T
E
L
KO
M
NI
K
A
,
V
ol
.
14,
N
o.
3,
S
ept
em
ber
20
16,
pp.
10
16
~
102
3
I
S
S
N
:
1
693
-
6
930
,
ac
c
r
edi
t
ed
A
b
y
D
IK
T
I,
D
e
c
r
e
e
N
o
:
58/
D
I
K
T
I
/
K
ep/
2013
D
O
I
:
10.
12928/
T
E
LK
O
M
N
I
K
A
.
v
1
4
i
3
.
3544
10
16
R
ec
ei
v
ed
M
ar
c
h 7
,
2
01
6
;
R
ev
i
s
ed
May
19
,
20
1
6
;
A
c
c
e
pt
ed
J
u
ne
3
,
2
01
6
Reco
g
n
it
i
o
n
o
f
F
issio
n
S
ig
n
als Based
o
n
W
a
v
e
l
et
A
n
a
l
y
sis
an
d
N
eu
r
al Ne
t
w
o
r
k
L
i
L
i
*
1
,
L
i
u
K
e
q
i
1
,
H
u
G
en
2
1
S
out
h
W
e
s
t
U
ni
v
er
s
i
t
y
of
s
c
i
e
nc
e &
t
ec
hnol
ogy
2
I
ns
t
i
t
ut
e of
N
u
c
l
ear
P
hy
s
i
c
s
a
nd
C
he
m
i
s
t
r
y
,
C
hi
n
a A
c
a
de
m
y
of
E
ng
i
ne
er
i
ng
P
hy
s
i
c
s
,
C
i
t
y
of
M
i
any
ang,
C
hi
n
a
,
t
e
l
p:
8
61388
117
3156
*
C
or
r
es
po
ndi
ng a
ut
hor
,
e
-
m
a
i
l:
l
i
l
i
_s
w
us
t
@
y
aho
o.
c
om
A
b
st
r
act
B
ec
au
s
e of
t
h
e par
t
i
c
ul
ar
i
t
y
of
t
he ur
ani
um
c
om
ponen
t
s
,
t
h
e
nonde
s
t
r
u
c
t
i
v
e m
eas
ur
i
ng t
e
c
hni
qu
e
is
ne
eded
t
o
de
t
ec
t
t
he
r
a
di
o
ac
t
i
v
i
t
y
of
t
he
c
om
pone
nt
i
n
c
er
t
ai
n
c
on
t
ai
n
er
an
d
i
de
nt
i
f
y
t
hei
r
pr
o
per
t
y
t
o
r
ec
og
ni
z
e al
l
k
i
nd
s
of
ur
ani
um
c
om
pone
nt
s
.
T
hi
s
pa
per
e
s
t
a
bl
i
s
he
s
a s
et
of
s
am
pl
e
s
w
i
t
h t
he s
am
e
s
ha
pe,
di
f
f
er
ent
w
ei
g
ht
and
ab
und
anc
e
of
ur
ani
um
by
s
i
m
ul
at
i
on.
S
ec
on
dl
y
t
he
c
r
o
s
s
-
c
or
r
e
l
at
i
on
f
unc
t
i
on
of
t
i
m
e
-
r
el
at
i
on s
i
g
nal
bet
w
een t
he s
o
ur
c
e det
e
c
t
or
an
d
t
h
e
d
et
ec
t
or
c
ou
l
d be c
al
c
ul
at
ed
.
L
as
t
l
y
t
h
e
r
es
ul
t
of
c
r
o
s
s
-
c
or
r
el
at
i
on f
un
c
t
i
ons
i
s
t
hr
ou
gh m
i
c
r
o
-
w
a
v
el
et
an
al
y
s
i
s
t
o
obt
ai
n f
eat
ur
e v
ec
t
or
w
hi
c
h
i
s
r
el
at
ed t
o t
he
qual
i
t
y
and a
bun
dan
c
e pr
oper
t
y
of
t
ar
get
ur
ani
um
c
om
pon
en
t
s
.
T
h
i
s
v
ec
t
or
i
s
u
s
ed t
o t
r
a
i
n
neur
al
net
w
or
k
and h
el
p
t
o i
dent
i
f
y
t
he
qua
l
i
t
y
and a
bund
an
c
e of
un
k
now
n
ur
ani
um
c
om
pone
nt
s
.
Ke
y
w
o
rd
s
:
m
i
c
ro
-
w
av
e
l
et
ana
l
y
s
i
s
,
neur
al
n
et
w
or
k
,
ur
a
ni
um
c
om
pone
nt
.
C
o
p
y
r
i
g
h
t
©
20
16 U
n
i
ver
si
t
a
s A
h
mad
D
ah
l
an
.
A
l
l
r
i
g
h
t
s r
eser
ved
.
1
.
I
n
tr
o
d
u
c
ti
o
n
T
he
det
ec
t
i
o
n
of
ur
an
i
um
c
om
ponent
s
i
s
an
apor
i
a
i
n
t
he
n
uc
l
ear
d
i
s
ar
m
a
m
ent
.
N
uc
l
ear
c
o
m
ponent
s
i
nc
l
u
de
ur
a
ni
um
c
o
m
ponent
s
an
d
p
l
ut
oni
um
c
o
m
ponent
s
.
B
ot
h
t
he
i
nt
ens
i
t
y
of
s
pont
a
neo
us
f
i
s
s
i
on
of
neu
t
r
ons
and
gam
m
a
-
r
a
y
i
n
m
et
al
ur
an
i
um
ar
e
w
e
ak
,
and
w
i
t
h
s
t
r
ong
l
y
s
e
lf
-
s
hi
el
d
i
ng
ef
f
ec
t
.
S
o i
t
i
s
di
f
f
i
c
ul
t
t
o
det
ec
t
m
et
al
ur
a
ni
um
t
hr
ough
pas
s
i
v
e
det
ec
t
i
ng
m
et
hods
[1
-
3]
.
S
om
e r
es
ear
c
hs
bas
ed o
n ac
t
i
v
e
i
nd
uc
i
ng
m
et
hod,
w
hi
c
h
m
a
k
es
t
he ex
og
e
nous
n
eut
r
o
ns
ent
er
ur
a
ni
um
c
o
m
ponent
s
and pr
o
duc
e s
c
at
t
er
i
ng
,
abs
or
bi
ng
and
i
n
duc
e f
i
s
s
i
ons
b
y
ac
t
i
v
e
i
nd
uc
i
ng
ur
an
i
um
c
o
m
pone
nt
s
[
4]
,
i
n
di
c
at
e
t
h
at
t
he
n
eur
ons
pr
o
duc
e
d
b
y
i
nduc
e
d
f
i
s
s
i
ons
ar
e
s
c
at
t
er
ed or
c
aus
e
nex
t
g
ener
a
t
i
o
n i
n
duc
ed
f
i
s
s
i
ons
,
m
ul
t
i
-
gener
a
t
i
o
n f
i
s
s
i
ons
c
oul
d f
or
m
a
f
i
s
s
i
on c
hai
n,
a
nd t
h
e n
eut
r
ons
on t
he s
am
e c
hai
n
ar
e
r
el
at
e
d
[5
,
6]
.
I
f
t
he r
el
at
i
on
-
s
hi
p c
ou
l
d b
e
s
a
m
pl
ed
,
des
c
r
i
be
d,
and r
egu
l
ar
,
t
he r
ec
og
ni
t
i
o
n of
n
uc
l
ear
c
om
ponent
s
c
oul
d b
e gi
v
en t
hr
o
ugh
ana
l
y
s
i
s
t
he r
e
l
at
i
o
ns
hi
p
be
t
w
een
t
he
c
har
ac
t
er
i
s
t
i
c
s
i
g
nal
of
c
o
m
ponent
s
and
t
he
i
r
f
eat
ur
es
.
B
eac
us
e
of
t
he ex
per
i
m
ent
al
da
t
a f
r
o
m
ac
t
i
v
e i
nduc
i
ng
m
et
hod ar
e d
i
s
per
s
e,
v
ar
i
e
d and
m
a
s
s
i
v
e.
