I
nte
rna
t
io
na
l J
o
urna
l o
f
I
nfo
rm
a
t
ics a
nd
Co
m
m
un
ica
t
io
n T
ec
hn
o
lo
g
y
(
I
J
-
I
CT
)
Vo
l.
1
5
,
No
.
1
,
Ma
r
ch
20
2
6
,
p
p
.
28
7
~
30
1
I
SS
N:
2252
-
8
7
7
6
,
DOI
:
1
0
.
1
1
5
9
1
/iji
ct
.
v
1
5
i
1
.
pp
28
7
-
30
1
287
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ict.
ia
esco
r
e.
co
m
Dev
elo
pment of
ma
chine learning
techniqu
es for au
tom
a
tic
mo
dula
tion cla
ss
ificatio
n and
p
erfo
rma
nce ana
ly
sis
under
AWG
N and
fadin
g
channels
P
.
G
.
Va
rna
K
u
m
a
r
Reddy
,
M
.
M
ee
na
D
e
p
a
r
t
me
n
t
o
f
EC
E
,
V
e
l
s I
n
st
i
t
u
t
e
o
f
S
c
i
e
n
c
e
,
Te
c
h
n
o
l
o
g
y
a
n
d
A
d
v
a
n
c
e
d
S
t
u
d
i
e
s,
C
h
e
n
n
a
i
,
I
n
d
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Oct
7
,
2
0
2
4
R
ev
is
ed
J
u
l 1
7
,
2
0
2
5
Acc
ep
ted
Oct
7
,
2
0
2
5
Au
to
m
a
ti
c
m
o
d
u
latio
n
c
las
sifica
ti
o
n
(AMC)
is
e
ss
e
n
ti
a
l
in
m
o
d
e
r
n
wire
les
s
c
o
m
m
u
n
ica
ti
o
n
f
o
r
o
p
t
imiz
in
g
sp
e
c
tru
m
u
sa
g
e
a
n
d
a
d
a
p
ti
v
e
sig
n
a
l
p
ro
c
e
ss
in
g
.
Th
is
st
u
d
y
e
x
p
l
o
re
s
th
e
u
se
o
f
v
a
ri
o
u
s
m
a
c
h
in
e
lea
r
n
in
g
(M
L)
m
e
th
o
d
s
f
o
r
AMC,
f
o
c
u
si
n
g
o
n
t
h
e
ir
p
e
rfo
rm
a
n
c
e
i
n
a
d
d
it
i
v
e
w
h
it
e
G
a
u
ss
ian
n
o
ise
(AWG
N)
a
n
d
fa
d
i
n
g
c
h
a
n
n
e
ls.
Th
is
stu
d
y
e
v
a
lu
a
tes
o
f
M
L
c
las
sifiers
su
c
h
a
s
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
s
(S
VM),
K
-
n
e
a
re
st
n
e
ig
h
b
o
rs
(KN
N),
d
e
c
isio
n
tree
s
(DT),
a
n
d
e
n
se
m
b
l
e
m
e
th
o
d
s
with
a
d
a
tas
e
t
sp
a
n
n
i
n
g
sig
n
a
l
-
to
-
n
o
ise
ra
ti
o
s
(S
NRs
)
fr
o
m
-
30
d
B
t
o
+
3
0
d
B
.
Hi
g
h
e
r
-
o
rd
e
r
sta
ti
stica
l
fe
a
tu
re
s
in
c
lu
d
in
g
m
o
m
e
n
ts
a
n
d
c
u
m
u
lan
ts
a
re
u
se
d
t
o
trai
n
th
e
c
las
sifiers
fo
r
AMC.
P
e
rfo
rm
a
n
c
e
is
m
e
a
su
re
d
in
term
s
o
f
c
las
sifica
ti
o
n
a
c
c
u
ra
c
y
a
n
d
c
o
m
p
u
tati
o
n
a
l
e
fficie
n
c
y
a
c
ro
ss
d
iffere
n
t
S
NR l
e
v
e
ls.
T
h
e
fin
d
i
n
g
s
sh
o
w t
h
a
t
li
n
e
a
r
S
VM,
fin
e
KN
N,
a
n
d
fin
e
tree
s
c
o
n
siste
n
tl
y
a
c
h
iev
e
d
h
i
g
h
c
las
sifica
ti
o
n
a
c
c
u
ra
c
y
,
e
v
e
n
a
t
l
o
w
S
NRs
.
F
r
o
m
t
h
e
a
n
a
ly
sis
,
it
i
s
o
b
se
rv
e
d
th
a
t
li
n
e
a
r
S
VM
a
n
d
fi
n
e
KNN
a
c
h
iev
e
o
v
e
r
9
6
%
a
c
c
u
ra
c
y
a
t
0
d
B
S
NR.
Th
e
se
c
las
sifiers
d
e
m
o
n
stra
te
sig
n
ifi
c
a
n
t
ro
b
u
st
n
e
ss
,
m
a
in
tain
i
n
g
p
e
rfo
rm
a
n
c
e
in
c
h
a
ll
e
n
g
i
n
g
n
o
is
e
c
o
n
d
it
i
o
n
s.
T
h
e
re
se
a
rc
h
h
ig
h
li
g
h
ts
t
h
e
p
ro
m
ise
o
f
M
L
tec
h
n
iq
u
e
s
in
imp
ro
v
in
g
AMC,
p
r
o
v
i
d
in
g
a
d
e
tailed
c
o
m
p
a
riso
n
o
f
c
las
sifiers
a
n
d
th
e
i
r
stre
n
g
t
h
s.
K
ey
w
o
r
d
s
:
AM
C
AW
GN
DT
E
n
s
em
b
le
class
if
ier
s
Fad
in
g
ch
an
n
els
KNN
SVM
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
P.
G.
Var
n
a
Ku
m
ar
R
ed
d
y
Dep
ar
tm
en
t o
f
E
C
E
,
Vels I
n
s
titu
te
o
f
Scien
ce
,
T
ec
h
n
o
lo
g
y
a
n
d
Ad
v
a
n
ce
d
Stu
d
ies
C
h
en
n
ai,
I
n
d
ia
E
m
ail: v
ar
u
n
k
u
m
ar
r
ed
d
y
p
g
.
p
h
d
1
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
I
n
m
o
d
er
n
wir
eless
co
m
m
u
n
i
ca
tio
n
s
y
s
tem
s
,
ef
f
icien
t
u
tili
za
tio
n
o
f
t
h
e
r
ad
i
o
f
r
e
q
u
en
c
y
s
p
ec
tr
u
m
is
p
ar
am
o
u
n
t.
T
h
e
in
cr
ea
s
in
g
d
e
m
an
d
f
o
r
wir
eless
s
er
v
ices
an
d
th
e
lim
ited
av
ailab
ilit
y
o
f
s
p
ec
tr
u
m
r
eso
u
r
ce
s
h
av
e
n
ec
ess
itated
th
e
d
ev
elo
p
m
en
t
o
f
ad
v
an
ce
d
tech
n
o
lo
g
ies
th
at
ca
n
o
p
tim
ize
s
p
ec
tr
u
m
u
s
ag
e.
On
e
s
u
ch
tech
n
o
lo
g
y
is
au
to
m
atic
m
o
d
u
latio
n
class
if
icatio
n
(
AM
C
)
,
wh
ich
p
lay
s
a
c
r
itical
r
o
le
in
c
o
g
n
itiv
e
r
a
d
io
(
C
R
)
n
etwo
r
k
s
an
d
v
ar
io
u
s
m
ilit
ar
y
an
d
civ
ilian
ap
p
licatio
n
s
.
AM
C
en
ab
les
th
e
id
e
n
tific
atio
n
o
f
th
e
m
o
d
u
latio
n
s
ch
em
e
o
f
r
ec
eiv
ed
s
ig
n
als
with
o
u
t
p
r
io
r
k
n
o
wled
g
e,
th
er
eb
y
f
ac
ilit
atin
g
ad
a
p
tiv
e
s
ig
n
al
p
r
o
ce
s
s
in
g
,
in
ter
f
er
en
ce
m
a
n
ag
em
e
n
t,
an
d
s
ec
u
r
e
co
m
m
u
n
icatio
n
[
1
]
.
T
h
e
s
ig
n
if
ican
ce
o
f
AM
C
s
tem
s
f
r
o
m
its
ab
ilit
y
to
en
h
an
ce
th
e
p
er
f
o
r
m
a
n
ce
an
d
r
eliab
ilit
y
o
f
wir
eless
co
m
m
u
n
icatio
n
s
y
s
tem
s
.
B
y
ac
cu
r
a
tely
id
en
tify
in
g
th
e
m
o
d
u
latio
n
ty
p
e,
AM
C
s
u
p
p
o
r
ts
d
y
n
am
ic
s
p
ec
tr
u
m
ac
ce
s
s
,
allo
win
g
u
n
licen
s
ed
u
s
er
s
to
o
p
p
o
r
tu
n
is
tically
u
tili
ze
th
e
licen
s
ed
s
p
ec
tr
u
m
with
o
u
t
ca
u
s
in
g
h
ar
m
f
u
l
in
ter
f
er
en
ce
to
p
r
im
ar
y
u
s
er
s
[
2
]
.
T
h
is
ca
p
ab
ilit
y
is
ess
en
tial
f
o
r
th
e
ef
f
icien
t
im
p
lem
en
tatio
n
o
f
CR
n
etwo
r
k
s
,
wh
ich
aim
to
ad
d
r
ess
th
e
s
p
ec
tr
u
m
s
ca
r
city
p
r
o
b
lem
.
I
n
m
ilit
ar
y
a
p
p
lic
atio
n
s
,
AM
C
is
v
ital
f
o
r
e
lectr
o
n
ic
war
f
ar
e
an
d
s
u
r
v
e
illan
ce
,
wh
er
e
th
e
id
en
tific
atio
n
o
f
en
e
m
y
co
m
m
u
n
icatio
n
s
ig
n
als
is
cr
u
cial.
I
n
civ
ilian
ap
p
licatio
n
s
,
A
MC
co
n
tr
ib
u
tes
to
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
1
,
Ma
r
ch
20
2
6
:
28
7
-
30
1
288
im
p
r
o
v
e
d
co
m
m
u
n
icatio
n
q
u
a
lity
,
r
ed
u
ce
d
laten
cy
,
an
d
en
h
an
ce
d
s
ec
u
r
ity
b
y
e
n
ab
lin
g
ad
ap
tiv
e
m
o
d
u
latio
n
an
d
co
d
i
n
g
s
ch
em
es.
T
h
e
n
e
ed
f
o
r
r
o
b
u
s
t
an
d
ac
cu
r
ate
A
MC
s
y
s
tem
s
is
u
n
d
er
s
co
r
ed
b
y
th
e
d
i
v
er
s
e
an
d
d
y
n
am
ic
n
atu
r
e
o
f
wir
eless
en
v
ir
o
n
m
en
ts
,
wh
er
e
s
ig
n
als
ar
e
o
f
ten
s
u
b
jecte
d
to
v
a
r
y
i
n
g
lev
els
o
f
n
o
is
e,
in
ter
f
er
en
ce
,
a
n
d
f
ad
in
g
[
3
]
,
[
4
]
.
Sev
er
al
tech
n
iq
u
es
h
av
e
b
ee
n
d
ev
elo
p
ed
f
o
r
AM
C
,
ea
ch
w
ith
its
o
wn
s
et
o
f
m
er
its
an
d
d
em
er
its
.
T
h
ese
tech
n
iq
u
es
ca
n
b
e
b
r
o
ad
ly
ca
teg
o
r
ized
i
n
to
lik
elih
o
o
d
-
b
ased
m
eth
o
d
s
,
f
ea
t
u
r
e
-
b
ased
m
eth
o
d
s
,
an
d
ML
-
b
ased
m
eth
o
d
s
[
5
]
.
L
ik
el
ih
o
o
d
-
b
ased
m
et
h
o
d
s
,
s
u
ch
a
s
th
e
m
ax
im
u
m
lik
elih
o
o
d
cl
ass
if
ier
,
o
f
f
er
h
ig
h
ac
cu
r
ac
y
in
i
d
en
tify
in
g
m
o
d
u
latio
n
s
ch
em
es.
T
h
ey
ar
e
th
e
o
r
etica
lly
o
p
tim
al
u
n
d
er
ce
r
t
ain
co
n
d
itio
n
s
a
n
d
p
r
o
v
id
e
a
s
o
lid
s
tatis
tic
al
f
o
u
n
d
atio
n
f
o
r
m
o
d
u
latio
n
class
if
icatio
n
.
T
h
ese
m
eth
o
d
s
o
f
ten
r
eq
u
ir
e
p
r
ec
is
e
k
n
o
wled
g
e
o
f
th
e
ch
an
n
el
co
n
d
itio
n
s
an
d
s
ig
n
al
p
ar
am
eter
s
,
wh
ich
m
ay
n
o
t
b
e
r
ea
d
ily
av
ailab
le
in
p
r
ac
tical
s
ce
n
ar
io
s
.
Ad
d
itio
n
ally
,
th
e
y
ar
e
co
m
p
u
tatio
n
ally
in
ten
s
iv
e,
m
ak
in
g
th
em
less
s
u
itab
le
f
o
r
r
ea
l
-
tim
e
ap
p
licatio
n
s
[
6
]
,
[
7
]
.
Featu
r
e
-
b
ased
m
eth
o
d
s
ex
tr
ac
t
s
p
ec
i
f
ic
s
ig
n
al
ch
ar
ac
ter
is
tics
,
s
u
c
h
as
h
ig
h
er
-
o
r
d
er
s
tatis
t
ics
to
d
is
tin
g
u
is
h
b
et
wee
n
d
if
f
er
en
t
m
o
d
u
latio
n
t
y
p
es.
T
h
ese
m
eth
o
d
s
ar
e
less
co
m
p
u
tatio
n
ally
d
em
an
d
in
g
co
m
p
a
r
ed
to
lik
el
ih
o
o
d
-
b
ased
m
eth
o
d
s
an
d
ca
n
b
e
e
f
f
ec
tiv
e
u
n
d
er
v
ar
i
o
u
s
ch
an
n
el
c
o
n
d
itio
n
s
.
T
h
e
p
er
f
o
r
m
a
n
ce
o
f
f
ea
tu
r
e
-
b
ased
m
eth
o
d
s
h
ea
v
ily
d
ep
e
n
d
s
o
n
th
e
q
u
ality
an
d
r
elev
an
c
e
o
f
th
e
e
x
tr
ac
ted
f
ea
tu
r
es.
