241
6.2104
5.5021
3.4230
7.1948
6.0964
4.5102
5
G
r
a
y
-
s
c
a
le
I
ma
ge
6.4252
5.6871
3.4241
11.2719
9.2904
7.1184
13.7848
12.7439
9.9832
10
8.8814
6.5924
3.8241
13.2718
12.4459
9.7784
1
4
.7848
1
3
.7439
18.3669
15
6.2524
4.
1265
3.0429
8.5430
8.0297
4.9035
11.3693
8.0657
5.7575
5
R
G
B
I
ma
ge
6.6524
4.
65
98
3.2546
1
2
.1241
11.0214
7.0805
12.9173
12.1966
11.0159
10
12.5219
8.7932
4.9287
1
2
.4210
11.9191
11.0731
13.0797
11.2950
22.6539
15
In
a
ddit
ion
,
F
igur
e
s
2
a
nd
3
s
how
c
ompar
is
on
o
f
r
e
c
ove
r
e
d
im
a
ge
s
a
t
the
s
ys
tem
r
e
c
e
iver
without
f
il
ter
,
with
media
n
a
nd
wie
ne
r
f
il
ter
s
f
o
r
di
f
f
e
r
e
nt
modul
a
ti
on
s
c
he
mes
ove
r
AW
GN
a
nd
R
a
yleigh
f
a
ding
c
ha
nne
ls
,
r
e
s
pe
c
ti
ve
ly.
How
e
ve
r
,
the
r
e
c
ove
r
e
d
i
m
a
ge
c
l
e
a
r
ne
s
s
c
ha
nge
d
by
the
modul
a
ti
on
type
a
nd
t
he
f
il
te
r
s
us
e
d
whic
h
gives
va
r
ious
va
lues
of
M
S
E
,
P
S
NR
a
nd
S
NR
f
o
r
AW
GN
c
ha
nne
l
a
nd
R
a
yleigh
f
a
ding
c
ha
nne
ls
,
r
e
s
pe
c
ti
ve
ly,
ba
s
e
d
on
the
s
im
ulation
r
e
s
ult
s
of
the
pr
opos
e
d
a
lgor
it
hm
to
dif
f
e
r
e
nt
va
lues
of
S
NR
.
I
t
is
s
e
e
n
that
a
s
e
xpe
c
ted
the
va
lues
of
the
P
S
NR
a
nd
S
NR
i
m
a
ge
of
the
r
e
c
e
ived
im
a
ge
s
incr
e
a
s
e
with
de
c
r
e
a
s
e
of
M
S
N
whe
n
incr
e
a
s
e
of
the
S
NR
va
lues
with
a
nd
witho
ut
f
il
ter
us
e
d,
whe
r
e
incr
e
a
s
e
the
va
lue
o
f
thes
e
p
a
r
a
mete
r
s
with
us
e
d
media
n
a
nd
W
iene
r
f
il
te
r
s
r
e
s
pe
c
ti
ve
ly.
F
ur
ther
mor
e
,
F
igur
e
4
dis
play
s
t
r
a
ns
mi
tt
e
d
inf
o
r
m
a
ti
on
a
s
digi
tal
s
ignal,
s
e
r
ial
s
ymbol
f
or
M
-
a
r
r
a
y
a
t
the
tr
a
ns
mi
tt
e
r
a
nd
wa
ve
f
or
m
f
or
M
-
a
r
r
a
y
a
c
c
or
d
ing
to
the
s
ymbol
ic
inf
or
mation
f
or
di
f
f
e
r
e
nt
mo
dulation
s
c
he
mes
.
F
inally,
F
igu
r
e
5
p
r
e
s
e
nts
r
e
-
obtain
s
ymbol
a
nd
r
e
c
e
ivi
ng
in
f
or
mation
a
s
digi
tal
s
ignal
a
f
ter
M
-
a
r
r
a
y
f
or
dif
f
e
r
e
nt
modul
a
ti
on
s
c
he
mes
.
How
e
ve
r
,
a
s
we
c
a
n
be
view
f
r
om
the
im
a
ge
s
r
e
c
e
ived,
the
ba
s
ic
pr
oblem
is
the
tr
a
ns
mi
s
s
ion
via
the
f
a
ding
c
ha
nne
ls
,
whic
h
r
e
pr
e
s
e
nt
the
a
c
tually
e
xis
ted
c
ha
nne
ls
in
the
wir
e
les
s
c
omm
unica
ti
o
ns
.