T
he r
el
at
i
on
-
s
h
i
p
i
s
v
er
y
h
ar
d t
o ex
t
r
ac
t
e
d and d
es
c
r
i
bed.
T
r
adi
t
on
al
m
at
h
m
at
i
c
s
t
heor
y
i
s
har
d
l
y
t
o w
or
k
out
t
he r
e
l
at
i
o
n
-
s
hi
p f
unc
t
i
on or
des
c
r
i
be m
odel
f
or
r
ec
ogni
t
i
on
of
ur
ani
um
c
o
m
ponent
s
i
n s
h
or
t
t
i
m
e and
ef
f
ec
t
i
v
e
[1
,
6]
.
T
hi
s
paper
d
ev
ot
es
w
or
k
s
t
o bu
i
l
t
w
a
v
e
l
e
t
pac
k
et
ana
l
y
s
i
s
m
et
hod t
o
ex
t
r
ac
t
t
he f
eat
ur
e s
i
gn
al
s
of
ur
an
i
um
c
o
m
ponent
s
a
nd
us
e a
B
-
P
neur
a
l
n
et
w
or
k
t
o i
den
t
i
f
y
t
he pr
o
per
t
y
of
ur
a
ni
um
c
om
ponent
s
.
T
he m
i
c
r
o
-
w
av
el
et
an
al
y
s
i
s
i
s
v
er
y
ef
f
i
c
i
ent
t
o dea
l
w
i
t
h t
he di
s
p
er
s
e s
i
gna
l
s
and n
eur
al
n
et
w
or
k
i
s
hel
pf
ul
t
o
bui
l
d des
c
r
i
b
e
m
odel
f
or
v
ar
i
e
d an
d l
es
s
-
r
el
at
ed dat
a
.
2.
R
e
sea
r
ch
m
et
h
o
d
T
he
m
eas
ur
e
m
ent
m
odel
i
s
as
F
i
gur
e 1 b
y
us
i
ng a
c
t
i
v
e m
et
hod t
o det
ec
t
ur
ani
um
c
o
m
ponent
s
.
I
n
F
i
g
ur
e 1,
t
he t
i
m
e det
ec
t
or
1 i
s
us
e
d t
o c
apt
ur
e t
h
e s
i
g
nal
s
s
uc
h as
f
i
s
s
i
on
f
r
ag
m
ent
s
,
al
p
ha
par
t
i
c
l
es
and
et
c
,
w
hi
c
h
al
l
i
nd
uc
ed
b
y
neu
t
r
on
s
o
ur
c
e.
T
he
d
e
t
ec
t
or
2
and
3
ar
e
us
ed
t
o
o
bt
a
i
n
t
he
n
eut
r
ons
and
γ
r
a
y
s
w
hi
c
h
i
ndu
c
ed
b
y
n
eut
r
on
s
o
ur
c
e
f
r
o
m
t
he
ur
ani
um
c
o
m
ponent
s
.
T
he neu
t
r
on
s
our
c
e i
s
m
ai
nl
y
m
ade o
f
2
52
C
f
,
w
h
i
c
h gen
er
at
es
4
neut
r
ons
a
nd
6γ
phot
ons
i
n
e
v
er
y
s
p
ont
a
neou
s
f
i
s
s
i
on.
T
he
neut
r
o
n
s
and
γ
r
a
y
s
f
r
o
m
s
pont
an
eous
f
i
s
s
i
on
of
neut
r
o
n s
our
c
e br
e
ak
i
nt
o t
he ur
an
i
um
c
o
m
ponent
s
an
d pr
oduc
e c
hai
n r
eac
t
i
on.
T
he de
t
ec
t
or
1
,
2,
3
w
ou
l
d c
ap
t
ur
e t
h
e t
i
m
e
r
el
at
ed
i
m
pul
s
e s
i
gna
l
of
new
n
eut
r
ons
pr
o
duc
e
d b
y
f
i
s
s
i
on of
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
K
A
I
S
S
N
:
1
693
-
6
930
R
ec
ogn
i
t
i
on
of
F
i
s
s
i
on
S
i
gn
al
s
B
as
ed o
n W
av
el
et
A
na
l
y
s
i
s
and N
e
ur
al
N
et
w
or
k
(
Li
Li
)
1017
ur
ani
um
c
o
m
ponent
s
.
U
r
an
i
um
c
o
m
ponent
s
ar
e t
he t
ar
get
of
f
i
s
s
i
on s
i
gnal
s
f
eat
ur
e ana
l
y
s
i
s
an
d
r
ec
ogni
t
i
o
n b
as
ed o
n
B
P
n
eur
al
net
w
or
k
i
n t
hi
s
pap
er
.
F
i
gur
e
1
.
M
eas
ur
em
ent
m
o
del
s
k
et
c
h m
ap of
ac
t
i
v
e
de
t
ec
t
i
n
g ur
an
i
um
c
o
m
ponent
s
2.
1.
T
h
e
S
i
m
u
l
a
ti
o
n
o
f th
e
F
i
ssi
o
n
S
ig
n
a
ls
b
y
A
c
ti
v
e
I
n
d
u
c
i
n
g
U
r
a
n
i
u
m
C
o
m
p
o
n
en
t
s
T
o pr
ot
ec
t
t
he pr
i
v
ac
y
of
s
our
c
e dat
a,
a
l
l
dat
a of
t
he
ac
t
i
v
e de
t
ec
t
ex
p
er
i
m
ent
i
n t
hi
s
paper
ar
e c
ha
nge
d a
nd
ad
ded t
he
no
i
s
e
-
par
t
o
n t
i
m
e
-
l
i
ne.
A
c
c
or
di
ng
t
o t
he ac
t
i
v
e
f
i
s
s
i
on t
he
or
y
,
4
k
i
nds
of
c
y
l
i
nd
er
s
am
p
l
e
w
i
t
h
di
f
f
er
ent
w
e
i
g
ht
ar
e
s
e
l
ec
t
ed
as
t
h
e
ex
p
er
i
m
ent
t
a
r
get
s
,
an
d
t
h
e
w
ei
g
ht
s
uc
h
as
2
0k
g,
16k
g,
12k
g
&
10k
g.
T
he
4
c
y
l
i
n
der
s
am
pl
es
ar
e
w
i
t
h
di
f
f
er
ent
h
ei
ght
an
d
r
adi
us
.
E
ac
h s
am
pl
e has
4 t
y
p
es
c
onc
ent
r
at
i
on of
U
23
5(
0.
2
w
t
%
,
36.
0
w
t
%
,
5
0.
0w
t
%
,
93.
15
w
t
%
)
.
T
hus
t
he c
r
os
s
-
c
or
r
el
at
i
on
f
unc
t
i
o
n s
ho
ul
d
be c
al
c
u
l
at
e
d t
o
w
ar
d t
h
e 1
6 gr
ou
p d
at
a.
F
i
gur
e
2 s
h
o
w
s
t
h
e c
or
r
el
at
i
on
-
f
unc
t
i
ons
of
f
our
ur
a
ni
u
m
c
o
m
ponent
s
s
am
pl
es
w
i
t
h f
i
x
ed
m
a
s
s
of
20k
g
and
d
i
f
f
er
ent
dens
i
t
i
es
---
0
.2
w
t%
,
3
6
.0
w
t%
,
5
0
.0
w
t%
,
9
3
.1
5
w
t%
,
F
i
gur
e
3
s
ho
w
s
t
he c
or
r
e
l
at
i
on
-
f
unc
t
i
ons
of
f
our
s
am
pl
es
w
i
t
h
f
i
x
ed
d
ens
i
t
y
of
36.
0
w
t
%
and
d
i
f
f
er
ent
w
ei
ght
--
-
20k
g,
16k
g,
12k
g and
10k
g.
I
n
F
i
gur
e
2,
t
he
c
or
r
el
at
i
on
-
f
unc
t
i
ons
c
ur
v
e
of
f
i
x
ed w
ei
ght
(
20k
g)
and
d
i
f
f
er
ent
dens
i
t
i
es
(
r
es
pec
t
i
v
e
l
y
0.
2
w
t
%
,
36.