T
h
e
y
m
ay
also
r
eq
u
ir
e
ex
ten
s
iv
e
f
ea
t
u
r
e
en
g
in
ee
r
in
g
an
d
d
o
m
ain
k
n
o
wled
g
e
t
o
id
en
tify
th
e
m
o
s
t
d
is
cr
im
in
ativ
e
f
ea
tu
r
es.
Ma
ch
i
n
e
lear
n
i
n
g
(
ML
)
-
b
ased
m
eth
o
d
s
h
av
e
g
ai
n
ed
p
o
p
u
lar
ity
d
u
e
to
th
ei
r
ab
ilit
y
to
lear
n
co
m
p
lex
p
atter
n
s
f
r
o
m
d
ata.
T
h
ese
m
eth
o
d
s
d
o
n
o
t
r
eq
u
ir
e
ex
p
licit
ch
an
n
el
k
n
o
wled
g
e
an
d
ca
n
ad
a
p
t
to
d
iv
er
s
e
an
d
d
y
n
am
ic
en
v
ir
o
n
m
en
ts
.
T
h
ey
o
f
f
er
h
ig
h
class
if
icatio
n
ac
cu
r
ac
y
a
n
d
r
o
b
u
s
tn
ess
ag
ain
s
t
n
o
is
e
an
d
in
ter
f
er
en
ce
.
T
h
e
p
r
im
ar
y
d
r
awb
ac
k
o
f
ML
-
b
ased
m
eth
o
d
s
is
th
eir
r
elian
ce
o
n
lar
g
e
lab
eled
d
atasets
f
o
r
tr
ain
in
g
,
wh
ich
m
ay
n
o
t
alwa
y
s
b
e
av
ailab
le
[
8
]
.
Ad
d
itio
n
a
lly
,
th
e
tr
ain
in
g
p
r
o
ce
s
s
ca
n
b
e
co
m
p
u
tatio
n
ally
in
ten
s
iv
e,
an
d
th
e
p
er
f
o
r
m
a
n
c
e
o
f
th
ese
m
eth
o
d
s
m
a
y
d
eg
r
a
d
e
if
th
e
tr
ai
n
in
g
d
ata
is
n
o
t r
e
p
r
esen
tativ
e
o
f
r
ea
l
-
wo
r
ld
co
n
d
itio
n
s
.
Giv
en
th
e
c
r
itical
r
o
le
o
f
A
MC
in
m
o
d
er
n
wir
eless
co
m
m
u
n
icatio
n
s
y
s
te
m
s
,
th
is
p
a
p
er
aim
s
to
ad
v
an
ce
th
e
f
ield
b
y
lev
er
a
g
in
g
ML
tech
n
iq
u
es
f
o
r
en
h
a
n
ce
d
m
o
d
u
latio
n
class
if
icatio
n
.
T
h
e
p
r
im
ar
y
f
o
cu
s
is
o
n
e
v
alu
atin
g
th
e
p
er
f
o
r
m
a
n
c
e
o
f
v
ar
i
o
u
s
ML
class
if
ier
s
ac
r
o
s
s
a
r
a
n
g
e
o
f
s
ig
n
al
-
to
-
n
o
i
s
e
r
atio
s
(
SNR
s
)
in
b
o
th
ad
d
itiv
e
wh
ite
g
a
u
s
s
ian
n
o
is
e
(
AW
GN
)
an
d
f
ad
in
g
ch
an
n
els.
T
h
is
p
a
p
er
m
a
k
es
s
ev
er
al
s
ig
n
if
ican
t
co
n
tr
ib
u
tio
n
s
an
d
ar
e
as f
o
llo
ws:
−
T
h
is
p
ap
er
co
n
d
u
cts
an
ex
te
n
s
iv
e
ev
alu
atio
n
o
f
v
ar
io
u
s
ML
class
if
ier
s
,
in
clu
d
in
g
s
u
p
p
o
r
t
v
ec
to
r
m
a
ch
in
es
(
SVM
)
,
K
-
n
ea
r
est
n
eig
h
b
o
u
r
s
(
KNN
)
,
d
ec
is
io
n
tr
ee
(
DT
)
,
an
d
en
s
em
b
le
m
eth
o
d
s
,
f
o
r
A
MC.
T
h
e
s
tu
d
y
s
y
s
tem
atica
lly
as
s
es
s
es
th
e
p
er
f
o
r
m
a
n
ce
o
f
th
ese
class
if
ier
s
ac
r
o
s
s
a
wid
e
r
an
g
e
SN
R
s
,
f
r
o
m
-
3
0
d
B
to
+3
0
d
B
,
p
r
o
v
id
in
g
v
alu
ab
le
in
s
ig
h
ts
in
to
th
eir
r
o
b
u
s
tn
ess
an
d
ef
f
ec
tiv
en
ess
u
n
d
er
d
if
f
er
en
t
n
o
is
e
co
n
d
itio
n
s
.
−
T
h
e
p
a
p
er
em
p
lo
y
s
h
ig
h
er
-
o
r
d
er
s
tatis
tical
f
ea
tu
r
es,
s
p
ec
if
i
ca
lly
m
o
m
e
n
ts
an
d
cu
m
u
lan
ts
,
ex
tr
ac
ted
f
r
o
m
m
o
d
u
latio
n
s
ig
n
als.
T
h
ese
f
ea
t
u
r
es a
r
e
cr
itical
f
o
r
d
is
tin
g
u
is
h
in
g
b
etwe
en
d
if
f
er
en
t
m
o
d
u
latio
n
s
ch
em
es.
−
T
h
e
p
ap
e
r
co
m
p
ar
es
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
ML
class
if
ier
s
u
n
d
er
b
o
th
AW
GN
an
d
f
ad
in
g
ch
a
n
n
els.
I
t
p
r
o
v
id
es
a
co
m
p
r
eh
e
n
s
iv
e
u
n
d
er
s
tan
d
in
g
o
f
h
o
w
th
es
e
class
if
ier
s
p
er
f
o
r
m
in
r
ea
l
-
wo
r
ld
wir
eless
en
v
ir
o
n
m
en
ts
,
wh
er
e
s
ig
n
als a
r
e
o
f
ten
s
u
b
jecte
d
to
v
a
r
io
u
s
i
m
p
air
m
en
ts
.
−
T
h
e
f
i
n
d
in
g
s
h
ig
h
lig
h
t
th
e
r
o
b
u
s
tn
ess
o
f
l
in
ea
r
SVM,
f
i
n
e
KNN,
a
n
d
f
in
e
tr
ee
s
,
wh
ich
co
n
s
is
ten
tly
ac
h
iev
e
h
ig
h
class
if
icatio
n
ac
cu
r
ac
y
ev
e
n
in
lo
w
SNR
co
n
d
itio
n
s
.
T
h
is
r
o
b
u
s
tn
ess
is
p
ar
ticu
lar
ly
s
ig
n
if
ican
t
f
o
r
p
r
ac
tical
ap
p
licatio
n
s
wh
er
e
wir
eless
co
m
m
u
n
icatio
n
s
y
s
tem
s
m
u
s
t
o
p
er
ate
r
eliab
ly
u
n
d
er
ch
allen
g
in
g
c
o
n
d
itio
n
s
.
B
y
ad
d
r
ess
in
g
th
ese
asp
ec
ts
,
th
is
p
ap
er
aim
s
to
p
r
o
v
id
e
v
alu
ab
le
in
s
ig
h
ts
an
d
p
r
ac
t
ical
g
u
id
elin
es
f
o
r
d
ev
elo
p
in
g
r
o
b
u
s
t
an
d
ef
f
icien
t
AM
C
s
y
s
tem
s
,
co
n
tr
ib
u
tin
g
to
th
e
o
p
tim
izatio
n
o
f
s
p
ec
t
r
u
m
u
s
ag
e
an
d
th
e
o
v
er
all
en
h
a
n
ce
m
en
t
o
f
wir
ele
s
s
co
m
m
u
n
icatio
n
tech
n
o
lo
g
ie
s
.
T
h
e
r
est
o
f
t
h
e
p
ap
er
is
o
r
g
a
n
ized
as
f
o
llo
ws:
s
ec
tio
n
2
p
r
o
v
id
es
a
d
etailed
liter
atu
r
e
r
ev
iew
o
f
ex
is
tin
g
AM
C
tech
n
iq
u
es,
h
ig
h
lig
h
tin
g
th
eir
m
e
r
its
an
d
d
em
er
its
.
Sectio
n
3
d
escr
ib
es
th
e
m
eth
o
d
o
lo
g
y
,
in
clu
d
in
g
f
ea
tu
r
e
e
x
tr
ac
tio
n
an
d
th
e
ML
alg
o
r
ith
m
s
em
p
l
o
y
ed
.
Sectio
n
4
p
r
esen
ts
th
e
s
im
u
latio
n
r
esu
lts
,
p
er
f
o
r
m
an
ce
a
n
aly
s
is
o
f
th
e
cl
ass
if
ier
s
u
n
d
er
v
ar
i
o
u
s
ch
an
n
e
l
co
n
d
itio
n
s
,
a
n
d
a
d
is
cu
s
s
io
n
o
n
th
e
im
p
licatio
n
s
o
f
th
e
f
in
d
in
g
s
.
Fin
ally
,
s
ec
tio
n
5
co
n
clu
d
es
th
e
p
ap
er
,
s
u
m
m
ar
izin
g
th
e
k
ey
co
n
tr
ib
u
ti
o
n
s
an
d
s
u
g
g
esti
n
g
d
ir
ec
tio
n
s
f
o
r
f
u
tu
r
e
s
tu
d
y
.
2.
L
I
T
E
R
AT
U
RE
R
E
VI
E
W
AM
C
h
as
witn
es
s
ed
co
n
s
id
er
ab
le
ad
v
a
n
ce
m
en
ts
,
tr
an
s
it
io
n
in
g
f
r
o
m
ea
r
ly
an
alo
g
m
o
d
u
latio
n
m
eth
o
d
s
to
s
o
p
h
is
ticated
d
ig
ital
m
o
d
u
latio
n
tech
n
iq
u
e
s
.
T
h
e
co
m
p
lex
ity
an
d
d
iv
er
s
ity
o
f
wir
eless
co
m
m
u
n
icatio
n
s
y
s
tem
s
h
av
e
d
r
iv
e
n
co
n
tin
u
o
u
s
r
esear
ch
a
n
d
in
n
o
v
atio
n
,
r
esu
ltin
g
in
a
b
r
o
ad
s
p
ec
tr
u
m
o
f
AM
C
s
tr
ateg
ie
s
.
T
h
ese
s
tr
at
eg
ies
ca
n
b
e
b
r
o
ad
ly
class
if
ied
in
to
d
ec
is
io
n
-
th
e
o
r
etic
(
DT
C
)
o
r
m
ax
im
u
m
lik
elih
o
o
d
(
ML
H)
m
eth
o
d
s
,
tr
an
s
f
o
r
m
d
o
m
ain
o
r
wav
elet
tr
an
s
f
o
r
m
(
W
T
)
m
eth
o
d
s
,
s
tatis
tical
m
eth
o
d
s
,
an
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
Dev
elo
p
men
t o
f m
a
ch
i
n
e
lea
r
n
in
g
tech
n
iq
u
es fo
r
a
u
t
o
ma
tic
mo
d
u
la
tio
n
…
(
P
.
G.
V
a
r
n
a
K
u
ma
r
R
ed
d
y
)
289
f
ea
tu
r
e
-
b
ased
(
FB
)
o
r
p
atter
n
r
ec
o
g
n
itio
n
(
PR
)
ap
p
r
o
ac
h
es
.
T
h
is
s
ec
tio
n
aim
s
to
h
ig
h
li
g
h
t
th
e
s
ig
n
if
ican
t
co
n
tr
ib
u
tio
n
s
m
ad
e
in
th
e
f
ield
,
ex
am
in
e
t
h
e
ev
o
lu
tio
n
o
f
A
MC
m
eth
o
d
o
lo
g
ies,
an
d
id
en
ti
f
y
th
e
s
tr
en
g
th
s
an
d
wea
k
n
ess
es o
f
ea
ch
ap
p
r
o
ac
h
.
2
.
1
.
Dec
is
io
n
-
t
heo
re
t
ic/ma
x
i
m
um
lik
eliho
o
d a
pp
ro
a
ches
DT
C
o
r
ML
H
ap
p
r
o
ac
h
es
a
r
e
am
o
n
g
t
h
e
m
o
s
t
p
o
p
u
lar
an
d
wid
ely
r
esear
ch
ed
m
eth
o
d
s
f
o
r
AM
C
.
T
h
ese
ap
p
r
o
ac
h
es
ar
e
m
o
tiv
a
ted
b
y
th
eir
o
p
tim
al
p
e
r
f
o
r
m
an
ce
f
o
r
k
n
o
wn
ch
an
n
el
p
ar
a
m
eter
s
an
d
m
o
d
els.
T
h
e
class
if
icatio
n
task
in
D
T
C
in
v
o
lv
es
two
p
r
im
ar
y
p
h
ases
.
I
n
th
e
f
ir
s
t
p
h
ase,
th
e
lik
elih
o
o
d
o
f
ea
c
h
m
o
d
u
latio
n
h
y
p
o
th
esis
is
ev
al
u
ated
b
ased
o
n
t
h
e
o
b
s
er
v
e
d
s
ig
n
al
s
am
p
les.
Featu
r
es
s
u
ch
as
s
q
u
ar
ed
s
ec
o
n
d
-
o
r
d
er
cy
clic
tem
p
o
r
al,
f
o
u
r
th
-
o
r
d
er
c
u
m
u
lan
ts
,
m
o
m
en
ts
,
in
s
tan
tan
eo
u
s
am
p
litu
d
e
,
p
h
a
s
e,
f
r
eq
u
en
c
y
,
an
d
s
ig
n
al
co
n
s
tellatio
n
s
ar
e
u
s
ed
to
co
n
s
tr
u
ct
lik
elih
o
o
d
r
atio
s
an
d
d
ec
is
io
n
b
o
u
n
d
ar
ies.
I
n
th
e
s
ec
o
n
d
p
h
ase,
th
ese
lik
elih
o
o
d
s
ar
e
co
m
p
ar
e
d
to
d
eter
m
i
n
e
th
e
m
o
s
t lik
ely
m
o
d
u
latio
n
s
ch
em
e.