T
hus
,
th
e
r
e
s
ult
s
e
xplain
that
the
r
e
c
e
ived
im
a
ge
s
qua
li
ty
de
c
r
e
a
s
e
s
due
to
t
he
r
a
pid
moveme
nt
of
the
r
e
c
e
iver
.
I
n
a
ddit
ion
,
Q
P
S
K
mo
dulation
tec
hnique
give
a
be
tt
e
r
r
e
s
ult
o
f
r
e
c
ove
r
e
d
im
a
ge
s
a
nd
it
is
mor
e
im
mune
to
the
nois
e
than
other
mod
ulation
tec
hniques
,
be
ca
us
e
whe
n
the
number
of
tr
a
ns
mi
tt
e
d
bit
s
is
incr
e
a
s
e
,
they
a
r
e
mor
e
s
us
c
e
pti
ble
to
the
int
e
r
f
e
r
e
nc
e
.
I
n
o
ther
wor
ds
,
64
-
QA
M
modul
a
ti
on
tec
hnique
give
a
wor
s
e
r
e
s
ult
be
c
a
us
e
it
be
mor
e
s
us
c
e
pti
ble
t
o
the
int
e
r
s
im
ple
int
e
r
f
e
r
e
nc
e
(
I
S
I
)
.
F
r
om
F
igu
r
e
3
,
the
f
i
r
s
t
im
a
ge
s
in
the
lef
t
c
olum
n
a
r
e
the
or
igi
na
l
input
im
a
ge
s
.
T
he
gr
a
y
s
c
a
le
a
nd
R
GB
im
a
ge
s
a
r
e
the
im
a
ge
s
us
e
d
by
dif
f
e
r
e
nt
mod
ulation
tec
hniques
,
li
ke
;
QPS
K
,
16
-
QA
M
a
nd
64
-
Q
AM
will
be
a
ppli
e
d
in
the
c
omm
unica
ti
on
s
ys
tem.
How
e
ve
r
,
the
e
xe
c
uti
on
of
the
s
ys
t
e
m
will
no
t
be
f
a
s
t
whe
n
r
a
is
e
d
the
nois
e
leve
l
a
bove
the
c
r
it
ica
l
leve
l.
T
he
f
e
a
tur
e
of
the
de
s
igned
s
ys
tem
in
thi
s
wor
k
is
that
if
the
c
ha
nne
l
is
with
high
nois
e
,
t
his
make
s
the
s
ys
tem
ge
ne
r
a
tes
the
im
a
ge
qua
li
ty
wor
s
e
r
a
ther
than
los
ing
the
c
o
mpl
e
tely
i
mage
s
e
nt.
F
ur
ther
mor
e
,
the
r
e
s
ult
s
of
the
s
im
ulat
ion
a
na
lys
is
f
or
(
2D
)
Gr
a
y
a
nd
R
GB
im
a
ge
s
whic
h
uti
li
z
e
s
f
or
tr
a
ns
mi
s
s
ion
da
ta
a
r
e
pr
e
s
e
nted
in
c
ompar
is
on
to
the
theor
e
ti
c
a
l
r
e
s
ult
s
.
T
hus
,
the
r
e
c
ove
r
e
d
im
a
ge
qua
li
ty
is
be
s
t
f
or
hi
ghe
r
va
lues
f
r
om
S
NR
r
e
ga
r
dles
s
of
t
he
modul
a
ti
on
method
uti
l
ize
d.
W
hil
e
,
in
c
a
s
e
of
l
ow
-
S
NR
,
the
r
e
c
ove
r
e
d
im
a
ge
qua
li
ty
is
ve
r
y
low
be
c
a
us
e
of
high
a
mount
of
AW
GN
no
is
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
3
,
J
une
2020:
1159
-
1168
1164
(
a
)
(
b)
(
c
)
F
igur
e
2.