0
w
t
%
,
50.
0
w
t
%
a
nd
9
3.
15
w
t
%
)
c
ons
i
s
t
s
of
t
w
o
peak
s
.
Mor
eov
er
,
f
our
c
ur
v
es
f
r
o
m
0ns
t
o
20ns
ha
v
e
no
ob
v
i
ous
d
i
f
f
er
enc
e,
and
f
r
o
m
20ns
t
o
60ns
t
h
e
c
ur
v
es
s
pl
i
t
.
T
he r
eas
on
i
s
t
hat
n
eut
r
o
ns
i
n t
h
i
s
phas
e
w
as
gen
er
at
e
d m
ai
n
l
y
b
y
f
i
s
s
i
o
n
of
ur
ani
um
c
o
m
ponent
s
,
an
d t
he s
i
g
na
l
s
of
t
hi
s
phas
e
ar
e r
el
at
ed t
o
dens
i
t
y
of
c
om
ponent
s
.
T
he
f
eat
ur
es
of
s
i
gnal
s
al
s
o ar
e af
f
ec
t
ed
b
y
t
i
m
e.
I
t
i
s
o
bv
i
ous
t
ha
t
i
n
F
ig
ur
e 3 t
h
e f
i
r
s
t
peak
of
c
or
r
el
at
i
on
-
f
unc
t
i
o
ns
c
ur
v
e
i
s
r
el
at
e
d t
o
t
h
e
w
ei
g
ht
of
c
o
m
ponent
s
,
and
c
or
r
el
at
i
o
n
-
f
unc
t
i
ons
f
r
o
m
20ns
t
o 60ns
ar
e al
s
o r
el
at
ed t
o t
he
w
e
i
g
ht
of
c
o
m
ponent
s
a
nd af
f
ec
t
ed b
y
t
i
m
e.
S
o b
y
ana
l
y
z
i
ng t
h
e c
or
r
el
at
i
on
-
f
unc
t
i
o
ns
of
ur
ani
um
c
o
m
p
onen
t
s
,
t
he f
eat
ur
es
r
ef
l
ec
t
ed t
he
w
e
i
g
ht
and
abu
nda
nc
e of
t
ar
ge
t
c
o
ul
d
be
got
[
7,
8
].
F
i
gur
e
2
.
S
our
c
e
-
d
et
ec
t
or
c
r
os
s
-
c
or
r
el
at
i
on
f
unc
t
i
ons
f
or
ur
ani
um
c
o
m
p
onen
t
s
of
di
f
f
er
ent
abun
danc
es
and f
i
x
e
d m
as
s
(
20k
g)
F
i
gur
e
3
.
S
our
c
e
-
d
et
ec
t
or
c
r
os
s
-
c
or
r
el
at
i
on
f
unc
t
i
ons
f
or
ur
ani
um
c
o
m
p
onen
t
s
of
di
f
f
er
ent
m
a
s
s
es
and f
i
x
ed de
ns
i
t
y
(
36%
w
t
U
-
235)
2.
2.
E
xt
r
act
s
th
e
F
e
a
tu
r
e
s
o
f
F
i
ssi
o
n
S
i
g
n
al
s
B
as
ed
o
n
W
av
el
et
P
a
cket
A
n
al
ysi
s M
et
h
o
d
W
av
el
et
t
r
ans
f
or
m
i
s
a pr
oc
es
s
i
ng a
nd an
al
y
z
i
ng m
et
hod f
or
s
i
gna
l
s
of
m
ul
t
i
-
sca
l
e
s w
i
t
h
good t
i
m
e
-
f
r
equenc
y
l
oc
a
l
i
z
ed f
eat
ur
es
[
9
-
11]
,
s
o
i
t
i
s
v
er
y
s
u
i
t
a
bl
e f
or
pr
oc
es
s
i
ng nuc
l
ear
s
i
gna
l
s
of
f
i
s
s
i
on neut
r
o
n
s
w
i
t
h t
r
ans
i
en
t
an
d t
i
m
e
-
v
ar
y
i
n
g c
har
ac
t
er
i
s
t
i
c
s
.
U
s
ual
l
y
,
w
a
v
e
l
et
dec
om
pos
i
t
i
on o
nl
y
dec
om
pos
es
l
o
w
-
f
r
equenc
y
c
o
ef
f
i
c
i
ent
s
i
nt
o t
w
o par
t
s
,
w
h
i
l
e
w
av
el
e
t
pac
k
et
ana
l
y
s
i
s
f
ur
t
her
dec
om
pos
es
f
r
equenc
y
of
w
av
el
e
t
s
pa
c
e bas
ed on b
i
nar
y
s
y
s
t
em
m
anner
w
hi
c
h
m
a
k
e
s
w
a
v
e
l
et
p
ac
k
et
ov
er
c
om
e
t
he
pr
obl
em
of
hi
gh
t
i
m
e r
es
ol
ut
i
on b
ut
l
o
w
f
r
eque
nc
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
1
6
9
3
-
6
930
T
E
L
KO
M
NI
K
A
V
o
l.
14
,
N
o
.
3,
S
ept
em
ber
2016
:
10
16
–
1
023
1018
r
es
ol
ut
i
on
,
t
hus
pr
o
v
i
di
n
g m
or
e pr
ec
i
s
i
on anal
y
s
i
s
m
e
t
hod f
or
s
i
gna
l
s
.
F
i
gur
e
4 i
s
t
he t
hr
ee
-
l
ev
el
w
av
el
e
t
pac
k
et
dec
om
pos
i
t
i
on s
t
r
uc
t
ur
e
di
a
gr
am
.
F
i
gur
e
4
.
T
hr
ee
-
l
e
v
e
l
w
av
el
et
pac
k
et
dec
om
pos
i
t
i
o
n s
t
r
uc
t
ur
e d
i
agr
am
A
s
i
t
c
oul
d
be
s
ee
n
f
r
o
m
t
he
w
av
el
et
pac
k
et
dec
om
p
os
i
t
i
on
s
t
r
uc
t
ur
e
di
a
gr
am
,
w
av
el
e
t
pac
k
et
dec
om
pos
i
t
i
on c
om
par
t
m
ent
al
i
z
es
f
r
eque
nc
y
-
ban
d t
o m
ul
t
i
-
l
e
v
e
l
s
and f
ur
t
he
r
dec
om
pos
es
hi
gh
-
f
r
eque
nc
y
w
h
i
c
h
i
s
no
s
u
bs
ec
t
i
on
i
n
w
av
el
et
d
ec
om
pos
i
t
i
on
.
W
av
e
l
et
pac
k
et
dec
om
pos
i
t
i
on c
ou
l
d s
el
f
-
a
dapt
i
ng c
ho
os
e hom
ol
og
ou
s
f
r
equenc
y
-
b
and
i
n ac
c
or
d
anc
e
w
i
t
h t
he
f
eat
ur
e of
anal
y
z
e
d s
i
g
na
l
s
and m
a
k
e i
t
m
at
c
h w
i
t
h f
r
equenc
y
s
pec
t
r
um
of
t
he s
i
gn
al
s
t
o
i
m
pr
ov
e t
he
r
es
ol
ut
i
ons
of
t
i
m
e do
m
ai
n and
f
r
equenc
y
dom
ai
n
[
12,
15
]
.
2.
3.
F
e
at
u
r
e
E
x
tr
a
c
ti
o
n
S
t
ep
s B
a
s
e
d
o
n
W
av
el
et
P
a
cket
A
n
a
ly
s
is
A
f
t
er
goi
n
g
t
hr
ou
gh
de
t
ec
t
e
d
s
y
s
t
em
,
t
he
out
put
s
i
g
na
l
s
of
di
f
f
er
ent
c
o
m
ponent
s
di
f
f
er
i
n
ener
g
y
s
pac
e d
i
s
t
r
i
b
ut
i
on,
i
n ot
her
w
or
ds
,
t
h
e t
r
ans
f
or
m
at
i
on of
out
put
ener
g
y
i
nc
l
ud
es
t
he
f
eat
ur
e i
nf
or
m
at
i
on
of
t
he
c
o
m
ponent
s
.