T
ab
le
1
p
r
esen
ts
s
o
m
e
o
f
th
e
DT
C
ap
p
r
o
ac
h
es
f
o
r
AM
C
,
h
ig
h
lig
h
tin
g
th
eir
k
ey
f
ea
tu
r
es,
m
o
d
u
latio
n
class
es,
ac
cu
r
ac
y
,
an
d
th
ei
r
m
er
its
an
d
d
em
er
its
.
T
h
ese
m
eth
o
d
s
o
f
f
er
h
ig
h
ac
cu
r
ac
y
a
n
d
o
p
tim
al
p
er
f
o
r
m
a
n
ce
wh
en
s
ig
n
al
p
a
r
am
eter
s
ar
e
k
n
o
wn
,
p
r
o
v
id
i
n
g
a
s
o
lid
s
tatis
tical
f
o
u
n
d
atio
n
.
H
o
wev
er
,
th
e
y
ar
e
co
m
p
u
tatio
n
ally
in
ten
s
iv
e,
r
e
q
u
ir
e
p
r
ec
is
e
p
r
io
r
k
n
o
wled
g
e
o
f
s
ig
n
al
p
ar
am
eter
s
,
an
d
a
r
e
less
p
r
ac
tical
f
o
r
r
ea
l
-
tim
e
ap
p
licatio
n
s
,
b
ei
n
g
s
en
s
itiv
e
to
p
h
ase
an
d
f
r
eq
u
en
c
y
o
f
f
s
ets.
T
ab
le
1
.
L
iter
atu
r
e
o
n
DT
C
/ML
H
t
ec
h
n
iq
u
es
Ref
.
K
e
y
f
e
a
t
u
r
e
s
N
a
me
o
f
t
h
e
c
l
a
s
s
i
f
i
e
r
S
N
R
(
d
B
)
M
o
d
u
l
a
t
i
o
n
c
l
a
ss
e
s
A
c
c
u
r
a
c
y
(
%)
P
r
o
s
C
o
n
s
[
9
]
Lo
g
l
i
k
e
l
i
h
o
o
d
f
u
n
c
t
i
o
n
s
A
LR
T
-
2
0
t
o
5
B
P
S
K
,
Q
P
S
K
99
H
i
g
h
a
c
c
u
r
a
c
y
i
n
l
o
w
S
N
R
e
n
v
i
r
o
n
m
e
n
t
s
C
o
m
p
u
t
a
t
i
o
n
a
l
l
y
i
n
t
e
n
s
i
v
e
[
1
0
]
Li
k
e
l
i
h
o
o
d
-
r
a
t
i
o
s
B
a
y
e
s
10
B
P
S
K
,
Q
P
S
K
90
-
1
0
0
Ef
f
e
c
t
i
v
e
f
o
r
k
n
o
w
n
si
g
n
a
l
l
e
v
e
l
s
R
e
q
u
i
r
e
s
p
r
i
o
r
k
n
o
w
l
e
d
g
e
o
f
si
g
n
a
l
l
e
v
e
l
s
[
1
1
]
R
a
t
i
o
o
f
v
a
r
i
a
n
c
e
o
f
e
n
v
e
l
o
p
e
t
o
sq
u
a
r
e
o
f
m
e
a
n
Li
k
e
l
i
h
o
o
d
7
-
10
A
M
,
F
M
,
S
S
B
,
D
S
B
80
-
95
S
i
mp
l
e
i
mp
l
e
m
e
n
t
a
t
i
o
n
Li
mi
t
e
d
mo
d
u
l
a
t
i
o
n
c
l
a
ss
e
s
[
1
2
]
H
i
st
o
g
r
a
m
o
f
i
n
s
t
a
n
t
a
n
e
o
u
s
a
mp
,
p
h
a
se
,
a
n
d
f
r
e
q
Li
k
e
l
i
h
o
o
d
10
-
20
A
S
K
2
,
P
S
K
2
,
P
S
K
4
,
P
S
K
8
,
F
S
K
2
,
F
S
K
4
95
-
96
H
i
g
h
a
c
c
u
r
a
c
y
f
o
r
mu
l
t
i
p
l
e
m
o
d
u
l
a
t
i
o
n
c
l
a
ss
e
s
R
e
q
u
i
r
e
s
h
i
g
h
e
r
S
N
R
f
o
r
a
c
c
u
r
a
t
e
c
l
a
ss
i
f
i
c
a
t
i
o
n
[
1
3
]
Li
k
e
l
i
h
o
o
d
10
-
20
A
S
K
2
,
P
S
K
2
,
P
S
K
4
,
P
S
K
8
,
F
S
K
2
,
F
S
K
4
9
5
.
4
-
1
0
0
B
r
o
a
d
a
p
p
l
i
c
a
t
i
o
n
a
c
r
o
ss
m
o
d
u
l
a
t
i
o
n
c
l
a
ss
e
s
P
e
r
f
o
r
ma
n
c
e
d
e
p
e
n
d
s
o
n
f
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
[
1
4
]
,
[
1
5
]
4
p
a
r
a
m
e
t
e
r
s
d
e
r
i
v
e
d
f
r
o
m
I
n
st
a
n
t
a
n
e
o
u
s
a
mp
l
i
t
u
d
e
a
n
d
p
h
a
se
Li
k
e
l
i
h
o
o
d
10
AM
-
F
C
,
D
S
B
-
S
C
,
S
S
B
,
V
S
B
,
LSB,
U
S
B
,
FM
91
-
1
0
0
H
i
g
h
a
c
c
u
r
a
c
y
f
o
r
a
v
a
r
i
e
t
y
o
f
m
o
d
u
l
a
t
i
o
n
t
y
p
e
s
R
e
q
u
i
r
e
s
p
r
e
c
i
s
e
f
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
[
1
6
]
,
[
1
7
]
Q
u
a
ss
i
l
i
k
e
l
i
h
o
o
d
r
a
t
i
o
o
f
p
h
a
se
Li
k
e
l
i
h
o
o
d
-
2
t
o
8
B
P
S
K
,
Q
P
S
K
1
0
0
O
p
t
i
mal
p
e
r
f
o
r
m
a
n
c
e
f
o
r
sp
e
c
i
f
i
c
mo
d
u
l
a
t
i
o
n
t
y
p
e
s
Li
mi
t
e
d
t
o
p
h
a
se
-
mo
d
u
l
a
t
e
d
si
g
n
a
l
s
Li
k
e
l
i
h
o
o
d
25
PSK
-
1
6
,
1
6
-
QAM
1
0
0
H
i
g
h
a
c
c
u
r
a
c
y
f
o
r
h
i
g
h
-
o
r
d
e
r
m
o
d
u
l
a
t
i
o
n
H
i
g
h
c
o
m
p
u
t
a
t
i
o
n
a
l
c
o
m
p
l
e
x
i
t
y
[
1
8
]
Li
k
e
l
i
h
o
o
d
-
r
a
t
i
o
s
Q
u
a
si
-
A
L
R
T
-
1
0
t
o
2
B
P
S
K
,
Q
P
S
K
55
-
1
0
0
R
e
d
u
c
e
d
c
o
m
p
u
t
a
t
i
o
n
a
l
c
o
m
p
l
e
x
i
t
y
c
o
mp
a
r
e
d
t
o
A
LR
T
P
e
r
f
o
r
ma
n
c
e
v
a
r
i
e
s
si
g
n
i
f
i
c
a
n
t
l
y
w
i
t
h
S
N
R
[
1
9
]
Li
k
e
l
i
h
o
o
d
-
r
a
t
i
o
s
H
LR
T
-
2
0
t
o
20
B
P
S
K
,
Q
P
S
K
,
O
Q
P
S
K
90
R
o
b
u
st
p
e
r
f
o
r
ma
n
c
e
a
c
r
o
ss
a
r
a
n
g
e
o
f
S
N
R
s
H
i
g
h
c
o
m
p
u
t
a
t
i
o
n
a
l
d
e
m
a
n
d
2
.
2
.
T
ra
ns
f
o
rm
do
ma
in a
pp
ro
a
ches
T
r
an
s
f
o
r
m
d
o
m
ain
m
eth
o
d
s
f
o
r
AM
C
u
tili
ze
tech
n
iq
u
es
s
u
ch
as
W
T
to
an
aly
ze
th
e
s
ig
n
al
in
d
if
f
er
en
t
d
o
m
ain
s
,
allo
win
g
f
o
r
th
e
ex
tr
ac
tio
n
o
f
tr
an
s
ien
t
in
f
o
r
m
atio
n
ab
o
u
t
v
a
r
iatio
n
s
in
s
ig
n
al
am
p
litu
d
e,
p
h
ase,
an
d
f
r
eq
u
en
cy
.
T
h
ese
m
eth
o
d
s
ar
e
ef
f
ec
tiv
e
in
class
if
y
in
g
s
ig
n
als u
n
d
er
v
ar
io
u
s
ch
an
n
el
co
n
d
itio
n
s
b
u
t
ca
n
b
e
in
f
lu
e
n
ce
d
b
y
th
e
c
h
o
i
ce
o
f
tr
an
s
f
o
r
m
p
a
r
am
eter
s
an
d
th
e
f
ix
ed
wi
n
d
o
w
len
g
th
.
T
a
b
le
2
p
r
esen
ts
s
o
m
e
o
f
th
e
tr
an
s
f
o
r
m
d
o
m
ain
a
p
p
r
o
ac
h
es
f
o
r
AM
C
,
h
ig
h
lig
h
tin
g
th
eir
k
e
y
f
ea
t
u
r
es,
m
o
d
u
latio
n
class
es,
ac
cu
r
ac
y
,
an
d
th
eir
m
er
its
an
d
d
em
er
its
.
T
r
an
s
f
o
r
m
d
o
m
ain
m
eth
o
d
s
o
f
f
er
a
g
o
o
d
s
ig
n
al
an
aly
s
is
to
o
l
f
o
r
tim
e
-
v
a
r
y
in
g
s
ig
n
als
b
u
t r
eq
u
ir
e
ca
r
ef
u
l sele
ctio
n
o
f
p
ar
am
eter
s
an
d
ca
n
b
e
co
m
p
u
tatio
n
ally
d
em
a
n
d
in
g
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
1
,
Ma
r
ch
20
2
6
:
28
7
-
30
1
290
T
ab
le
2
.
C
o
m
p
a
r
ativ
e
an
aly
s
is
o
f
tr
an
s
f
o
r
m
d
o
m
ain
tech
n
iq
u
es
Ref
.
K
e
y
f
e
a
t
u
r
e
s
N
a
me
o
f
t
h
e
c
l
a
ss
i
f
i
e
r
S
N
R
(
d
B
)
M
o
d
u
l
a
t
i
o
n
c
l
a
sses
A
c
c
u
r
a
c
y
(
%)
P
r
o
s
C
o
n
s
[
2
0
]
S
p
e
c
t
r
o
g
r
a
ms
R
u
l
e
b
a
se
d
2
B
A
S
K
,
B
F
S
K
,
4
F
S
K
90
S
i
mp
l
e
i
mp
l
e
m
e
n
t
a
t
i
o
n
Li
mi
t
e
d
t
o
sp
e
c
i
f
i
c
S
N
R
[
2
1
]
I
n
st
a
n
t
a
n
e
o
u
s
f
r
e
q
u
e
n
c
y
,
m
a
i
n
-
l
o
b
e
w
i
d
t
h
s
,
p
e
a
k
t
o
si
d
e
-
l
o
b
e
r
a
t
i
o
S
TFT
0
-
12
A
S
K
,
M
-
a
r
y
F
S
K
,
PSK
90
-
1
0
0
H
i
g
h
a
c
c
u
r
a
c
y
F
i
x
e
d
w
i
n
d
o
w
l
e
n
g
t
h
l
i
mi
t
s
d
e
t
e
c
t
i
o
n
[
2
2
]
S
mo
o
t
h
-
w
i
n
d
o
w
e
d
W
i
g
n
e
r
V
i
l
l
e
b
i
s
p
e
c
t
r
u
m
R
u
l
e
b
a
se
d
0
-
26
A
S
K
,
M
-
a
r
y
F
S
K
80
-
95
Ef
f
e
c
t
i
v
e
a
c
r
o
ss
w
i
d
e
S
N
R
r
a
n
g
e
C
o
m
p
u
t
a
t
i
o
n
a
l
l
y
c
o
m
p
l
e
x
[
2
3
]
I
n
st
a
n
t
a
n
e
o
u
s
f
r
e
q
u
e
n
c
y
S
-
t
r
a
n
sf
o
r
m
(
S
T)
3
-
15
C
P
,
LF
M
,
B
F
S
K
,
B
P
S
K
9
7
.
2
5
H
i
g
h
a
c
c
u
r
a
c
y
Li
mi
t
e
d
t
o
sp
e
c
i
f
i
c
mo
d
u
l
a
t
i
o
n
s
[
2
4
]
N
u
mb
e
r
o
f
p
e
a
k
s
S
TFT,
H
T
2
Li
n
e
a
r
a
n
d
n
o
n
-
l
i
n
e
a
r
F
M
-
Ef
f
e
c
t
i
v
e
f
o
r
FM
s
i
g
n
a
l
s
Li
mi
t
e
d
t
o
F
M
si
g
n
a
l
s
[
2
5
]
I
n
st
a
n
t
a
n
e
o
u
s
a
mp
l
i
t
u
d
e
,
f
r
e
q
u
e
n
c
y
W
T,
S
T
10
-
15
A
M
,
F
M
,
M
A
S
K
,
M
F
S
K
,
P
S
K
(
M
=
2
,
4
a
n
d
8
)
99
-
1
0
0
H
i
g
h
a
c
c
u
r
a
c
y
f
o
r
a
v
a
r
i
e
t
y
o
f
mo
d
u
l
a
t
i
o
n
t
y
p
e
s
C
o
m
p
u
t
a
t
i
o
n
a
l
l
y
i
n
t
e
n
si
v
e
[
2
6
]
Ti
me
-
f
r
e
q
u
e
n
c
y
c
o
n
t
o
u
r
s
(
TFC)
M
o
d
i
f
i
e
d
S
T
-
10
t
o
2
0
M
A
S
K
,
M
F
S
K
,
P
S
K
(
M
=
2
,
4
a
n
d
8
)
-
Ef
f
e
c
t
i
v
e
f
o
r
l
o
w
S
N
R
Li
mi
t
e
d
mo
d
u
l
a
t
i
o
n
c
l
a
sses
[
2
7
]
TFC,
S
c
a
l
o
g
r
a
ms
M
o
d
i
f
i
e
d
S
T
-
20
t
o
2
0
-
-
Ef
f
e
c
t
i
v
e
f
o
r
a
w
i
d
e
r
a
n
g
e
o
f
S
N
R
s
C
o
m
p
l
e
x
i
mp
l
e
m
e
n
t
a
t
i
o
n
[
2
8
]
h
i
g
h
e
r
-
o
r
d
e
r
st
a
t
i
st
i
c
s
(
H
o
S
)
,
TFC
M
o
d
i
f
i
e
d
S
T,
F
F
T
0
t
o
20
M
A
S
K
,
M
F
S
K
,
P
S
K
(
M
=
2
,
4
a
n
d
8
)
90
-
1
0
0
H
i
g
h
a
c
c
u
r
a
c
y
H
i
g
h
c
o
m
p
u
t
a
t
i
o
n
a
l
c
o
m
p
l
e
x
i
t
y
2
.