C
ompar
is
on
of
r
e
c
ove
r
e
d
i
mage
s
a
t
the
s
ys
tem
r
e
c
e
iver
without
f
il
ter
,
with
media
n
f
il
ter
a
n
d
with
wie
ne
r
f
il
ter
f
or
dif
f
e
r
e
nt
modul
a
ti
on
s
c
he
mes
ove
r
AW
GN
c
ha
nne
l
;
(
a
)
QPS
K
modul
a
ti
on
s
c
he
me,
(
b)
16
-
QA
M
modul
a
ti
on
s
c
he
me,
a
nd
(
c
)
64
-
QA
M
modul
a
ti
on
s
c
he
me
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
P
e
r
for
manc
e
analys
is
of
image
tr
ans
mis
s
ion
w
it
h
v
ar
ious
c
hanne
l
…
(
M
ar
w
a
J
alee
l
M
ohs
in
)
1165
(
a
)
(
b)
(
c
)
F
igur
e
3
.
C
ompar
is
on
of
r
e
c
ove
r
e
d
i
mage
s
a
t
t
he
s
ys
tem
r
e
c
e
iver
without
f
il
ter
,
with
media
n
f
il
ter
a
n
d
with
wie
ne
r
f
il
ter
f
or
dif
f
e
r
e
nt
modul
a
ti
on
s
c
he
mes
ove
r
R
a
yleigh
f
a
ding
c
ha
nne
l
;
(
a
)
QPS
K
modul
a
ti
on
s
he
me,
(
b)
16
-
QA
M
modul
a
ti
on
s
c
he
me,
a
nd
(
c
)
64
-
QA
M
modul
a
ti
on
s
he
me
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
3
,
J
une
2020:
1159
-
1168
1166
F
igur
e
4
.
T
r
a
ns
mi
tt
e
d
inf
o
r
mation
a
s
digi
tal
s
ignal,
s
e
r
ial
s
ymbol
f
or
M
-
a
r
r
a
y
a
t
the
t
r
a
ns
mi
tt
e
r
a
nd
wa
ve
f
or
m
f
o
r
M
-
a
r
r
a
y
a
c
c
or
ding
to
s
ymbol
ic
inf
or
mation
f
or
dif
f
e
r
e
nt
modul
a
ti
on
s
c
he
mes
F
igur
e
5
.
Re
-
obtain
s
ymbol
a
nd
r
e
c
e
ivi
ng
inf
or
mation
a
s
dig
it
a
l
s
ignal
a
f
te
r
M
-
a
r
r
a
y
f
or
dif
f
e
r
e
nt
modu
lation
s
c
he
mes
F
igur
e
6
de
mons
tr
a
tes
the
P
S
NR
vs
.
S
NR
c
ur
ve
s
ove
r
AW
GN
c
ha
nne
l
f
or
c
ompar
is
on
be
twe
e
n
gr
a
y
s
c
a
le
a
nd
R
G
B
im
a
ge
s
f
or
dif
f
e
r
e
nt
modul
a
ti
on
s
c
he
mes
.
I
t
is
s
hown
that
the
higher
P
S
NR
va
lue
c
a
n
be
ge
t
it
whe
n
us
e
d
QPS
K
c
ompar
e
with
other
modul
a
ti
on
s
c
he
mes
.
I
n
a
ddit
ion,
F
igur
e
7
e
xplains
the
P
S
NR
vs
.
S
NR
c
ur
ve
s
ove
r
AW
GN
c
ha
nne
l
without
f
il
ter
,
with
me
dian
f
il
ter
a
nd
with
W
iene
r
f
il
ter
,
r
e
s
pe
c
ti
ve
ly,
us
ing
QPS
K
modul
a
ti
on
s
c
he
me.
As
e
xpe
c
ted,
it
is
s
e
e
n
that
incr
e
a
s
e
of
the
P
S
NR
f
or
the
r
e
c
e
ived
i
mage
without
a
nd
with
(
media
n
a
nd
W
iene
r
)
f
il
te
r
s
us
e
d,
whe
r
e
th
is
is
im
pr
oving
the
qua
li
ty
of
im
a
ge
f
or
highe
r
P
S
NR
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
P
e
r
for
manc
e
analys
is
of
image
tr
ans
mis
s
ion
w
it
h
v
ar
ious
c
hanne
l
…
(
M
ar
w
a
J
alee
l
M
ohs
in
)
1167
F
igur
e
6.
P
S
NR
vs
.
S
NR
f
or
AW
GN
c
ha
nne
l
a
nd
dif
f
e
r
e
nt
modu
lation
s
c
he
mes
F
igur
e
7.
P
S
NR
vs
.