T
her
ef
or
e,
ex
t
r
ac
t
i
ng t
h
e f
eat
ur
es
of
c
o
m
ponent
s
f
r
o
m
s
i
gna
l
s
ener
g
y
i
n e
v
er
y
s
u
bs
pac
e di
s
t
r
i
but
i
on
,
nam
el
y
us
i
n
g w
av
el
et
pac
k
et
t
r
ans
f
or
m
at
i
on t
o
ana
l
y
z
e
s
i
gna
l
s
i
n
d
i
f
f
er
ent
f
r
equenc
y
-
ba
nd
of
m
ul
t
i
-
l
e
v
e
l
dec
om
pos
i
t
i
o
n
c
oul
d
m
ak
e
unobv
i
ous
s
i
gna
l
s
f
r
equenc
y
f
eat
ur
es
v
i
s
i
bl
e
w
i
t
h m
ar
k
ed ener
g
y
-
v
ar
y
i
ng f
or
m
i
n
m
an
y
s
ubs
pac
es
of
di
f
f
er
ent
r
es
ol
ut
i
o
ns
.
T
he f
ol
l
o
w
i
n
g i
s
bas
i
c
s
t
e
ps
of
t
hi
s
m
et
hod
[1
3
-
1
4]
,
[1
6
].
St
e
p
1
:
D
ec
om
pos
e s
a
m
pl
i
ng s
i
g
nal
s
bas
ed
on
w
av
el
et
pac
k
et
dec
om
pos
i
t
i
on
.
T
he
dec
o
m
pos
i
t
i
on
l
a
y
er
s
ar
e
i
n
l
i
ne
w
i
t
h
t
he
c
om
pl
ex
i
t
y
de
gr
ee
of
s
i
g
nal
s
.
E
x
t
r
ac
t
s
i
g
nal
f
eat
ur
es
of
ev
er
y
f
r
equ
enc
y
e
l
em
ent
s
i
n t
he l
as
t
l
a
y
er
.
I
t
ado
pt
s
t
h
r
ee
-
l
a
y
er
dec
om
pos
i
t
i
oni
n t
hi
s
pap
er
,
and
ex
t
r
ac
t
s
f
r
equenc
y
e
l
em
ent
s
of
t
he t
hi
r
d
l
a
y
er
r
es
p
ec
t
i
v
el
y
f
r
o
m
lo
w
-
f
r
equenc
y
t
o h
i
gh
-
f
r
eque
nc
y
.
F
i
gur
e
4 i
s
t
h
e dec
om
pos
i
t
i
on s
t
r
uc
t
ur
e
:
i
n t
he f
i
g
ur
e,
(
i
,
j
)
r
epr
es
ent
s
t
he
j
-
t
h
n
ode of
t
he
i
-
th
l
a
y
er
w
her
e
=
,
,
,
.
=
,
,
,
,
,
,
,
,
and
r
epr
es
ent
s
t
he d
ec
om
pos
i
t
i
on c
o
e
f
f
i
c
i
ent
at
t
h
e j
-
t
h no
de of
t
h
e i
-
t
h
l
ay
er
.
S
t
ep
2:
R
es
t
r
uc
t
ur
i
n
g
w
a
v
e
l
et
pac
k
et
dec
om
pos
es
c
oef
f
i
c
i
ent
a
nd ex
t
r
ac
t
s
i
g
na
l
f
r
om
ev
er
y
f
r
equ
enc
y
b
and.
r
e
pr
es
ent
s
r
es
t
r
uc
t
ur
e
d s
i
gn
al
of
I
n t
h
i
s
pap
er
,
o
nl
y
no
des
of
t
h
e
t
hi
r
d
l
a
y
er
ar
e an
al
y
z
ed,
s
o
t
he t
ot
a
l
s
i
g
na
l
S
c
oul
d b
e
r
epr
es
ent
e
d b
y
f
or
m
ul
a (
1)
:
3
0
31
32
33
34
35
36
37
S
S
S
SS
SS
SS
=
+
+
+
+
+
+
+
(
1)
St
e
p
3
:
C
om
put
e t
he
t
ot
a
l
e
ner
g
y
of
eac
h f
r
eque
nc
y
ba
nd.
S
up
pos
e t
hat
i
s
t
he
e
ner
gy
of
r
es
t
r
uc
t
ur
ed s
i
gn
al
s
, th
e
n
:
22
1
|
(
)|
|
(
)|
n
i
j
i
j
i
j
k
E
S
t
dt
x
k
=
=
=
∑
∫
(
2)
I
n f
or
m
ul
a (
2)
,
()
i
j
xk
r
epr
es
ent
s
t
he k
-
t
h di
s
c
r
et
e poi
nt
am
pl
i
t
ud
e of
t
he r
es
t
r
uc
t
ur
ed
s
i
gna
l
at
t
he
i
-
t
h
l
a
y
er
and
t
he t
h
no
de,
and
n r
epr
es
en
t
s
t
he n
um
ber
of
di
s
c
r
et
e p
oi
nt
s
.
S
t
ep
4:
S
t
r
uc
t
ur
e f
eat
ur
e
v
e
c
t
or
.
01
(
2
1)
[
,
,
,
]
i
ii
i
T
E
E
E
−
=
(
3)
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
K
A
I
S
S
N
:
1
693
-
6
930
R
ec
ogn
i
t
i
on
of
F
i
s
s
i
on
S
i
gn
al
s
B
as
ed o
n W
av
el
et
A
na
l
y
s
i
s
and N
e
ur
al
N
et
w
or
k
(
Li
Li
)
1019
W
h
en ener
g
y
i
s
t
oo
po
w
er
f
ul
,
t
he
v
a
l
ue
of
(
2
1)
i
i
E
−
get
s
b
i
g
ger
t
oo,
w
hi
c
h
w
i
l
l
br
i
n
g
i
nc
on
v
e
ni
e
nc
e f
or
dat
a an
al
y
s
i
s
,
s
o
i
t
i
s
nee
ded
t
o u
ni
f
or
m
i
z
e
ener
g
y
.
T
he t
ot
a
l
en
er
g
y
of
s
i
gna
l
E
is
：
21
0
i
i
j
j
EE
−
=
=
∑
(
4)
'
01
(
2
1)
[
,
,
,
]
/
i
ii
i
T
E
E
E
E
−
=
(
5)
'
T
i
s
t
he
v
ec
t
or
af
t
er
uni
f
or
m
i
z
at
i
o
n
A
s
i
t
c
ou
l
db
e s
een f
r
om
f
o
r
m
ul
a (
4)
,
ener
g
y
d
i
s
t
anc
e
i
j
M
not
on
l
y
c
o
unt
s
i
n en
er
g
y
s
i
z
e
but
a
l
s
o t
he
ener
g
y
di
s
t
r
i
b
ut
i
on s
t
at
us
on t
i
m
e ax
es
---
r
epr
es
ent
i
ng
b
y
p
ar
am
et
er
t
i
n
f
or
m
ul
a.
I
n
t
hi
s
c
as
e,
c
om
par
ed
w
i
t
h f
or
m
ul
a (
1)
,
i
t
c
oul
d
be
t
t
er
pos
t
di
s
t
r
i
but
i
on f
eat
ur
e of
ener
g
y
,
t
hus
i
s
h
el
p t
o
e
x
t
r
ac
t
c
o
m
ponent
s
at
t
r
i
b
ut
i
ons
f
eat
ur
es
.
S
o f
or
m
ul
a (
5)
c
oul
d
be
r
ew
r
i
t
t
e
n as
f
or
m
ul
a (
6)
:
(
2
1)
01
(
2
1)
1
'
[,,
,
]
/
i
i
i
i
i
j
i
j
T
MM
M
M
−
−
=
=
∑
(
6)
2.
4.