3
.
St
a
t
is
t
ica
l
m
et
ho
ds
Statis
t
ical
m
eth
o
d
s
f
o
r
AM
C
f
o
cu
s
o
n
e
x
tr
ac
tin
g
m
o
d
u
latio
n
ty
p
es
b
y
lev
er
ag
in
g
th
e
Ho
S o
f
s
ig
n
als,
s
u
ch
as
m
o
m
en
ts
a
n
d
c
u
m
u
lan
ts
,
in
th
eir
c
o
m
p
lex
en
v
e
lo
p
es.
T
h
ese
m
eth
o
d
s
ca
n
c
lass
if
y
s
ig
n
als
b
y
co
n
s
id
er
in
g
n
o
n
-
lin
ea
r
ity
p
r
o
p
er
ties
an
d
cy
clo
s
tatio
n
ar
y
s
ta
tis
tics
.
W
h
ile
th
ey
ar
e
r
o
b
u
s
t
to
v
ar
io
u
s
ch
an
n
e
l
co
n
d
itio
n
s
,
th
eir
p
e
r
f
o
r
m
an
ce
is
h
ea
v
ily
d
ep
en
d
e
n
t
o
n
th
e
s
elec
tio
n
o
f
th
e
r
ig
h
t
f
ea
t
u
r
e
s
et.
T
ab
le
3
p
r
esen
ts
s
o
m
e
o
f
th
e
s
tatis
tical
ap
p
r
o
ac
h
es
f
o
r
AM
C
,
h
ig
h
lig
h
tin
g
t
h
eir
k
ey
f
ea
tu
r
es,
m
o
d
u
latio
n
class
es
,
ac
cu
r
ac
y
,
an
d
th
eir
m
er
its
an
d
d
e
m
er
i
ts
.
Statis
t
ical
m
eth
o
d
s
ar
e
r
elativ
ely
ea
s
y
to
im
p
lem
en
t
an
d
p
r
o
v
id
e
q
u
ick
r
ec
o
g
n
itio
n
o
f
m
o
d
u
latio
n
ty
p
es,
b
u
t c
ar
ef
u
l
f
ea
tu
r
e
s
elec
tio
n
is
cr
u
cial
f
o
r
o
p
tim
al
p
e
r
f
o
r
m
an
ce
.
T
ab
le
3
.
Statis
tical
m
eth
o
d
s
R
e
f
.
K
e
y
f
e
a
t
u
r
e
s
S
N
R
(
d
B
)
M
o
d
u
l
a
t
i
o
n
c
l
a
sses
A
c
c
u
r
a
c
y
(
%)
P
r
o
s
C
o
n
s
[
2
9
]
h
i
g
h
e
r
-
o
r
d
e
r
c
o
r
r
e
l
a
t
i
o
n
(
H
O
C
o
)
-
3
t
o
6
M
F
S
K
55
-
95
R
o
b
u
st
t
o
f
r
e
q
u
e
n
c
y
o
f
f
set
e
r
r
o
r
s
Li
mi
t
e
d
t
o
sp
e
c
i
f
i
c
mo
d
u
l
a
t
i
o
n
t
y
p
e
s
[
3
0
]
H
O
C
o
,
a
v
e
r
a
g
e
l
i
k
e
l
i
h
o
o
d
-
r
a
t
i
o
f
u
n
c
t
i
o
n
(
A
LF)
-
6
t
o
10
M
F
S
K
20
-
95
Ef
f
e
c
t
i
v
e
f
o
r
M
F
S
K
si
g
n
a
l
s
P
e
r
f
o
r
ma
n
c
e
v
a
r
i
e
s
si
g
n
i
f
i
c
a
n
t
l
y
w
i
t
h
S
N
R
[
3
1
]
H
O
C
o
4
-
12
M
F
S
K
65
-
99
H
i
g
h
a
c
c
u
r
a
c
y
i
n
sp
e
c
i
f
i
e
d
S
N
R
r
a
n
g
e
Li
mi
t
e
d
m
o
d
u
l
a
t
i
o
n
c
l
a
ss
e
s
[
3
2
]
M
o
me
n
t
s
-
5
.
8
t
o
5
.
5
2
A
S
K
,
2
P
S
K
,
4
P
S
K
,
M
S
K
,
2
F
S
K
9
7
.
9
-
1
0
0
H
i
g
h
a
c
c
u
r
a
c
y
a
c
r
o
ss
m
u
l
t
i
p
l
e
mo
d
u
l
a
t
i
o
n
s
R
e
q
u
i
r
e
s
p
r
e
c
i
s
e
m
o
me
n
t
c
a
l
c
u
l
a
t
i
o
n
[
3
3
]
M
o
me
n
t
s
-
1
0
t
o
10
M
P
S
K
50
-
95
Ef
f
e
c
t
i
v
e
f
o
r
M
P
S
K
si
g
n
a
l
s
S
e
n
s
i
t
i
v
e
t
o
n
o
i
s
e
a
t
l
o
w
e
r
S
N
R
s
[
3
4
]
C
y
c
l
i
c
c
u
mu
l
a
n
t
s
3
M
P
S
K
,
M
S
K
40
-
1
0
0
H
i
g
h
a
c
c
u
r
a
c
y
a
t
sp
e
c
i
f
i
e
d
S
N
R
P
e
r
f
o
r
ma
n
c
e
d
e
p
e
n
d
s
o
n
c
y
c
l
i
c
f
e
a
t
u
r
e
e
x
t
r
a
c
t
i
o
n
[
3
5
]
HOC
-
5
t
o
25
P
A
M
,
Q
A
M
,
M
P
S
K
55
-
1
0
0
B
r
o
a
d
a
p
p
l
i
c
a
b
i
l
i
t
y
,
h
i
g
h
a
c
c
u
r
a
c
y
R
e
q
u
i
r
e
s
l
a
r
g
e
n
u
m
b
e
r
o
f
sy
m
b
o
l
s f
o
r
h
i
g
h
a
c
c
u
r
a
c
y
[
3
6
]
HOC
5
,
1
0
Q
A
M
,
P
S
K
,
A
S
K
55
-
95
Ef
f
e
c
t
i
v
e
f
o
r
Q
A
M
a
n
d
P
S
K
P
e
r
f
o
r
ma
n
c
e
d
r
o
p
s a
t
l
o
w
e
r
S
N
R
s
[
3
7
]
HOC
5
,
7
,
1
0
4
-
Q
A
M
,
1
6
-
QAM
8
4
.
5
-
99
H
i
g
h
a
c
c
u
r
a
c
y
f
o
r
QAM
Li
mi
t
e
d
t
o
Q
A
M
si
g
n
a
l
s
[
3
8
]
HOC
-
1
0
t
o
10
16
-
Q
A
M
,
6
4
-
Q
A
M
,
B
P
S
K
,
Q
P
S
K
45
-
90
B
r
o
a
d
mo
d
u
l
a
t
i
o
n
c
l
a
ss
c
o
v
e
r
a
g
e
P
e
r
f
o
r
ma
n
c
e
d
r
o
p
s
si
g
n
i
f
i
c
a
n
t
l
y
a
t
l
o
w
e
r
S
N
R
s
2
.
4
.
F
e
a
t
ure
-
ba
s
ed
a
pp
ro
a
c
hes
T
h
e
lim
itatio
n
s
o
f
DT
C
an
d
s
tatis
tical
ap
p
r
o
ac
h
es
h
av
e
led
to
th
e
d
ev
elo
p
m
en
t
o
f
FB
ap
p
r
o
ac
h
es,
wh
ich
p
r
o
v
id
e
s
u
b
o
p
tim
al
p
er
f
o
r
m
an
ce
with
f
ewe
r
c
o
m
p
u
tatio
n
s
an
d
d
o
n
o
t
r
eq
u
i
r
e
p
r
io
r
in
f
o
r
m
atio
n
a
b
o
u
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
Dev
elo
p
men
t o
f m
a
ch
i
n
e
lea
r
n
in
g
tech
n
iq
u
es fo
r
a
u
t
o
ma
tic
mo
d
u
la
tio
n
…
(
P
.
G.
V
a
r
n
a
K
u
ma
r
R
ed
d
y
)
291
th
e
s
ig
n
al
an
d
ch
a
n
n
el.
T
h
es
e
ap
p
r
o
ac
h
es
ar
e
p
r
ac
tically
r
ea
lizab
le
an
d
ca
n
wo
r
k
u
n
d
er
d
if
f
er
e
n
t
n
o
is
y
co
n
d
itio
n
s
.
FB
ap
p
r
o
ac
h
es a
r
e
b
r
o
ad
l
y
ca
teg
o
r
ize
d
in
to
ML
ap
p
r
o
ac
h
es,
d
ee
p
lear
n
in
g
a
p
p
r
o
ac
h
es,
an
d
n
eu
r
al
n
etwo
r
k
m
eth
o
d
s
.
T
h
e
p
e
r
f
o
r
m
an
ce
o
f
FB
ap
p
r
o
ac
h
es
d
ep
e
n
d
s
o
n
th
e
ch
o
ice
o
f
th
e
f
ea
t
u
r
e
s
et
d
e
r
iv
ed
f
r
o
m
th
e
s
ig
n
als.
T
ab
le
4
p
r
esen
ts
s
o
m
e
o
f
th
e
FB
ap
p
r
o
ac
h
es
f
o
r
AM
C
,
h
ig
h
lig
h
tin
g
t
h
eir
k
ey
f
ea
tu
r
es,
m
o
d
u
latio
n
class
es,
ac
cu
r
ac
y
,
an
d
t
h
eir
m
e
r
its
an
d
d
em
er
i
ts
.
FB
ap
p
r
o
ac
h
es
elim
in
ate
th
e
n
ee
d
f
o
r
p
r
i
o
r
k
n
o
wled
g
e,
m
ak
in
g
t
h
em
s
u
itab
le
f
o
r
d
y
n
am
ic
an
d
n
o
is
y
en
v
ir
o
n
m
e
n
ts
.
Ho
wev
er
,
th
e
ch
o
ice
o
f
f
ea
tu
r
e
s
et
is
cr
u
cial
f
o
r
t
h
eir
ef
f
ec
tiv
e
n
ess
.
T
ab
le
4
.
Patter
n
r
ec
o
g
n
itio
n
m
eth
o
d
s
R
e
f
.