S
NR
f
or
AW
GN
c
ha
nne
l
a
nd
QPS
K
modul
a
ti
on
s
c
he
me
4.
CONC
L
UD
I
NG
RE
M
AR
KS
I
n
thi
s
pa
pe
r
,
we
a
na
lyze
the
pe
r
f
o
r
manc
e
of
im
a
ge
tr
a
ns
mi
s
s
ion
ove
r
wir
e
les
s
digi
tal
c
omm
unica
ti
ons
thr
ough
pr
e
s
e
nt
a
c
ompar
is
on
be
twe
e
n
da
ta
ha
ndli
n
g
us
ing
dif
f
e
r
e
nt
modul
a
ti
on
s
c
he
m
e
s
unde
r
va
r
ious
c
ha
nne
l
c
ondit
ions
a
nd
to
identi
f
y
the
be
tt
e
r
a
ppr
op
r
ia
te
m
odulation
tec
hnique
f
or
tr
a
ns
mi
tt
ing
im
a
ge
s
.
Ge
ne
r
a
ll
y,
we
e
xa
mi
ne
d
the
inf
luenc
e
of
two
va
r
ios
c
ha
nne
ls
of
AW
GN
a
nd
R
a
yleigh
f
a
ding
on
im
a
ge
tr
a
ns
f
e
r
in
the
wir
e
les
s
tr
a
ns
c
e
iver
m
ode
l.
How
e
ve
r
,
the
t
r
a
ns
mi
s
s
ion
of
i
mage
s
thr
ough
the
c
omm
unica
ti
on
s
ys
t
e
m
is
pe
r
f
o
r
med
by
us
e
d
digi
tal
modul
a
ti
on
tec
hniques
a
s
we
ll
a
s
th
e
r
e
s
ult
s
a
r
e
got
th
r
ough
us
e
d
M
a
tl
a
b
s
of
twa
r
e
t
ha
t
wa
s
int
r
oduc
e
d
to
de
s
ign
a
nd
im
pleme
nt
the
wir
e
les
s
digi
tal
c
omm
u
nica
ti
on
s
ys
tem.
T
hus
,
the
c
a
r
r
ied
ou
t
of
thi
s
s
ys
tem
whic
h
r
e
a
c
he
s
to
a
c
e
r
tain
leve
l
of
nois
e
f
r
om
the
c
r
it
ica
l
c
ha
nne
l
is
a
c
c
e
ptable
.
I
n
a
nother
mea
ning,
whe
n
r
a
is
e
d
the
nois
e
leve
l
a
bove
the
c
r
it
ica
l
leve
l,
the
s
ys
tem
pe
r
f
o
r
manc
e
c
a
nnot
be
ve
r
y
quickly
.
T
he
s
im
ulation
r
e
s
ult
s
s
howe
d
that
the
AW
GN
c
h
a
nne
l
gives
the
be
s
t
r
e
s
ult
than
the
R
a
yleigh
f
a
din
g
thr
ough
us
e
d
the
s
a
me
S
NR
,
whic
h
c
omes
with
a
les
s
e
f
f
ic
ient
r
e
s
ult
s
.
I
n
a
ddit
ion
,
whe
n
the
c
ha
nne
l
of
the
c
ur
r
e
ntl
y
de
s
igned
s
y
s
tem
is
in
high
nois
e
,
the
s
ys
tem
ge
n
e
r
a
tes
wor
s
ing
in
im
a
ge
qua
li
ty
ins
tea
d
of
los
ing
the
s
e
nt
i
mage
c
ompl
e
tely.
AC
KNOWL
ODGME
NT
T
his
wor
k
is
s
uppor
ted
in
pa
r
t
by
E
nginee
r
ing
T
e
c
h
nica
l
C
oll
e
ge
-
Na
jaf
,
Al
-
F
ur
a
t
Al
-
Aw
s
a
t
T
e
c
hnica
l
Unive
r
s
it
y
a
nd
in
pa
r
t
by
the
Unive
r
s
it
i
M
a
lays
ia
P
a
ha
ng
(
ww
w.
ump
.
e
du.
my)
via
R
e
s
e
a
r
c
h
Gr
a
nt
U
M
P
-
I
B
M
C
e
ntr
e
of
E
xc
e
ll
e
nc
e
R
DU
190304.
Evaluation Warning : The document was created with Spire.PDF for Python.