F
eat
u
r
e
E
x
tr
a
c
ti
o
n
o
f
U
r
a
n
i
u
m
C
o
m
p
o
n
e
n
ts
F
i
s
s
i
o
n
S
i
g
n
a
l
s
B
a
s
e
d
o
n
W
a
v
el
et
P
a
c
k
e
t D
e
c
o
m
p
o
s
i
ti
o
n
I
n t
h
e pr
oc
es
s
of
ac
t
i
v
e
i
nduc
i
ng
ur
an
i
um
c
o
m
pone
nt
s
f
i
s
s
i
on,
f
i
s
s
i
on
s
i
g
na
l
s
r
ef
l
ec
t
abun
dan
t
i
nf
or
m
at
i
on of
ur
ani
um
c
o
m
ponent
s
at
t
r
i
but
i
ons
.
S
o
i
t
i
s
f
eas
i
bl
e t
o
g
et
at
t
r
i
but
i
o
ns
i
nf
or
m
at
i
on
of
ur
an
i
um
c
om
ponent
s
t
hr
ou
gh
an
al
y
z
i
n
g
f
i
s
s
i
on
s
i
gn
al
s
.
B
ut
w
h
en
f
i
s
s
i
on
s
i
gna
l
s
ar
e s
t
oc
has
t
i
c
an
d
non
-
b
al
a
nc
e,
i
t
i
s
d
i
f
f
i
c
ul
t
t
o
obs
er
v
e
i
nner
f
i
s
s
i
on r
u
l
es
of
ur
an
i
um
c
o
m
ponent
s
f
r
om
ex
t
er
nal
c
hanges
.
I
n
or
der
t
o f
i
nd
i
t
s
i
nt
r
i
ns
i
c
r
ul
es
,
i
t
ne
eds
us
i
ng m
oder
n
t
ec
hno
l
og
y
t
o pr
oc
es
s
pr
i
m
al
s
i
gnal
s
.
T
her
ef
or
e,
h
o
w
t
o ex
t
r
ac
t
at
t
r
i
but
i
on
f
eat
ur
e i
nf
or
m
at
i
on
f
r
o
m
f
i
s
s
i
on
s
i
gna
l
s
i
s
t
he
k
e
y
of
s
uc
c
es
s
f
ul
r
ec
o
gni
t
i
o
n
of
at
t
r
i
but
i
ons
i
n
t
he
at
t
r
i
but
i
on
s
r
ec
ogni
z
i
ng
pr
oc
es
s
.
T
he
ex
t
r
ac
t
ed
f
eat
ur
es
s
houl
d
be
s
uf
f
i
c
i
ent
s
ens
i
t
i
v
e
t
o
di
f
f
er
ent
at
t
r
i
b
ut
i
ons
,
w
h
i
c
h r
equ
i
r
es
t
he
f
eat
ur
e ex
t
r
ac
t
i
on m
et
hod c
ou
l
d ex
t
r
ac
t
at
t
r
i
but
i
o
n i
nf
or
m
at
i
on
t
hat
hi
des
i
n
s
i
g
na
l
s
.
W
av
el
et
t
r
ans
f
or
m
at
i
on
i
s
f
i
t
f
or
f
eat
ur
e
ex
t
r
ac
t
i
on
of
non
-
ba
l
anc
e
s
i
g
na
l
s
.
A
t
pr
es
ent
,
w
a
v
e
l
et
ana
l
y
s
i
s
has
b
een
ap
pl
i
ed
t
o f
eat
ur
e
ex
t
r
ac
t
i
ons
of
m
any
s
i
gna
l
s
an
d
ac
hi
e
v
e
d go
od r
es
u
l
t
s
.
I
n t
he
ac
t
i
v
e
i
nd
uc
i
ng
ur
a
ni
um
c
o
m
ponent
s
f
i
s
s
i
on,
i
t
s
ener
g
y
s
pac
e
di
s
t
r
i
bu
t
i
on
of
out
p
ut
s
i
gn
al
s
i
nc
l
ud
es
abu
ndan
t
at
t
r
i
but
i
on
f
eat
ur
e
i
nf
or
m
at
i
on.
T
her
ef
or
e,
i
t
m
a
k
es
unob
v
i
ou
s
s
i
gna
l
s
f
r
equenc
y
f
ea
t
ur
es
t
o ap
pear
w
i
t
h m
ar
k
ed ener
g
y
-
v
ar
y
i
n
g f
or
m
i
n m
an
y
s
ubs
pac
es
of
di
f
f
er
ent
r
es
ol
ut
i
ons
b
y
ex
t
r
ac
t
i
ng a
t
t
r
i
b
ut
i
on f
eat
ur
es
f
r
o
m
s
i
gnal
s
e
ner
g
y
i
n e
v
er
y
s
u
bs
pac
e
di
s
t
r
i
b
ut
i
on,
nam
el
y
a
nal
y
z
i
n
g s
i
gn
al
s
of
di
f
f
er
ent
f
r
equenc
y
ba
nd af
t
e
r
m
ul
t
i
-
l
a
ye
r
dec
om
pos
i
t
i
on
w
i
t
h
w
a
v
e
l
e
t
pac
k
et
t
r
ans
f
or
m
at
i
on,
an
d t
hen
ex
t
r
ac
t
f
eat
ur
e i
nf
or
m
at
i
on w
hi
c
h
r
ef
l
ec
t
s
ur
ani
um
c
o
m
ponen
t
s
at
t
r
i
b
ut
i
ons
.
F
ol
l
o
w
i
ng t
h
e an
al
y
s
i
s
s
t
eps
i
n
I
II
,
pr
oc
es
s
i
n
g s
o
ur
c
e
-
det
ec
t
or
’
s
c
r
os
s
-
c
or
r
el
at
i
o
n
f
unc
t
i
ons
of
f
our
c
om
ponent
s
s
a
m
pl
es
w
i
t
h f
i
x
ed m
as
s
(
20k
g)
and d
i
f
f
er
ent
dens
i
t
y
(
r
es
pec
t
i
v
e
l
y
0.
2
w
t
%
,
36.
0
w
t
%
,
50.
0
w
t
%
and 9
3.
1
5
w
t
%
)
g
et
t
h
ei
r
ener
g
y
di
s
t
anc
e d
i
s
t
r
i
b
ut
i
on d
i
agr
am
as
F
i
gur
e 5(
a
)
an
d (
b
).
P
r
oc
es
s
i
ng s
our
c
e
-
de
t
ec
t
or
'
s
c
r
os
s
-
c
or
r
el
at
i
o
n f
unc
t
i
ons
of
f
our
c
o
m
ponent
s
s
am
pl
es
w
i
t
h
di
f
f
er
ent
m
as
s
es
(
r
es
pec
t
i
v
el
y
20k
g,
16k
g,
12k
g
and
1
0k
g)
and
f
i
x
ed
dens
i
t
y
(
3
6.
0
w
t
%
)
get
t
he
i
r
ener
g
y
d
i
s
t
anc
e
di
s
t
r
i
but
i
o
n
di
a
gr
am
as
F
i
gur
e 6
(
a
)
an
d (
b
)
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
1
6
9
3
-
6
930
T
E
L
KO
M
NI
K
A
V
o
l.
14
,
N
o
.
3,
S
ept
em
ber
2016
:
10
16
–
1
023
1020
F
i
gur
e
5(
a
)
.
E
ner
g
y
d
i
s
t
anc
e di
s
t
r
i
but
i
o
n
di
a
gr
am
of
ur
ani
um
c
o
m
po
nent
s
w
i
t
h f
i
x
ed
m
a
s
s
(
20
k
g)
and d
i
f
f
er
ent
dens
i
t
i
es
(
r
es
pec
t
i
v
e
0.
2
w
t
%
2
35U
an
d 36.
0
w
t
%
2
35U
)
F
i
gur
e 5(
b
)
.
E
ner
g
y
d
i
s
t
anc
e di
s
t
r
i
but
i
o
n
di
a
gr
am
f
or
ur
ani
um
c
o
m
ponent
s
w
i
t
h f
i
x
ed
m
a
s
s
(
20
k
g)
and d
i
f
f
er
ent
dens
i
t
i
es
(
r
es
pec
t
i
v
e 50
.
0
w
t
%
235
U
a
nd
93.
1
5
w
t
%
2
35U
)
F
i
gur
e
6(
a
)
.