K
e
y
f
e
a
t
u
r
e
s
S
N
R
(
d
B
)
M
o
d
u
l
a
t
i
o
n
c
l
a
sses
A
c
c
u
r
a
c
y
(
%)
P
r
o
s
C
o
n
s
[
3
9
]
2
D
f
u
z
z
y
s
e
t
s
6
-
14
1
6
Q
A
M
,
3
2
Q
A
M
80
-
95
B
e
t
t
e
r
p
e
r
f
o
r
ma
n
c
e
o
v
e
r
t
r
a
d
i
t
i
o
n
a
l
M
L
i
n
n
o
i
s
y
e
n
v
i
r
o
n
m
e
n
t
s
Li
mi
t
e
d
t
o
t
w
o
-
c
l
a
ss
p
r
o
b
l
e
m
[
4
0
]
S
t
a
t
i
st
i
c
a
l
mo
m
e
n
t
s
>5
A
S
K
,
M
F
S
K
,
P
S
K
,
1
6
Q
A
M
82
-
96
Lo
w
c
o
m
p
l
e
x
i
t
y
,
n
o
p
r
i
o
r
a
w
a
r
e
n
e
ss
o
f
S
N
R
r
e
q
u
i
r
e
d
R
e
q
u
i
r
e
s
h
i
g
h
e
r
S
N
R
f
o
r
p
r
o
p
e
r
f
u
n
c
t
i
o
n
[
4
1
]
Tw
o
-
l
a
y
e
r
p
e
r
c
e
p
t
r
o
n
w
i
t
h
b
a
c
k
p
r
o
p
a
g
a
t
i
o
n
N
o
t
sp
e
c
i
f
i
e
d
R
e
c
t
-
Q
P
S
K
,
si
n
c
-
Q
P
S
K
,
S
Q
P
S
K
,
M
S
K
80
-
95
N
o
s
y
n
c
h
r
o
n
i
z
a
t
i
o
n
w
i
t
h
si
g
n
a
l
a
r
r
i
v
a
l
t
i
me
n
e
e
d
e
d
I
n
a
d
e
q
u
a
t
e
p
e
r
f
o
r
m
a
n
c
e
a
t
l
o
w
e
r
S
N
R
s
[
4
2
]
N
N
w
i
t
h
H
O
S
p
a
r
a
m
e
t
e
r
s
0
-
20
2
,
4
,
8
-
P
S
K
,
2
,
4
,
8
-
F
S
K
,
1
6
,
6
4
,
2
5
6
-
QAM
75
-
95
Ef
f
e
c
t
i
v
e
i
n
v
a
r
y
i
n
g
p
r
o
p
a
g
a
t
i
o
n
e
n
v
i
r
o
n
m
e
n
t
s
P
e
r
f
o
r
ma
n
c
e
d
e
g
r
a
d
e
s
i
n
mu
l
t
i
p
a
t
h
e
n
v
i
r
o
n
m
e
n
t
s
[
4
3
]
B
i
n
a
r
y
f
e
a
t
u
r
e
v
e
c
t
o
r
s fr
o
m
T
-
F
i
ma
g
e
s
N
o
t
sp
e
c
i
f
i
e
d
B
P
S
K
,
F
M
C
W
,
F
r
a
n
k
,
P
4
,
P
T1
93
A
u
t
o
n
o
m
o
u
s PR
a
l
g
o
r
i
t
h
m
P
o
o
r
e
r
p
e
r
f
o
r
ma
n
c
e
w
i
t
h
a
d
a
p
t
i
v
e
b
i
n
a
r
i
z
a
t
i
o
n
[
4
4
]
M
u
l
t
i
p
l
i
c
a
t
i
o
n
o
f
c
o
n
se
c
u
t
i
v
e
si
g
n
a
l
v
a
l
u
e
s
0
2
A
S
K
,
4
A
S
K
,
P
S
K
2
,
P
S
K
4
,
1
6
Q
A
M
5
0
a
t
0
dB
B
e
t
t
e
r
p
e
r
f
o
r
ma
n
c
e
t
h
a
n
t
r
a
d
i
t
i
o
n
a
l
M
L
c
l
a
ss
i
f
i
e
r
s
I
n
a
d
e
q
u
a
t
e
a
t
0
d
B
S
N
R
f
o
r
p
r
a
c
t
i
c
a
l
a
p
p
l
i
c
a
t
i
o
n
s
[
4
5
]
S
e
v
e
n
s
t
a
t
i
s
t
i
c
a
l
si
g
n
a
l
f
e
a
t
u
r
e
s
N
o
t
sp
e
c
i
f
i
e
d
C
W
,
A
M
,
LSB
,
U
S
B
,
FM
-
N
B
,
2
F
S
K
,
4
F
S
K
,
2
P
S
K
,
4
P
S
K
75
-
95
S
u
p
e
r
i
o
r
t
o
D
T
b
a
se
d
c
l
a
ss
i
f
i
e
r
s a
t
l
o
w
e
r
S
N
R
s
N
o
Q
A
M
si
g
n
a
l
s
i
n
c
l
u
d
e
d
[
4
6
]
W
T
c
o
e
f
f
i
c
i
e
n
t
s
10
2
A
S
K
,
4
A
S
K
,
2
F
S
K
,
4
F
S
K
,
2
P
S
K
,
4
P
S
K
,
M
S
K
,
1
6
Q
A
M
75
-
95
B
e
t
t
e
r
p
e
r
f
o
r
ma
n
c
e
a
t
1
0
d
B
S
N
R
I
n
a
d
e
q
u
a
t
e
a
t
S
N
R
l
e
ss
t
h
a
n
1
0
dB
[
4
7
]
B
o
o
t
s
t
r
a
p
t
e
c
h
n
i
q
u
e
a
n
d
r
a
d
i
a
l
b
a
si
s
f
u
n
c
t
i
o
n
N
N
Lo
w
V
a
r
i
o
u
s
a
n
a
l
o
g
a
n
d
d
i
g
i
t
a
l
80
-
95
Ef
f
e
c
t
i
v
e
i
n
l
o
w
S
N
R
a
n
d
f
a
d
i
n
g
c
h
a
n
n
e
l
s
Li
mi
t
e
d
m
o
d
u
l
a
t
i
o
n
c
l
a
ss
e
s
i
n
v
e
s
t
i
g
a
t
e
d
T
h
is
s
u
r
v
ey
u
n
d
er
s
co
r
es
th
e
n
ee
d
f
o
r
f
u
r
t
h
er
r
esear
ch
to
en
h
an
ce
AM
C
p
er
f
o
r
m
an
ce
in
d
y
n
am
ic
an
d
n
o
is
y
en
v
ir
o
n
m
e
n
ts
,
ad
d
r
ess
in
g
th
e
lim
itatio
n
s
id
e
n
tifie
d
in
ex
is
tin
g
ap
p
r
o
ac
h
es.
T
h
is
wo
r
k
aim
s
to
en
h
an
ce
AM
C
u
s
in
g
ML
tech
n
iq
u
es
b
y
f
o
c
u
s
in
g
o
n
r
o
b
u
s
t
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
s
elec
tio
n
,
i
m
p
lem
en
tin
g
d
iv
er
s
e
class
if
ier
s
s
u
ch
as
SVM,
K
NN,
an
d
en
s
em
b
le
class
if
ier
s
.
E
x
ten
s
iv
e
d
atasets
an
d
d
ata
au
g
m
e
n
tatio
n
tech
n
iq
u
es
will
b
e
em
p
lo
y
e
d
f
o
r
tr
ain
in
g
a
n
d
v
alid
atio
n
to
i
m
p
r
o
v
e
g
en
er
aliza
tio
n
.
Per
f
o
r
m
an
ce
o
p
tim
izatio
n
wi
ll
tar
g
et
h
y
p
er
p
ar
am
eter
tu
n
in
g
an
d
h
a
n
d
lin
g
class
im
b
alan
ce
s
to
ac
h
iev
e
a
b
alan
ce
b
etwe
en
ac
cu
r
ac
y
an
d
co
m
p
u
tatio
n
al
e
f
f
icien
cy
.
3.
M
E
T
H
O
D
T
h
is
s
ec
tio
n
p
r
esen
ts
th
e
d
etailed
f
r
am
ewo
r
k
u
s
ed
in
th
is
p
a
p
er
f
o
r
AM
C
an
d
d
escr
ib
es
th
e
d
if
f
er
en
t
ty
p
es o
f
ML
alg
o
r
ith
m
s
em
p
l
o
y
ed
.
3
.
1
.
F
r
a
m
ewo
r
k
Fig
u
r
e
1
illu
s
tr
ates
th
e
o
v
er
all
f
r
am
ewo
r
k
o
f
th
e
p
r
o
p
o
s
ed
AM
C
u
s
in
g
ML
tec
h
n
iq
u
es.
I
n
th
is
f
r
am
ewo
r
k
,
th
er
e
ar
e
s
ev
er
al
im
p
o
r
tan
t
s
tep
s
th
at
im
p
r
o
v
e
ac
cu
r
ac
y
an
d
ef
f
icien
c
y
in
th
e
d
etec
tio
n
an
d
class
if
icatio
n
o
f
v
ar
io
u
s
m
o
d
u
latio
n
class
ty
p
es
f
r
o
m
th
e
r
ec
eiv
ed
n
o
is
y
d
ata.
T
h
e
m
ajo
r
task
s
an
d
s
tep
s
in
th
is
p
r
o
p
o
s
ed
f
r
am
ewo
r
k
ar
e
d
etailed
in
s
u
b
s
ec
tio
n
s
b
elo
w
.
3
.
1
.
1
.
Da
t
a
s
et
g
ener
a
t
io
n wi
t
h SDR
t
estbed
I
n
m
o
d
er
n
co
m
m
u
n
icatio
n
s
y
s
tem
s
,
o
b
tain
in
g
r
ea
lis
tic
d
atasets
f
o
r
AM
C
is
cr
u
cial
f
o
r
d
ev
elo
p
in
g
r
o
b
u
s
t
ML
m
o
d
els.
A
SDR
te
s
tb
ed
p
r
o
v
id
es
t
h
e
f
le
x
ib
ilit
y
an
d
c
o
n
tr
o
l
n
ee
d
ed
f
o
r
r
ea
l
-
ti
m
e
d
ataset
cr
ea
tio
n
.
T
h
e
u
s
e
o
f
two
u
n
iv
er
s
al
s
o
f
t
war
e
r
ad
io
p
e
r
ip
h
er
als
(
USR
Ps
)
in
th
e
test
b
ed
s
etu
p
allo
ws
f
o
r
co
n
tr
o
lled
s
ig
n
al
tr
an
s
m
is
s
io
n
an
d
r
ec
ep
tio
n
,
en
ab
lin
g
th
e
g
e
n
er
atio
n
o
f
a
d
iv
e
r
s
e
s
et
o
f
m
o
d
u
lated
s
ig
n
al
class
es u
n
d
er
v
ar
y
in
g
ch
an
n
el
c
o
n
d
itio
n
s
.
T
h
is
s
etu
p
is
ess
en
tial
f
o
r
b
u
ild
i
n
g
a
d
ataset
th
at
m
im
ics
r
ea
l
-
w
o
r
ld
c
o
n
d
itio
n
s
a
n
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
1
,
Ma
r
ch
20
2
6
:
28
7
-
30
1
292
s
u
p
p
o
r
ts
th
e
d
e
v
elo
p
m
e
n
t
o
f
ef
f
ec
tiv
e
AM
C
alg
o
r
ith
m
s
.
Fig
u
r
e
2
p
r
esen
ts
th
e
SDR
t
estb
ed
f
o
r
d
ataset
g
en
er
atio
n
[
4
8
]
.
T
h
e
SDR
test
b
ed
co
m
p
r
is
es
two
USR
Ps
,
wh
er
e
o
n
e
ac
ts
as
th
e
tr
an
s
m
i
tter
an
d
th
e
o
th
er
as
th
e
r
ec
eiv
e
r
.
T
h
e
SDR
p
latf
o
r
m
p
r
o
v
id
es
a
f
le
x
ib
le
way
to
g
en
er
ate,
m
an
ip
u
late,
an
d
a
n
al
y
ze
r
a
d
io
s
ig
n
als
in
s
o
f
twar
e,
m
ak
in
g
it
h
ig
h
l
y
a
d
ap
tab
le
f
o
r
ex
p
er
im
en
tatio
n
an
d
d
ataset
cr
ea
tio
n
.
T
h
e
tr
an
s
m
itter
USR
P
i
s
p
r
o
g
r
a
m
m
ed
t
o
g
en
e
r
ate
d
if
f
e
r
en
t
class
es
o
f
m
o
d
u
lated
s
ig
n
als
s
u
ch
as
M
-
ar
y
P
SK
(
M=
2
,
4
,
an
d
8
)
,
4
-
QAM
,
16
-
QAM
,
an
d
6
4
-
QAM
.
T
h
es
e
s
ig
n
als
ar
e
tr
an
s
m
itted
o
v
er
th
e
air
in
r
ea
l
-
tim
e
to
th
e
r
ec
e
iv
er
USR
P,
wh
ich
ca
p
tu
r
es th
e
s
ig
n
als f
o
r
an
aly
s
is
an
d
s
to
r
ag
e.
Fig
u
r
e
1
.
Fra
m
ewo
r
k
f
o
r
AM
C
Fig
u
r
e
2
.
SDR
t
estb
ed
f
o
r
AM
C
d
ataset
cr
ea
tio
n
T
o
s
im
u
late
r
ea
l
-
w
o
r
ld
c
o
m
m
u
n
icatio
n
e
n
v
ir
o
n
m
en
ts
,
AW
GN
is
in
tr
o
d
u
ce
d
d
u
r
in
g
th
e
tr
an
s
m
is
s
io
n
p
h
ase.
T
h
is
n
o
is
e
m
im
ics
th
e
b
ac
k
g
r
o
u
n
d
in
ter
f
er
en
ce
co
m
m
o
n
ly
f
o
u
n
d
in
wir
eless
co
m
m
u
n
icatio
n
ch
an
n
els.
I
n
ad
d
itio
n
to
AW
GN,
f
ad
in
g
n
o
is
e
is
g
e
n
er
ate
d
b
y
v
ar
y
in
g
th
e
d
is
tan
ce
b
etwe
en
th
e
tr
an
s
m
itter
an
d
r
ec
eiv
er
.
As
th
e
d
is
tan
ce
in
cr
ea
s
es,
s
ig
n
al
s
tr
en
g
th
wea
k
e
n
s
,
in
tr
o
d
u
cin
g
f
ad
i
n
g
ef
f
ec
ts
s
u
ch
as
m
u
ltip
ath
in
ter
f
er
e
n
ce
an
d
s
ig
n
al
atten
u
atio
n
,
wh
ic
h
ar
e
co
m
m
o
n
i
n
r
ea
l
co
m
m
u
n
icatio
n
s
ce
n
ar
i
o
s
.
T
h
e
v
ar
iatio
n
in
d
is
tan
ce
,
co
m
b
in
ed
with
th
e
in
tr
o
d
u
cti
o
n
o
f
AW
GN,
allo
ws
f
o
r
th
e
cr
ea
tio
n
o
f
r
ea
lis
tic
s
ig
n
al
en
v
ir
o
n
m
en
ts
.
T
h
ese
ch
an
g
es
s
im
u
late
d
if
f
er
e
n
t
p
r
o
p
ag
atio
n
c
o
n
d
itio
n
s
,
s
u
ch
as
th
o
s
e
en
c
o
u
n
ter
e
d
in
u
r
b
a
n
,
r
u
r
al,
o
r
in
d
o
o
r
en
v
ir
o
n
m
en
ts
,
wh
er
e
s
ig
n
als
m
ay
ex
p
e
r
ien
ce
v
a
r
y
in
g
lev
els o
f
f
ad
i
n
g
a
n
d
in
ter
f
er
en
ce
.
At
th
e
r
ec
eiv
er
USR
P,
th
e
tr
an
s
m
itted
s
ig
n
als
n
o
w
em
b
e
d
d
ed
with
n
o
is
e
an
d
f
ad
in
g
ef
f
ec
ts
ar
e
ca
p
tu
r
ed
.
T
h
e
r
ec
eiv
e
d
s
ig
n
als
ar
e
s
to
r
e
d
i
n
r
ea
l
tim
e,
a
n
d
th
i
s
p
r
o
ce
s
s
is
r
ep
ea
te
d
f
o
r
ea
ch
class
o
f
m
o
d
u
lated
s
ig
n
als
at
d
if
f
er
e
n
t
SNR
lev
els.
B
y
s
y
s
tem
atica
lly
ad
ju
s
tin
g
th
e
SNR
,
th
e
d
ataset
ca
p
t
u
r
es
s
ig
n
als
u
n
d
er
v
ar
io
u
s
n
o
is
e
co
n
d
itio
n
s
,
e
n
a
b
lin
g
th
e
AM
C
s
y
s
tem
to
b
e
tr
ain
ed
a
n
d
e
v
alu
ated
u
n
d
er
r
ea
lis
tic
co
n
d
itio
n
s
.
T
h
is
ap
p
r
o
ac
h
en
s
u
r
es
th
at
t
h
e
m
ac
h
in
e
lear
n
in
g
m
o
d
els
ca
n
g
en
e
r
alize
well
to
d
if
f
er
en
t
co
m
m
u
n
icatio
n
en
v
ir
o
n
m
en
ts
.