E
ner
g
y
d
i
s
t
anc
e di
s
t
r
i
but
i
o
n
di
a
gr
am
f
or
ur
ani
um
c
o
m
ponent
s
w
i
t
h di
f
f
er
ent
m
a
s
s
es
(
r
es
pec
t
i
v
e 2
0k
g and16k
g)
and
f
i
x
ed
dens
i
t
y
(
3
6.
0
w
t
%
)
F
i
gur
e
6(
b
)
.
E
ner
g
y
d
i
s
t
anc
e di
s
t
r
i
but
i
o
n
di
a
gr
am
f
or
ur
ani
um
c
o
m
ponent
s
w
i
t
h
di
f
f
er
ent
m
a
s
s
es
(
r
es
pec
t
i
v
e 1
2k
g and10k
g)
and
f
i
x
ed
dens
i
t
y
(
3
6.
0
w
t
%
)
T
e1,
T
e
2,
T
e3
and
T
e4
r
es
pec
t
i
v
el
y
r
e
pr
es
ent
en
er
g
y
di
s
t
a
nc
e
f
eat
ur
e
v
ec
t
or
s
of
ur
ani
um
c
o
m
ponent
s
of
di
f
f
er
ent
dens
i
t
i
es
(
r
es
pec
t
i
v
e
l
y
0
.
2
w
t
%
,
36.
0
w
t
%
,
50.
0
w
t
%
an
d
93.
1
5
w
t
%
2
35U
)
an
d
f
i
x
ed
m
a
s
s
(
20
k
g)
.
T
m
1
、
T
m
2
、
T
m
3
and
T
m
4
r
es
pec
t
i
v
e
l
y
r
epr
es
e
nt
ener
g
y
d
i
s
t
anc
e
f
eat
ur
e
v
e
c
t
or
s
of
ur
ani
um
c
o
m
ponen
t
s
of
di
f
f
er
ent
m
as
s
es
(
r
es
pec
t
i
v
el
y
2
0k
g,
16k
g,
12k
g and1
0k
g)
and f
i
x
ed de
ns
i
t
i
es
(
36.
0
w
t
%
2
35
U
)
.
2.
5
. Id
e
n
t
if
ie
s
M
asses a
n
d
A
b
u
n
d
a
n
c
e
o
f
U
r
a
n
i
u
m
C
o
m
p
o
n
e
n
ts
B
a
s
e
d
o
n
B
P
N
eu
r
a
l
N
et
w
o
r
k
I
n ac
c
or
danc
e
w
i
t
h t
op
ol
o
gi
c
al
s
t
r
uc
t
ur
e and op
er
at
i
on
m
ode of
net
w
or
k
,
neur
al
ne
t
w
or
k
m
odel
s
i
nc
l
ud
e f
eed
-
f
or
w
a
r
d m
ul
t
i
-
l
a
y
er
net
w
or
k
m
o
del
,
f
eed
bac
k
r
ec
ur
s
i
on net
w
or
k
m
odel
,
r
andom
net
w
or
k
m
odel
and et
c
.
T
he m
at
ur
e
m
od
el
i
n
pat
t
er
n
r
ec
ogn
i
t
i
on
i
s
er
r
or
B
ac
k
P
r
opa
gat
i
on
(
B
P
)
m
odel
,
w
hi
c
h
i
s
on
e
of
f
eed
-
f
or
w
ar
d
m
ul
t
i
-
l
a
y
er
n
et
w
or
k
m
odel
s
.
B
P
ne
t
w
or
k
has
not
onl
y
i
np
ut
an
d
ou
t
p
ut
no
des
,
bu
t
al
s
o
m
ak
es
o
ne
l
a
y
er
or
m
an
y
l
a
y
er
s
w
i
t
h
c
onno
t
at
i
v
e
nodes
.
F
or
i
np
ut
i
nf
or
m
at
i
on,
B
P
pr
o
pag
at
es
t
o c
o
n
not
at
i
v
e
l
a
y
er
s
n
od
es
at
f
i
r
s
t
,
and
t
h
en
pr
opa
g
at
es
t
o
o
ut
pu
t
nod
e
s
t
hr
ough
ou
t
put
i
nf
or
m
at
i
on
of
c
onnot
at
i
v
e
nod
es
,
an
d f
i
nal
l
y
g
i
v
es
out
p
ut
r
es
ul
t
s
.
N
et
w
or
k
l
ear
ni
n
g pr
oc
es
s
c
ons
i
s
t
s
of
t
w
o par
t
s
---
f
or
w
ar
d pr
op
ag
at
i
o
n an
d bac
k
pr
opa
gat
i
on.
Mas
s
es
of
s
a
m
pl
e
c
y
l
i
ndr
i
c
al
c
om
ponent
s
i
n
t
h
i
s
pap
er
ar
e
r
es
pec
t
i
v
e
2
0k
g,
16k
g,
12k
g
and
10k
g,
an
d t
h
e d
ens
i
t
i
e
s
of
eac
h d
i
f
f
er
ent
m
as
s
c
om
ponent
ar
e r
es
p
ec
t
i
v
e
0.
2
w
t
%
,
3
6.
0
w
t
%
,
50.
0
w
t
%
an
d
93.
15
w
t
%
,
w
hi
c
h
am
ount
t
o
1
6
c
om
ponent
s
s
am
pl
es
.
I
t
i
nc
l
u
des
t
w
o
s
im
u
la
t
io
n
s
:
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
K
A
I
S
S
N
:
1
693
-
6
930
R
ec
ogn
i
t
i
on
of
F
i
s
s
i
on
S
i
gn
al
s
B
as
ed o
n W
av
el
et
A
na
l
y
s
i
s
and N
e
ur
al
N
et
w
or
k
(
Li
Li
)
1021
t
he f
i
x
ed
m
as
s
w
i
t
h d
i
f
f
er
ent
dens
i
t
i
es
an
d t
he
di
f
f
er
ent
m
as
s
es
w
i
t
h f
i
x
ed
de
ns
i
t
y
.
I
t
f
i
r
s
t
l
y
pr
oc
es
s
es
and an
al
y
z
es
r
el
e
v
a
nt
s
i
m
ul
at
i
o
n da
t
a,
a
nd t
he
n pr
oc
es
s
es
w
i
t
h
w
av
e
l
et
pac
k
et
dec
om
pos
i
t
i
on
,
a
nd
f
i
nal
l
y
get
s
t
he
i
r
e
ner
g
y
di
s
t
a
n
c
e
f
eat
ur
e
v
ec
t
or
s
w
h
i
c
h
ar
e
t
ak
en
as
l
ear
n
i
n
g s
am
pl
es
of
B
P
n
eur
al
net
w
or
k
.
T
ar
get
out
p
ut
T
1[
2×
1
6]
a
nd T
2[
2×
16]
r
epr
es
ent
t
h
e
m
a
s
s
es
and U
23
5ab
und
an
c
es
of
ev
er
y
s
am
pl
e.
20 20 20 20 16 16 16 16 12 12 12 12 10 1
0 10 10;
1
0.2 36 50 93.15 0.2 36 50 93.15 0.2 36 5
0 93.15 0.2 36 50 93.15
T
=
20 16 12 10 20 16 12 10 20 16 12 10 20 1
6 12 10;
2
0.2 0.2 0.2 0.2 36 36 36 36 50 50 50 50
93.15 93.15 93.15 93.15
T
=
F
i
gur
e
7
(
a
)
.
T
r
ai
ni
ng c
ur
v
e
of
f
i
x
ed m
as
s
c
o
m
ponent
s
F
i
gur
e
7
(
b
)
.
R
ec
o
gn
i
t
i
on r
e
s
ul
t
s
of
f
i
x
ed m
as
s
c
o
m
ponent
s
F
i
gur
e
8
(
a
).
T
r
ai
ni
ng c
ur
v
e
of
f
i
x
ed abun
danc
e
c
o
m
ponent
s
F
i
gur
e
8
(
b
).