T
h
e
r
e
p
ea
ted
p
r
o
ce
s
s
o
f
t
r
an
s
m
itti
n
g
a
n
d
r
ec
eiv
in
g
s
ig
n
als
u
n
d
e
r
v
a
r
y
in
g
c
o
n
d
itio
n
s
r
esu
lts
in
a
d
iv
er
s
e
d
ataset,
co
n
tain
in
g
s
ig
n
als
with
d
if
f
er
en
t
m
o
d
u
latio
n
s
ch
em
es,
n
o
is
e
lev
els,
an
d
ch
an
n
el
im
p
air
m
en
ts
.
T
h
is
d
ataset
is
th
en
lab
eled
ac
co
r
d
in
g
t
o
th
e
m
o
d
u
latio
n
t
y
p
e
an
d
SNR
v
alu
es,
p
r
o
v
id
i
n
g
a
co
m
p
r
eh
e
n
s
iv
e
tr
ain
in
g
s
et
f
o
r
th
e
AM
C
alg
o
r
ith
m
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
Dev
elo
p
men
t o
f m
a
ch
i
n
e
lea
r
n
in
g
tech
n
iq
u
es fo
r
a
u
t
o
ma
tic
mo
d
u
la
tio
n
…
(
P
.
G.
V
a
r
n
a
K
u
ma
r
R
ed
d
y
)
293
3
.
1
.
2
.
F
ea
t
ure
e
x
t
ra
ct
i
o
n
a
nd
m
o
del selec
t
io
n
I
t
d
ir
ec
tly
ex
tr
ac
ts
f
ea
tu
r
es,
wh
i
ch
in
clu
d
e
am
p
litu
d
e
,
p
h
a
s
e,
an
d
f
r
eq
u
e
n
cy
ch
ar
ac
te
r
is
tics
o
f
th
e
tim
e
-
d
o
m
ain
s
ig
n
al.
All
o
f
th
ese
f
ea
tu
r
es
ca
tch
s
o
m
e
o
f
th
e
ess
en
tial
p
r
o
p
er
ties
o
f
th
e
s
ig
n
al
u
s
ef
u
l
f
o
r
class
if
icatio
n
.
Fre
q
u
en
cy
-
d
o
m
ain
f
ea
tu
r
es
g
iv
e
i
n
f
o
r
m
atio
n
r
eg
ar
d
in
g
th
e
c
o
m
p
o
n
en
ts
th
at
co
n
s
titu
te
a
s
ig
n
al
an
d
th
eir
d
is
tr
ib
u
tio
n
.
C
o
n
s
id
er
h
ig
h
er
-
o
r
d
er
s
tatis
tics
,
s
u
c
h
as
cu
m
u
lan
ts
an
d
m
o
m
en
ts
,
wh
ich
ca
p
tu
r
e
th
e
n
o
n
lin
ea
r
r
elatio
n
s
h
ip
an
d
d
ep
en
d
e
n
cies
in
th
e
s
ig
n
al
to
p
r
o
v
id
e
m
o
r
e
d
is
cr
im
in
ativ
e
f
ea
tu
r
es
f
o
r
class
if
icatio
n
.
I
n
th
e
m
o
d
el
s
elec
tio
n
p
h
ase,
a
p
p
r
o
p
r
iate
m
a
ch
in
e
lear
n
i
n
g
alg
o
r
ith
m
s
f
o
r
AM
C
ar
e
s
elec
ted
.
T
h
e
ch
o
ice
o
f
alg
o
r
ith
m
d
ep
en
d
s
o
n
th
e
n
atu
r
e
o
f
th
e
class
if
icatio
n
task
an
d
s
p
ec
if
ic
n
ee
d
s
f
r
o
m
th
e
d
ataset.
3
.
1
.
3
.
M
o
del
t
ra
ini
ng
At
th
is
s
tag
e,
7
0
%,
1
5
%,
an
d
1
5
%
o
f
th
e
d
ataset
will
b
e
u
s
ed
f
o
r
tr
ai
n
in
g
,
v
alid
atio
n
,
a
n
d
test
in
g
,
r
esp
ec
tiv
ely
.
T
h
e
m
o
d
els
will
b
e
tr
ain
e
d
o
n
th
e
tr
ain
in
g
s
et.
C
r
o
s
s
-
v
alid
atio
n
will
b
e
u
s
ed
to
tu
n
e
th
e
h
y
p
er
p
ar
am
eter
s
,
an
d
t
h
e
b
est
m
o
d
el
will
b
e
ch
o
s
en
b
ased
o
n
v
alid
atio
n
p
er
f
o
r
m
an
ce
.
Fin
a
lly
,
th
e
test
s
et
p
er
f
o
r
m
an
ce
will b
e
ev
al
u
ated
with
th
e
m
o
d
el.
3
.
1
.
4
.
M
o
del
e
v
a
lua
t
io
n
I
n
th
is
s
tag
e,
test
th
e
tr
ain
ed
m
o
d
els
o
n
th
e
test
s
et,
ev
alu
atin
g
th
em
b
y
m
etr
ics
s
u
ch
as
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
th
e
F1
-
s
co
r
e.
C
o
m
p
ar
e
th
e
d
if
f
er
e
n
t
m
o
d
els
b
u
ilt
to
d
eter
m
i
n
e
wh
ich
is
th
e
b
est
alg
o
r
ith
m
f
o
r
a
p
ar
ticu
lar
m
o
d
u
latio
n
s
ch
em
e
to
ac
h
iev
e
o
p
tim
u
m
p
er
f
o
r
m
a
n
ce
in
p
r
ac
tice.
3
.
2
.
Dec
is
io
n
t
rees
DT
s
ar
e
p
o
p
u
lar
n
o
n
-
lin
ea
r
p
r
ed
ictio
n
m
o
d
els
in
ML
a
n
d
d
ata
m
in
in
g
,
u
s
ed
f
o
r
eith
e
r
cla
s
s
if
icatio
n
o
r
r
eg
r
es
s
io
n
p
r
o
b
lem
s
.
T
h
e
b
asic
id
ea
o
f
DT
i
s
to
r
e
cu
r
s
iv
ely
p
ar
titi
o
n
th
e
g
iv
en
d
ataset
in
to
n
o
n
-
o
v
er
lap
p
i
n
g
s
u
b
s
ets
b
ased
o
n
th
e
m
o
s
t
im
p
o
r
tan
t
f
ea
t
u
r
es
[
4
9
]
.
I
n
DT
s
,
a
tr
ee
s
tr
u
ctu
r
e
is
m
ad
e
wh
er
e
th
e
in
ter
n
al
n
o
d
e
wh
ich
ev
er
y
n
o
d
e
o
f
a
DT
r
e
p
r
esen
ts
a
f
ea
tu
r
e
an
d
ea
c
h
b
r
an
ch
r
ep
r
esen
ts
a
p
o
s
s
ib
le
o
u
tco
m
e
o
f
th
at
f
ea
tu
r
e.
T
h
e
lea
v
es
o
f
th
e
t
r
ee
r
e
p
r
esen
t
f
in
al
class
o
r
v
alu
e
p
r
ed
ictio
n
s
.
T
h
e
p
r
o
ce
s
s
o
f
p
a
r
titi
o
n
in
g
co
n
tin
u
es
u
n
til
eith
er
th
e
d
ata
v
alu
es
ar
e
s
ep
ar
ated
o
r
s
o
m
e
cr
iter
ia,
lik
e
m
ax
im
izin
g
th
e
d
ep
th
o
f
th
e
tr
ee
,
m
in
im
izin
g
th
e
s
am
p
les
p
er
leaf
o
r
s
o
f
o
r
th
,
is
m
et
o
r
.
I
t
allo
ws
th
e
m
o
d
el
to
m
ak
e
p
r
e
d
ictio
n
s
b
y
f
o
llo
win
g
th
r
o
u
g
h
th
e
b
r
a
n
ch
es
o
f
th
e
t
r
ee
b
ased
o
n
f
ea
tu
r
e
v
alu
es
u
n
til
it
ar
r
iv
es
at
a
leaf
t
h
at
co
n
tain
s
a
class
o
r
esti
m
ated
v
alu
e
to
p
r
ed
ic
t.
T
h
e
DT
s
h
o
u
ld
b
e
ea
s
ily
in
ter
p
r
etab
le
an
d
tr
a
n
s
p
ar
en
t,
wh
ich
is
u
s
ef
u
l
f
o
r
in
tu
itiv
ely
ex
p
lain
i
n
g
w
h
y
th
e
m
o
d
el
m
ak
es
ce
r
tain
d
ec
is
io
n
s
.
T
h
ey
h
elp
ex
p
lain
h
o
w
d
if
f
er
e
n
t
f
ea
tu
r
es
co
n
tr
ib
u
te
to
m
ak
in
g
a
p
r
e
d
i
ctio
n
,
th
u
s
m
ak
i
n
g
t
h
e
p
r
ed
i
ctio
n
tr
an
s
p
a
r
en
t
a
n
d
tr
u
s
tw
o
r
th
y
.
DT
s
ca
n
b
e
ap
p
lied
to
b
o
t
h
n
u
m
e
r
ical
an
d
ca
teg
o
r
ical
d
ata
an
d
ca
n
w
o
r
k
with
o
u
tlier
s
.
DT
s
ar
e
h
i
g
h
ly
v
u
l
n
er
ab
le
to
o
v
er
f
itti
n
g
,
esp
ec
ially
wh
en
a
tr
ee
is
p
ar
ticu
lar
ly
d
ee
p
.
T
h
e
y
m
ay
also
s
h
o
w
h
ig
h
v
ar
ian
ce
,
wh
er
eb
y
s
m
all
ch
an
g
es
in
th
e
d
ata
lead
to
q
u
ite
d
if
f
er
en
t
tr
ee
s
.
So
m
e
f
ea
tu
r
es
o
f
s
o
m
e
d
if
f
er
e
n
t
DT
cl
ass
if
ier
s
r
eg
ar
d
in
g
m
ax
im
u
m
d
ep
th
,
class
if
icatio
n
s
p
ee
d
,
class
if
y
in
g
ac
c
u
r
ac
y
,
an
d
f
lex
ib
ilit
y
.
T
h
e
f
in
e
DT
class
if
ier
,
w
ith
a
m
ax
im
u
m
d
ep
t
h
o
f
1
0
0
,
is
th
e
-
f
astes
t
cla
s
s
if
y
in
g
,
ac
cu
r
ate,
an
d
h
ig
h
l
y
v
iab
le
o
n
e.
T
h
e
m
ed
iu
m
DT
class
if
ier
h
as
a
m
ax
im
u
m
d
ep
th
o
f
2
0
;
it
is
also
f
ast
in
class
if
icatio
n
,
ac
cu
r
ate
,
b
u
t
with
m
ed
iu
m
f
lex
ib
ilit
y
.
L
ast
b
u
t
n
o
t
least,
t
h
e
co
a
r
s
e
DT
class
if
ier
h
as
a
m
a
x
im
u
m
d
e
p
th
o
f
4
;
its
clas
s
if
icatio
n
is
s
til
l f
ast,
b
u
t it
is
m
ed
iu
m
in
ac
c
u
r
ac
y
a
n
d
lo
w
in
f
lex
i
b
ilit
y
.
3
.
3
.
K
-
nea
re
s
t
neig
hb
o
rs
KNN
is
a
s
im
p
le,
h
en
ce
ef
f
ec
tiv
e,
n
o
n
p
a
r
am
etr
ic
class
if
icatio
n
an
d
r
eg
r
ess
io
n
alg
o
r
it
h
m
.
KNN
o
p
er
ates
b
ased
o
n
th
e
s
im
ilar
ity
p
r
in
cip
le:
class
if
y
in
g
a
d
ata
p
o
in
t
b
y
co
n
s
id
er
in
g
th
e
m
aj
o
r
ity
class
am
o
n
g
s
t
its
n
ea
r
est
n
eig
h
b
o
r
s
.
Fo
r
clas
s
if
icatio
n
,
KNN
ass
ig
n
s
a
cla
s
s
to
a
q
u
er
y
p
o
in
t
b
ased
o
n
th
e
m
ajo
r
ity
class
o
f
its
KNN
[
4
9
]
.
T
h
e
v
alu
e
o
f
K
tu
r
n
s
in
to
a
v
e
r
y
im
p
o
r
tan
t
h
y
p
er
p
a
r
am
eter
.
A
s
m
all
K
th
e
n
r
en
d
e
r
s
th
e
m
o
d
el
s
en
s
itiv
e
to
n
o
is
e,
w
h
ile
a
lar
g
e
K
p
r
o
v
id
es
a
s
m
o
o
th
e
r
d
ec
is
io
n
b
o
u
n
d
ar
y
b
u
t
p
r
o
b
a
b
ly
at
an
ad
d
itio
n
al
co
m
p
u
tatio
n
al
c
o
s
t.
So
m
e
o
f
th
e
ad
v
a
n
tag
es
o
f
KNN
i
n
clu
d
e
h
o
w
ea
s
y
it
is
to
im
p
lem
en
t
an
d
g
et
in
ter
p
r
etab
ilit
y
o
f
r
esu
lts
,
h
an
d
le
n
o
n
lin
ea
r
r
elatio
n
s
h
ip
s
,
an
d
it
g
en
er
ally
f
ar
es
well
with
m
u
lti
-
clas
s
p
r
o
b
lem
s
.
On
th
e
o
th
er
s
id
e,
it
ca
n
b
e
co
m
p
u
tatio
n
ally
r
a
th
er
ex
p
en
s
iv
e
o
n
lar
g
e
d
atas
ets,
is
s
en
s
itiv
e
to
o
u
tlier
s
,
an
d
r
e
q
u
ir
es
p
r
o
p
er
s
ca
lin
g
o
f
f
ea
tu
r
es
f
o
r
o
p
tim
a
l
p
er
f
o
r
m
an
ce
.
C
lass
if
icatio
n
s
o
f
d
if
f
er
en
t
KNN
class
if
ier
s
b
a
s
ed
o
n
th
e
n
u
m
b
e
r
o
f
n
ei
g
h
b
o
r
s
(
K)
an
d
d
is
tan
c
e
f
u
n
ctio
n
u
s
ed
.
T
h
e
f
in
e
KNN
class
if
ier
u
s
es
1
n
eig
h
b
o
r
with
th
e
E
u
clid
ea
n
d
is
tan
ce
f
u
n
ctio
n
.
T
h
e
m
e
d
iu
m
KNN
class
if
ier
u
s
es
1
0
n
eig
h
b
o
r
s
,
a
ls
o
with
th
e
E
u
clid
ea
n
d
is
tan
ce
f
u
n
ctio
n
.
T
h
e
c
o
ar
s
e
KNN
class
if
ier
u
s
es
1
0
0
n
eig
h
b
o
r
s
with
th
e
s
am
e
d
is
tan
ce
f
u
n
ctio
n
.
T
h
e
co
s
in
e
KN
N
class
if
ier
m
ak
es
u
s
e
o
f
1
0
n
eig
h
b
o
r
s
with
th
e
co
s
in
e
d
is
tan
ce
f
u
n
ctio
n
,
an
d
th
e
cu
b
ic
KNN
clas
s
if
ier
u
s
es
1
0
n
eig
h
b
o
r
s
with
th
e
Min
k
o
wsk
i
d
is
tan
ce
f
u
n
ctio
n
.
Fin
ally
,
t
h
e
weig
h
ted
KNN
class
if
ier
u
s
es
1
0
n
eig
h
b
o
r
s
with
a
weig
h
ted
E
u
clid
ea
n
d
is
tan
ce
f
u
n
ctio
n
b
y
1
/d
².
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
1
,
Ma
r
ch
20
2
6
:
28
7
-
30
1
294
3.
4
.
Su
pp
o
rt
v
ec
t
o
r
ma
chine
SVM
is
a
p
o
wer
f
u
l
s
u
p
er
v
is
e
d
lear
n
in
g
alg
o
r
ith
m
u
s
ed
f
o
r
b
o
th
class
if
icatio
n
an
d
r
e
g
r
ess
io
n
task
s
.
SVM
aim
s
to
f
in
d
th
e
o
p
tim
al
h
y
p
er
p
lan
e
th
at
b
est
s
ep
ar
ates
th
e
d
ata
in
to
d
if
f
er
en
t
class
es.
T
h
e
d
etailed
wo
r
k
in
g
a
n
d
ch
a
r
ac
ter
is
tics
o
f
SVM
ar
e
as
f
o
llo
ws:
SVM
s
f
in
d
an
o
p
tim
al
h
y
p
er
p
lan
e
t
o
class
if
y
d
a
ta,
u
s
in
g
k
er
n
el
f
u
n
ctio
n
s
f
o
r
n
o
n
-
lin
ea
r
s
ep
ar
ab
ilit
y
,
an
d
p
er
f
o
r
m
ef
f
ec
tiv
ely
in
h
ig
h
-
d
im
en
s
io
n
al
s
p
ac
es
with
r
o
b
u
s
t
g
en
er
aliza
tio
n
.
SVM
ca
n
b
e
s
en
s
itiv
e
to
n
o
is
y
d
ata,
r
eq
u
ir
e
s
ca
r
ef
u
l
tu
n
in
g
o
f
h
y
p
er
p
ar
a
m
eter
s
,
an
d
ca
n
b
e
co
m
p
u
tatio
n
ally
ex
p
e
n
s
iv
e
f
o
r
lar
g
e
d
atasets
[
5
0
]
.
T
h
e
class
if
icatio
n
o
f
SVM
class
if
ier
s
b
ased
o
n
s
p
ee
d
,
m
e
m
o
r
y
u
s
ag
e
f
o
r
b
in
ar
y
an
d
m
u
lti
-
class
class
if
icatio
n
,
an
d
f
lex
ib
ilit
y
.
T
h
e
lin
ea
r
SVM
class
if
ier
is
f
ast,
with
m
ed
iu
m
m
em
o
r
y
u
s
ag
e
f
o
r
b
o
th
b
in
ar
y
an
d
m
u
lti
-
class
task
s
,
b
u
t
o
f
f
er
s
lo
w
f
lex
ib
ilit
y
.
B
o
th
th
e
c
u
b
ic
an
d
q
u
ad
r
atic
SVM
clas
s
if
ier
s
ar
e
also
f
ast,
with
m
ed
iu
m
m
em
o
r
y
u
s
ag
e
f
o
r
b
in
ar
y
class
if
icatio
n
an
d
la
r
g
e
m
em
o
r
y
u
s
ag
e
f
o
r
m
u
lti
-
c
lass
clas
s
if
icatio
n
,
p
r
o
v
id
i
n
g
m
ed
iu
m
f
lex
i
b
ilit
y
.
T
h
e
f
in
e
g
au
s
s
ian
SVM
clas
s
if
ier
m
ain
tain
s
f
ast
s
p
ee
d
an
d
m
ed
i
u
m
m
em
o
r
y
u
s
ag
e
f
o
r
b
in
ar
y
task
s
,
with
lar
g
e
m
em
o
r
y
u
s
ag
e
f
o
r
m
u
lti
-
cl
ass
task
s
,
an
d
o
f
f
er
s
h
ig
h
.
3.
5
.
E
ns
em
ble
cl
a
s
s
if
iers
E
n
s
em
b
le
lear
n
in
g
is
a
p
o
we
r
f
u
l
m
eth
o
d
o
f
ML
wh
er
e
s
e
v
er
al
m
o
d
els
ar
e
c
o
m
b
in
e
d
t
o
g
eth
er
in
o
r
d
er
to
h
av
e
a
m
o
r
e
ac
cu
r
ate
an
d
r
eliab
le
p
r
ed
ictiv
e
m
o
d
el
.
I
n
h
er
e
n
t
in
th
e
u
s
e
o
f
d
if
f
e
r
en
t
m
o
d
els
is
th
eir
co
llectiv
e
k
n
o
wled
g
e
to
im
p
r
o
v
e
g
en
er
al
p
e
r
f
o
r
m
an
ce
an
d
r
e
d
u
ce
th
e
r
is
k
o
f
o
v
er
f
itti
n
g
.
Ke
y
e
n
s
em
b
le
m
eth
o
d
s
in
cl
u
d
e
b
ag
g
i
n
g
an
d
b
o
o
s
tin
g
.
T
h
e
b
ag
g
in
g
/b
o
o
ts
tr
ap
ag
g
r
eg
atin
g
tech
n
iq
u
e
is
b
ased
o
n
tr
ain
in
g
s
ev
er
al
m
o
d
els
i
n
d
i
f
f
er
en
t
s
u
b
s
ets
o
f
th
e
tr
ain
i
n
g
s
et,
cr
e
ated
b
y
b
o
o
ts
tr
ap
p
in
g
,
a
n
d
th
en
ag
g
r
eg
atin
g
th
e
p
r
ed
ictio
n
s
o
f
th
ese
m
o
d
els
th
r
o
u
g
h
av
er
ag
e
o
r
m
ajo
r
ity
v
o
te.
B
ag
g
in
g
r
ed
u
ce
s
v
ar
ian
ce
an
d
m
ak
es
th
e
m
o
d
els
m
o
r
e
s
tab
le.
B
o
o
s
tin
g
,
o
n
t
h
e
o
t
h
er
h
a
n
d
,
d
ea
ls
with
th
e
s
eq
u
e
n
tial
tr
ain
in
g
o
f
wea
k
m
o
d
els
s
o
th
at
ev
er
y
s
u
b
s
eq
u
e
n
t
m
o
d
el
g
iv
e
s
m
o
r
e
weig
h
t
to
th
e
in
s
tan
c
es
m
is
clas
s
if
ied
b
y
p
r
ev
io
u
s
m
o
d
els.
T
h
e
f
in
a
l
p
r
ed
ictio
n
is
a
weig
h
ted
s
u
m
o
f
th
e
i
n
d
iv
id
u
al
m
o
d
els
th
en
,
g
iv
in
g
m
o
r
e
weig
h
t
to
th
o
s
e
p
er
f
o
r
m
in
g
well
o
n
h
ar
d
ex
a
m
p
les.
E
n
s
em
b
le
m
eth
o
d
s
g
en
e
r
ally
p
r
o
v
id
e
b
et
ter
p
er
f
o
r
m
an
ce
t
h
an
in
d
iv
id
u
al
m
o
d
els,
r
e
d
u
ce
o
v
er
f
itti
n
g
,
a
n
d
in
c
r
ea
s
e
m
o
d
el
r
o
b
u
s
tn
ess
[
5
1
]
,
[
5
2
]
.
E
n
s
e
m
b
le
m
eth
o
d
s
ca
n
b
e
co
m
p
u
t
atio
n
ally
in
ten
s
iv
e,
co
m
p
lex
to
im
p
lem
en
t,
a
n
d
h
a
r
d
er
to
i
n
ter
p
r
et
th
a
n
s
in
g
le
m
o
d
els.
T
h
e
class
if
icatio
n
o
f
v
a
r
io
u
s
en
s
em
b
le
class
if
ier
s
is
as
f
o
llo
ws.
T
h
e
b
o
o
s
ted
class
if
ier
u
s
es
th
e
Ad
aBo
o
s
t
en
s
em
b
le
m
eth
o
d
with
DT
s
as
wea
k
lear
n
er
s
,
o
f
f
er
in
g
h
ig
h
f
lex
i
b
ilit
y
.
T
h
e
b
ag
g
in
g
class
if
ier
em
p
lo
y
s
th
e
r
an
d
o
m
f
o
r
es
t
m
eth
o
d
,
also
with
DT
s
as
wea
k
lear
n
er
s
,
an
d
p
r
o
v
id
es
h
ig
h
f
lex
ib
ilit
y
.
T
h
e
s
u
b
s
p
ac
e
d
is
cr
im
in
an
t
class
if
ier
u
s
es
th
e
s
u
b
s
p
ac
e
m
et
h
o
d
with
d
is
cr
im
in
an
t
an
aly
s
is
as
wea
k
lear
n
er
s
,
p
r
o
v
id
i
n
g
m
ed
iu
m
f
lex
ib
ilit
y
.
L
astl
y
,
th
e
s
u
b
s
p
ac
e
class
if
ier
u
tili
ze
s
th
e
Su
b
s
p
ac
e
m
et
h
o
d
with
KNN
as
wea
k
lear
n
er
s
,
o
f
f
e
r
in
g
m
ed
iu
m
f
lex
ib
ilit
y
.
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
NS
T
h
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
p
r
o
p
o
s
ed
DT
,
KN
N,
SVM,
a
n
d
en
s
em
b
le
-
b
ased
ML
cla
s
s
if
ier
s
i
s
th
o
r
o
u
g
h
ly
ev
alu
ated
u
n
d
er
n
o
n
-
id
ea
l
ch
an
n
el
co
n
d
itio
n
s
,
s
im
u
latin
g
r
ea
l
-
wo
r
ld
s
ce
n
ar
io
s
with
s
ig
n
als
s
u
b
jecte
d
to
n
o
is
e
an
d
c
h
an
n
e
l
im
p
air
m
en
ts
.
T
h
e
m
o
d
u
latio
n
s
ch
em
es
co
n
s
id
er
e
d
f
o
r
th
e
s
im
u
latio
n
s
in
clu
d
e
M
-
a
r
y
PS
K
(
M=
2
,
4
,
an
d
8
)
,
4
-
QAM
,
1
6
-
QAM
,
an
d
6
4
-
QAM
,
cr
ea
tin
g
a
co
m
p
r
eh
en
s
iv
e
test
b
ed
f
o
r
ass
es
s
in
g
th
e
class
if
ier
s
’
ef
f
ec
tiv
en
ess
u
n
d
er
d
iv
er
s
e
co
n
d
itio
n
s
.
T
h
e
m
o
d
u
latio
n
class
if
icatio
n
d
ataset
is
well
-
d
esig
n
ed
to
in
clu
d
e
a
wid
e
v
ar
iety
o
f
SNR
v
alu
es
r
an
g
i
n
g
f
r
o
m
-
3
0
d
B
to
+3
0
d
B
.
T
h
is
r
an
g
e
co
v
er
s
v
er
y
n
o
is
y
co
n
d
itio
n
s
an
d
id
ea
l
o
n
es,
h
en
ce
p
r
o
d
u
cin
g
a
s
o
lid
b
ase
f
o
r
b
o
th
th
e
tr
ain
in
g
an
d
test
in
g
o
f
m
ac
h
in
e
lear
n
in
g
m
o
d
els.
T
h
i
s
en
s
u
r
es
th
at
th
e
s
am
p
led
SN
R
lev
els
ar
e
with
in
th
e
–
3
0
d
B
to
+3
0
d
B
in
ter
v
al
w
ith
a
co
n
s
tan
t
5
d
B
s
tep
s
ize.
Hav
in
g
a
c
o
n
s
tan
t
s
tep
s
ize
o
f
5
d
B
,
th
e
d
ataset
in
clu
d
es
an
eq
u
ab
ly
d
is
tr
ib
u
ted
n
u
m
b
er
o
f
s
am
p
les
o
n
a
u
s
ad
o
s
ca
le
o
f
t
h
e
SNR
.
T
h
e
d
ata
s
et
in
clu
d
es
n
eg
ativ
e
d
B
v
alu
es
co
r
r
esp
o
n
d
in
g
to
h
ig
h
ly
n
o
is
y
o
r
s
ig
n
if
ican
tly
in
ter
f
er
ed
s
ce
n
ar
io
s
.
Sp
ec
ial
atten
tio
n
is
g
iv
en
to
ac
cu
r
a
tely
s
im
u
late
s
u
ch
co
n
d
itio
n
s
.
All sig
n
als in
th
e
s
et
o
f
d
ata
ar
e
r
ep
r
esen
ted
b
y
1
0
,
0
0
0
s
am
p
les.
T
h
e
lar
g
e
s
am
p
le
s
ize
is
ch
o
s
en
to
p
r
o
v
id
e
ad
e
q
u
ate
d
ata
f
o
r
th
e
m
ac
h
in
e
lea
r
n
in
g
m
o
d
e
ls
to
lear
n
th
e
ch
ar
ac
ter
is
tics
o
f
ev
er
y
m
o
d
u
latio
n
ty
p
e
ef
f
ec
tiv
ely
.
T
h
e
s
ig
n
als
wer
e
g
en
er
ated
ac
r
o
s
s
d
if
f
er
en
t
SNR
s
an
d
m
u
ltip
ath
f
ad
in
g
ch
an
n
els
to
m
o
d
el
r
ea
lis
tic
co
m
m
u
n
i
ca
tio
n
en
v
ir
o
n
m
en
ts
.
T
h
is
d
iv
er
s
ity
in
th
e
d
ataset
aid
s
in
b
u
ild
i
n
g
r
o
b
u
s
t
m
o
d
e
ls
th
at
g
en
er
alize
well
ac
r
o
s
s
d
if
f
er
e
n
t
co
n
d
itio
n
s
.