R
ec
o
gn
i
t
i
on r
e
s
ul
t
s
of
f
i
x
ed
abun
danc
e
c
om
ponent
s
A
c
c
or
di
n
g t
o K
ol
m
ogor
o
v
t
heor
em
of
neur
al
net
w
or
k
t
heor
y
,
3
-
l
a
y
er
B
P
net
w
or
k
w
i
t
h
f
ul
l
l
ear
n
i
n
g
m
odel
c
oul
da
ppr
o
x
i
m
at
e
t
o
an
y
f
unc
t
i
on.
T
hi
s
paper
c
hoos
es
3
-
l
a
y
er
B
P
net
w
or
k
,
and
nam
e a B
P
net
w
or
k
w
i
t
h s
i
ngl
e h
i
dd
en
l
a
y
er
.
I
npu
t
l
a
y
er
nod
es
of
t
he
net
w
or
k
ar
e det
er
m
i
ned
b
y
t
he
i
n
put
v
ec
t
or
d
i
m
ens
i
on
w
h
i
c
h
i
s
8
her
e,
s
o
t
he
i
nput
l
a
y
er
no
des
ar
e
8 t
o
o.
O
ut
put
l
a
y
er
nodes
ar
e
det
er
m
i
ned b
y
t
he
out
p
ut
v
ec
t
or
di
m
ens
i
o
n
w
h
i
c
h
i
s
2
her
e.
T
her
e
i
s
no
t
heor
et
i
c
al
gui
danc
e
i
n a
l
l
B
P
net
w
or
k
about
t
h
e s
e
l
ec
t
i
on
of
h
i
dd
en
nod
es
.
E
x
c
es
s
i
v
e
n
et
w
or
k
nodes
w
oul
d
ex
p
and
t
h
e t
r
a
i
ni
ng
t
i
m
e and d
i
m
i
ni
s
h g
ener
a
l
i
z
at
i
on
c
apac
i
t
y
as
w
e
l
l
as
t
he pr
edi
c
t
i
v
e
abi
l
i
t
y
of
t
he
n
et
w
or
k
,
w
h
i
l
e
i
ns
uf
f
i
c
i
ent
n
et
w
or
k
nodes
c
oul
d
not
e
r
ef
l
ec
t
t
h
e
r
el
at
i
o
ns
hi
p
bet
w
e
en t
he f
ol
l
o
w
-
up
v
a
l
u
e and t
he pr
ec
ur
s
or
v
a
l
ue,
w
hi
c
h
l
ea
ds
i
ns
uf
f
i
c
i
en
c
y
of
m
odel
i
ng.
T
he
num
ber
of
hi
dden
l
a
y
er
n
o
de i
s
s
et
a
t
am
ount
of
10 i
n t
hi
s
pa
per
af
t
er
r
epe
at
ed
t
r
a
i
ni
ng,
and
an8
-
10
-
2
B
P
neur
a
l
net
w
or
k
i
s
des
i
g
ned
f
r
om
hi
dden
l
a
y
e
r
s
.
T
he
B
P
neur
a
l
net
w
or
k
i
s
pr
ogr
am
m
ed
and t
ak
en i
nt
es
t
r
u
nni
ng b
y
t
he M
A
T
LA
B
R
201
0a.
T
he
neur
a
l
net
w
or
k
des
i
gne
d i
n t
hi
s
p
aper
t
o
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
1
6
9
3
-
6
930
T
E
L
KO
M
NI
K
A
V
o
l.
14
,
N
o
.
3,
S
ept
em
ber
2016
:
10
16
–
1
023
1022
t
r
ai
n s
i
gna
l
en
er
g
y
di
s
t
anc
e v
ec
t
or
of
f
our
ur
ani
um
c
om
ponent
s
w
i
t
h f
i
x
ed m
as
s
and di
f
f
er
ent
dens
i
t
i
es
,
of
w
hi
c
h t
r
a
i
n
i
n
g pr
oc
es
s
and r
es
ul
t
s
ar
e r
e
p
r
es
ent
ed
as
F
i
gur
e 7(
a
)
an
d (
b
)
,
and t
h
en
t
o t
r
ai
n
s
i
g
na
l
s
en
er
g
y
di
s
t
anc
e
v
ec
t
or
of
f
our
ur
an
i
u
m
c
o
m
ponent
s
w
i
t
h
d
i
f
f
er
ent
m
as
s
es
and
f
i
x
ed dens
i
t
i
es
,
w
hi
c
h
t
r
ai
ni
ng pr
oc
es
s
a
nd r
es
u
l
t
s
ar
e
r
epr
es
ent
e
d as
F
i
gur
e 8(
a
)
and (
b
).
3.
R
esu
l
t
s
a
n
d
A
n
a
l
y
s
i
s
3.
1.
T
h
e an
al
ysi
s
o
f
B
P
D
ue
t
o
par
t
i
c
ul
ar
i
t
y
,
s
ens
i
t
i
v
i
t
y
an
d
c
om
pl
ex
i
t
y
of
t
he
r
es
ear
c
h
o
bj
ec
t
s
,
t
hi
s
pape
r
c
hoos
es
a t
r
ai
n
i
n
g m
et
hod
c
om
bi
ni
ng r
eal
l
y
m
eas
ur
ed dat
a
w
i
t
h
s
om
e noi
s
e
a
nd s
i
m
ul
at
i
on
dat
a.
P
ap
er
t
he
n
c
hoos
es
one
gr
o
up
of
s
i
m
ul
at
i
on
da
t
a
w
i
t
h
no
i
s
e
t
o
t
es
t
t
h
e
ne
t
w
or
k
.
T
he
t
es
t
r
es
ul
t
of
m
as
s
es
and ab
un
danc
es
ar
e as
F
i
gur
e
7
(
b
)
and
F
i
gur
e
8(
b
).
F
r
om
t
he
ana
l
y
s
i
s
r
es
ul
t
s
in
T
ab
l
e
1 a
nd
T
abl
e
2,
i
t
i
n
di
c
at
es
t
ha
t
pr
e
di
c
t
e
d
v
a
l
u
e
bas
e
d o
n
B
P
ne
ur
al
n
et
w
o
r
k
des
i
gn
ed
i
n
t
hi
s
pap
er
i
s
c
l
os
er
t
o t
he t
r
ue v
a
l
u
e.
A
v
er
a
ge r
el
at
i
v
e
er
r
or
i
n
T
abl
e
1
is
o
n
l
y
0.
48
%
,
and 1.
96
%
in
T
abl
e
2
.
A
s
f
or
t
he
de
ns
i
t
y
pr
e
di
c
t
i
on
,
t
h
e
av
er
age
r
e
l
at
i
v
e er
r
or
i
s
r
es
pec
t
i
v
el
y
t
o 27
.
49%
an
d
25.
3
5%
.
I
t
a
l
s
o i
nd
i
c
at
es
f
r
om
t
he t
abl
e t
hat
f
or
hi
gh a
bund
anc
e ur
ani
um
,
t
he pr
e
di
c
t
ed
v
a
l
u
e of
i
t
s
abu
nda
nc
e i
s
c
l
os
e t
o t
he t
r
ue
v
al
ue,
b
ut
i
t
i
s
not
f
i
t
f
or
t
he l
o
w
a
bun
danc
e
ur
ani
um
.
T
he
r
eas
on
i
s
t
h
at
onl
y
t
he
da
t
a
of
s
i
m
ul
at
ed f
i
s
s
i
o
n f
or
ur
a
ni
um
c
o
m
ponent
s
i
s
us
e
d f
or
B
P
b
ut
l
ac
k
of
r
eal
i
t
y
dat
a t
o adj
us
t
.
I
f
enou
gh s
am
pl
e of
l
ow
den
s
i
t
y
c
ou
l
d be us
e
d i
n
B
P
,
t
he m
et
hod of
dec
r
eas
i
n
g
t
he
n
oi
s
e
an
d
i
nc
r
eas
i
ng
t
he
ac
c
ur
ac
y
f
or
l
ow
abu
nda
nc
e
ur
an
i
um
w
ou
l
d
be
dev
el
ope
d.
T
abl
e 1.