I
t
m
ea
n
s
th
at
th
e
m
ac
h
i
n
e
lear
n
in
g
m
o
d
els
to
b
e
tr
ain
ed
b
ased
o
n
th
is
d
ataset
will
b
e
ap
p
licab
le
an
d
ef
f
ec
tiv
e
ag
ain
s
t
all
ty
p
es
o
f
r
ea
l
-
life
s
itu
atio
n
s
,
in
clu
d
in
g
th
o
s
e
with
lar
g
e
n
o
is
e
an
d
i
n
ter
f
er
e
n
ce
.
I
n
th
is
p
ap
er
,
cr
o
s
s
-
v
ali
d
atio
n
h
as
b
ee
n
u
s
ed
in
o
r
d
e
r
to
en
s
u
r
e
a
r
o
b
u
s
t e
v
a
lu
atio
n
an
d
v
alid
atio
n
f
o
r
ML
m
o
d
els.
Fig
u
r
e
3
s
h
o
ws
th
e
p
r
o
ce
s
s
o
f
cr
o
s
s
-
v
alid
atio
n
.
Su
c
h
m
eth
o
d
s
p
r
o
v
i
d
e
f
u
ll
in
s
ig
h
t
in
to
t
h
e
p
er
f
o
r
m
an
ce
o
f
th
e
m
o
d
el
an
d
p
r
e
v
en
t
o
v
er
f
itti
n
g
.
C
r
o
s
s
-
v
alid
atio
n
,
o
n
th
e
o
th
er
h
an
d
,
is
m
o
r
e
r
ig
o
r
o
u
s
,
en
tailin
g
th
e
d
iv
is
io
n
o
f
d
ata
i
n
to
a
n
u
m
b
er
o
f
f
o
ld
s
a
n
d
iter
ativ
ely
tr
ain
in
g
an
d
test
in
g
a
m
o
d
el
o
n
d
if
f
er
e
n
t
s
u
b
s
ets o
f
d
ata.
T
h
is
g
iv
es a
m
o
r
e
r
ea
lis
tic
m
o
d
el
p
e
r
f
o
r
m
an
ce
esti
m
ate.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
I
SS
N:
2252
-
8
7
7
6
Dev
elo
p
men
t o
f m
a
ch
i
n
e
lea
r
n
in
g
tech
n
iq
u
es fo
r
a
u
t
o
ma
tic
mo
d
u
la
tio
n
…
(
P
.
G.
V
a
r
n
a
K
u
ma
r
R
ed
d
y
)
295
Fig
u
r
e
3
.
C
r
o
s
s
v
ali
d
atio
n
p
r
o
ce
s
s
T
h
e
d
ataset
is
d
iv
id
ed
in
to
K
eq
u
ally
s
ized
f
o
ld
s
(
co
m
m
o
n
l
y
5
o
r
1
0
)
.
E
ac
h
f
o
ld
s
er
v
es
a
s
a
test
in
g
s
et
o
n
ce
,
wh
ile
th
e
r
em
ain
in
g
K
-
1
f
o
ld
s
ar
e
c
o
m
b
in
e
d
to
f
o
r
m
th
e
tr
ain
in
g
s
et
[
4
9
]
.
T
h
e
m
o
d
el
is
tr
ain
ed
an
d
test
ed
K
tim
es,
ea
ch
tim
e
u
s
in
g
a
d
if
f
er
e
n
t
f
o
ld
as
th
e
test
in
g
s
et.
T
h
e
r
esu
lts
ar
e
t
h
en
av
e
r
ag
ed
t
o
p
r
o
v
i
d
e
a
co
m
p
r
eh
e
n
s
iv
e
ev
alu
atio
n
m
et
r
ic.
T
ab
le
5
p
r
esen
ts
th
e
p
er
f
o
r
m
an
ce
o
f
ML
class
if
ier
s
at
1
0
d
B
SNR
,
lin
ea
r
SVM
,
an
d
b
a
g
g
ed
tr
e
es
ac
h
iev
ed
g
o
o
d
ac
cu
r
ac
y
th
an
o
th
er
s
in
AM
C
,
ac
h
iev
in
g
th
e
h
ig
h
est
ac
cu
r
ac
ies
ac
r
o
s
s
b
o
t
h
5
-
f
o
l
d
an
d
1
0
-
f
o
ld
cr
o
s
s
-
v
alid
atio
n
.
Fin
e
t
r
ee
an
d
weig
h
ted
KNN
also
s
h
o
w
ed
co
n
s
is
ten
t
p
er
f
o
r
m
a
n
ce
,
wh
ile
b
o
o
s
ted
tr
ee
s
p
r
o
v
id
e
d
s
o
lid
p
er
f
o
r
m
an
ce
,
p
a
r
ticu
lar
ly
in
5
-
f
o
ld
C
V.
Ov
er
all,
en
s
em
b
le
m
eth
o
d
s
an
d
SVMs
p
r
o
v
ed
h
ig
h
l
y
ef
f
ec
tiv
e.
T
ab
le
5
.
Per
f
o
r
m
an
ce
o
f
ML
c
lass
if
ier
s
at
1
0
d
B
SNR
M
L
c
l
a
ssi
f
i
e
r
H
y
p
e
r
p
a
r
a
me
t
e
r
%
o
f
a
c
c
u
r
a
c
y
5
-
f
o
l
d
C
V
10
-
f
o
l
d
C
V
DT
FT
9
6
.
7
9
5
.
8
MT
9
4
.
7
9
3
.
9
CT
9
4
.
7
9
4
.
6
S
V
M
Li
n
e
a
r
9
9
.
8
9
9
.
5
Q
u
a
d
r
a
t
i
c
9
6
.
6
9
7
.
3
C
u
b
i
c
9
4
.
3
9
4
.
4
F
i
n
e
g
a
u
ssi
a
n
9
2
.
3
9
2
.
9
M
e
d
i
u
m
g
a
u
ss
i
a
n
9
2
.
1
9
2
.
4
C
o
a
r
se
g
a
u
ss
i
a
n
9
2
.
3
9
3
.
1
K
N
N
F
i
n
e
9
7
.
3
9
3
.
5
M
e
d
i
u
m
9
1
.
6
9
1
.
5
C
o
a
r
se
9
4
.
5
9
5
7
C
o
s
i
n
e
9
3
.
8
9
4
.
2
C
u
b
i
c
9
6
.
3
9
5
.
8
W
e
i
g
h
t
e
d
9
7
.
2
9
6
.
9
En
se
mb
l
e
c
l
a
ssi
f
i
e
r
s
B
o
o
st
e
d
t
r
e
e
s
9
7
.
3
9
5
.
9
B
a
g
g
e
d
t
r
e
e
s
9
7
.
8
9
7
.
4
S
u
b
s
p
a
c
e
K
N
N
9
4
.
6
9
4
.
1
S
u
b
s
p
a
c
e
d
i
scri
mi
n
a
n
t
9
1
.
3
9
2
.
8
R
u
s
b
o
o
s
t
e
d
t
r
e
e
s
8
9
.
6
8
9
.
5
Fig
u
r
e
4
p
r
esen
ts
th
e
p
er
f
o
r
m
an
ce
o
f
d
if
f
er
en
t
DT
class
if
ier
s
u
n
d
er
v
ar
io
u
s
SNR
s
.
I
t
is
o
b
s
er
v
ed
th
at
f
in
e
tr
ee
o
u
t
p
er
f
o
r
m
e
d
o
th
e
r
class
if
ier
s
with
m
o
r
e
th
an
3
%
ac
cu
r
ac
y
at
all
SNR
s
.
Fu
r
th
er
is
o
b
s
er
v
ed
th
at
ev
en
at
0
d
B
SNR
f
in
e
tr
ee
c
lass
if
ier
ac
h
iev
ed
an
ac
cu
r
ac
y
o
f
8
4
.
2
%
an
d
at
1
0
d
B
SNR
it
i
s
ac
h
iev
ed
an
ac
cu
r
ac
y
o
f
9
6
.
7
% with
5
-
f
o
ld
C
V.
Simi
lar
ly
Fig
u
r
e
5
p
r
esen
ts
th
e
p
er
f
o
r
m
a
n
ce
o
f
v
ar
i
o
u
s
KNN
clas
s
if
ier
s
at
d
if
f
e
re
n
t
SNR
s
.
Fro
m
Fig
u
r
e
5
,
it
is
o
b
s
er
v
ed
th
at
f
in
e
KN
N
an
d
weig
h
ted
KNN
ac
h
iev
ed
an
ac
c
u
r
ac
y
o
f
m
o
r
e
th
an
8
0
% a
n
d
at
1
0
d
B
,
t
h
ese
class
if
ier
s
ar
e
ac
h
iev
ed
an
ac
cu
r
ac
y
o
f
9
7
.
3
% a
n
d
9
7
.
2
% r
esp
ec
tiv
ely
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
7
7
6
I
n
t J I
n
f
&
C
o
m
m
u
n
T
ec
h
n
o
l
,
Vo
l.
1
5
,
No
.
1
,
Ma
r
ch
20
2
6
:
28
7
-
30
1
296
Fig
u
r
e
4
.
Per
f
o
r
m
an
c
e
o
f
DT
s
at
v
ar
io
u
s
SNR
s
Fig
u
r
e
5
.
Per
f
o
r
m
an
c
e
o
f
KN
Ns at
v
ar
io
u
s
SNR
s
Fig
u
r
e
6
p
r
esen
ts
th
e
p
er
f
o
r
m
an
ce
o
f
v
ar
io
u
s
SVM
cla
s
s
if
ie
r
s
u
n
d
er
d
if
f
e
r
en
t
SNR
s
.
Fro
m
Fig
u
r
e
6
,
it
is
o
b
s
er
v
ed
th
at
SVM
cla
s
s
i
f
ier
with
lin
ea
r
k
er
n
el
an
d
co
u
r
s
e
g
au
s
s
ian
k
er
n
el
o
u
tp
er
f
o
r
m
ed
all
o
th
er
k
er
n
el
f
u
n
ctio
n
s
.
I
t
is
o
b
s
er
v
e
d
th
at
at
0
d
B
SNR
,
lin
ea
r
SVM
an
d
co
u
r
s
e
g
a
u
s
s
ian
SVM
class
if
ier
s
ac
h
iev
ed
an
ac
cu
r
ac
y
o
f
8
4
.
3
%
an
d
8
4
.
1
%
r
esp
ec
tiv
ely
.
Fu
r
t
h
er
,
it
is
o
b
s
er
v
e
d
th
at
th
ese
class
if
ier
s
ar
e
ac
h
ie
v
ed
an
ac
cu
r
ac
y
o
f
9
9
.
8
% a
n
d
9
9
.
6
% a
t 1
0
d
B
SNR
.
Fig
u
r
e
7
p
r
esen
ts
th
e
p
er
f
o
r
m
an
ce
o
f
v
ar
io
u
s
en
s
em
b
le
class
if
ier
s
u
n
d
er
d
if
f
er
e
n
t
SNR
s
.
Fro
m
Fig
u
r
e
7
,
it
is
o
b
s
er
v
ed
th
at
b
ag
g
e
d
tr
ess
an
d
b
o
o
s
ted
tr
ess
o
u
tp
er
f
o
r
m
e
d
all
o
th
e
r
cl
ass
if
ier
s
.
I
t
is
also
o
b
s
er
v
ed
t
h
at
at
0
d
B
SNR
,
b
ag
g
ed
tr
ee
s
a
n
d
b
o
o
s
ted
tr
ee
s
c
lass
if
ier
s
ar
e
ac
h
iev
ed
an
a
cc
u
r
ac
y
o
f
7
9
.
3
%
an
d
7
8
.
7
%
r
esp
ec
tiv
ely
.
Fu
r
th
er
,
it
is
o
b
s
er
v
ed
th
at
t
h
ese
class
if
ier
s
ar
e
ac
h
iev
e
d
an
ac
cu
r
ac
y
o
f
9
7
.
4
%
a
n
d
9
7
.
1
% a
t 1
0
d
B
SNR
.
T
ab
le
6
p
r
esen
ts
th
e
p
er
f
o
r
m
an
ce
co
m
p
a
r
is
o
n
o
f
p
r
o
p
o
s
ed
ML
class
if
ier
s
with
o
th
er
ex
is
tin
g
m
eth
o
d
s
.
T
h
is
s
tu
d
y
co
n
f
ir
m
s
th
at
ML
tech
n
iq
u
es,
p
ar
ticu
lar
ly
SVMs
an
d
en
s
em
b
le
m
eth
o
d
s
,
ar
e
h
ig
h
ly
ef
f
ec
tiv
e
f
o
r
AM
C
task
s
,
esp
ec
ially
at
m
o
d
er
ate
to
h
ig
h
SN
R
lev
el
s
.
T
h
e
s
tr
en
g
th
o
f
lin
ea
r
SVM
an
d
b
a
g
g
ed
tr
ee
s
in
r
ea
lizin
g
h
ig
h
ac
cu
r
ac
y
ac
r
o
s
s
d
if
f
er
e
n
t
SNR
s
m
ak
es
th
em
p
r
etty
s
u
itab
le
f
o
r
r
ea
l
-
wo
r
ld
ap
p
licatio
n
s
with
h
ig
h
ly
v
a
r
y
in
g
s
ig
n
al
q
u
ality
.
Dec
is
io
n
tr
ee
s
a
r
e
r
elev
a
n
t
f
o
r
in
te
r
p
r
etatio
n
,
r
elativ
ely
ea
s
y
to
ap
p
ly
,
an
d
less
ac
cu
r
ate
th
an
SVM
s
an
d
th
e
en
s
em
b
le
m
eth
o
d
s
.
Oth
er
p
r
o
m
is
in
g
ap
p
r
o
ac
h
es
in
clu
d
e
KNN
clas
s
if
ier
s
,
esp
ec
ially
f
in
e
KNN
a
n
d
wei
g
h
ted
KNN,
w
o
r
k
in
g
r
ea
lly
w
ell,
esp
ec
ially
at
h
i
g
h
er
SNR
s
.
T
h
ese
r
esu
lts
th
u
s
s
h
o
w
th
e
p
r
o
p
e
r
u
s
e
o
f
class
if
ier
s
an
d
t
u
n
in
g
t
h
eir
h
y
p
er
-
p
a
r
am
eter
s
in
o
r
d
er
to
attain
o
p
tim
al
p
er
f
o
r
m
an
ce
in
AM
C
task
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.