P
r
ed
i
c
t
i
on r
es
u
l
t
s
f
or
ur
ani
um
c
o
m
ponent
s
of
di
f
f
er
ent
abu
nda
nc
es
an
d f
i
x
ed m
as
s
es
t
ak
i
ng en
er
g
y
d
i
s
t
a
nc
e of
de
t
e
c
t
ed s
i
gna
l
s
as
f
eat
ur
e
v
e
c
t
or
s
K
g
/
%
wt
R
eal
K
g
/
%
wt
BP
K
g
/
%
wt
E
rro
r
1
20/
0.
2
19.
95/
-
1.
20
0.
25/
116.
6
2
20/
36.
0
20.
12/
40.
42
0.
61/
10.
94
3
20/
50.
0
20.
07/
48.
03
0.
36/
4.
11
4
20/
93.
15
20.
04/
92.
19
0.
20/
1.
04
5
16/
0.
2
15.
93/
-
2.
72
0.
44/
107.
3
6
16/
36.
0
15.
74/
34.
70
1.
66/
3.
75
7
16/
50.
0
15.
98/
50.
07
0.
12/
0.
15
8
16/
93.
15
15.
93/
93.
13
0.
43/
0.
02
T
abl
e
2.
P
r
ed
i
c
t
i
on r
es
u
l
t
s
f
or
ur
an
i
um
c
o
m
ponent
s
of
di
f
f
er
ent
m
as
s
es
and f
i
x
ed
abun
danc
es
t
ak
i
ng en
er
g
y
di
s
t
anc
e of
d
et
ec
t
ed
s
i
gn
al
s
as
f
eat
ur
e
v
ec
t
or
s
%w
t/K
g
R
eal
%w
t/K
g
BP
K
g
/
%
wt
E
rro
r
1
0.
2/
20
0.
17/
20.
03
0.
15/
20.
91
2
0.
2/
16
0.
19/
16.
10
0.
63/
8.
11
3
0.
2/
12
-
0.
13/
12.
00
0.
02/
256.
8
4
0.
2/
10
-
5.
00/
9.
20
8.
65/
104.
0
5
36.
0/
20
35.
88/
19.
93
0.
35/
0.
33
6
36.
0/
16
35.
96/
15.
91
0.
54/
0.
11
7
36.
0/
12
36.
04/
11.
91
0.
77/
0.
11
8
36.
0/
10
36.
00/
10.
00
0.
00/
0.
00
3.
2.
T
h
e
C
o
m
p
a
r
i
s
o
n
T
abl
e
3
i
s
a
c
om
par
i
s
on
be
t
w
een
t
r
ad
i
t
i
ona
l
ex
p
er
i
m
ent
al
av
er
ag
e
-
v
al
ue
(
A
V
E
)
m
et
hod
B
P
m
et
hod.
B
P
m
et
hod i
s
f
as
t
er
and m
or
e ef
f
ec
t
i
v
e.
I
f
put
t
h
e c
ons
i
der
at
i
on
on t
h
e c
os
t
s
t
o get
t
he bas
i
c
dat
a
f
r
om
t
he a
c
i
t
v
e
-
i
nd
uc
e s
y
s
t
em
(
20G
poi
nt
s
w
ou
l
d
be
r
ec
or
de
d
i
n on
e
ac
t
i
v
e
-
i
nd
uc
e)
,
B
P
m
et
hod
i
s
m
or
e v
al
u
ab
l
e a
nd s
a
v
es
a l
ot
of
m
one
y
a
nd t
i
m
e.
T
abl
e
3.
T
he c
om
par
s
i
on of
A
V
E
a
nd
B
P
S
a
m
pl
e
poi
nt
s
(
G
)
T
im
e
(
s
)
A
c
c
ur
ac
y
(%
)
E
rro
r
(%
)
BP
20G
100
92%
11%
A
VE
300G
6000
91.
6%
10.
7%
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
K
A
I
S
S
N
:
1
693
-
6
930
R
ec
ogn
i
t
i
on
of
F
i
s
s
i
on
S
i
gn
al
s
B
as
ed o
n W
av
el
et
A
na
l
y
s
i
s
and N
e
ur
al
N
et
w
or
k
(
Li
Li
)
1023
4.
C
o
n
c
l
u
s
i
o
n
T
hi
s
paper
app
l
i
es
ex
t
r
ac
t
i
on al
gor
i
t
hm
of
ener
g
y
d
i
s
t
anc
e v
ec
t
or
s
of
w
a
v
e
l
et
p
ac
k
et
dec
om
pos
i
t
i
on
an
d
B
P
neu
r
al
n
et
w
or
k
t
ec
hno
l
og
y
t
o
f
eat
ur
e
an
al
y
s
i
s
a
nd
r
ec
o
gn
i
t
i
on
f
or
ac
t
i
v
e
i
nd
uc
i
ng
ur
a
ni
um
c
o
m
ponent
s
f
i
s
s
i
on s
i
gna
l
s
o
n t
he
b
as
i
s
of
m
eas
ur
e
m
ent
pr
i
nc
i
pl
es
and
s
i
g
nal
c
har
ac
t
er
i
s
t
i
c
s
ana
l
y
s
i
s
f
r
o
m
ac
t
i
v
e
i
n
duc
i
n
g
ur
an
i
um
c
o
m
ponent
f
i
s
s
i
on
s
i
gn
al
s
.
T
he
r
es
ear
c
h
r
es
ul
t
i
nd
i
c
at
es
t
h
at
B
P
n
eur
al
net
w
or
k
c
oul
d be
t
a
k
en as
an ana
l
y
s
i
s
m
et
hod f
or
det
ec
t
i
ng
ur
ani
um
c
o
m
ponent
s
at
t
r
i
b
ut
i
o
n,
w
hi
c
h
has
hi
gh
f
eas
i
bi
l
i
t
y
i
n
t
er
m
s
of
ei
t
her
m
odel
c
ons
t
i
t
ut
i
on
or
ana
l
y
s
i
s
r
es
ul
t
s
.
T
he n
u
m
er
i
c
al
s
i
m
ul
at
i
on
r
es
u
l
t
s
c
oul
d
pr
o
v
i
d
e r
ef
er
enc
e f
or
v
er
i
f
i
c
at
i
on
of
ur
ani
um
c
o
m
ponent
at
t
r
i
b
u
t
i
on
i
n
ac
t
i
v
e
i
n
duc
i
ng
ur
a
n
i
um
c
o
m
ponent
s
,
and
a
l
s
o
of
f
e
r
t
ec
hni
c
al
s
uppor
t
f
or
f
ur
t
her
r
es
ear
c
h
.
A
c
k
n
o
w
l
e
d
g
e
m
e
n
t
s
T
hi
s
w
or
k
i
s
s
up
por
t
e
d b
y
t
he N
S
A
F
11
176
03
，
“
Meas
ur
em
ent
s
of
t
he Q
u
al
i
t
y
a
n
d t
h
e
A
bu
nda
nc
e of
U
r
an
i
um
C
om
ponent
bas
e
d o
n
T
i
m
e
-
r
el
at
ed S
i
gna
l
b
y
A
c
t
i
v
e
-
i
nduc
e
d
T
ec
hni
que”
R
ef
er
en
ces
[1
]
S
z
abo J
L,
S
im
on
A
C
,
J
unc
a
R
.
N
on
-
d
es
t
r
uc
t
i
v
e A
na
l
y
s
i
s
of
U
r
an
i
um
and
P
l
u
t
oni
um
U
s
i
ng
X
-
ra
y
F
l
uor
e
s
c
enc
e E
x
c
i
t
ed by
S
eal
ed S
our
c
e or
X
-
r
ay
T
ubes
.
N
uc
l
ear
I
ns
t
r
um
ent
s
and
M
et
hods
i
n
P
hy
s
i
c
s
R
es
ear
c
h.
1
994;
353(
1)
:
668
-
671
.
[2
]
Z
eng
K
F
,
O
n
g K
G
,
M
ugl
e C
,
G
r
i
m
e
s
C
A
. Ti
m
e
D
om
ai
n
C
har
ac
t
er
i
z
at
i
on
of
O
s
c
i
l
l
at
i
n
g S
en
s
or
:
A
ppl
i
c
at
i
on
of
F
r
e
que
nc
y
d
et
er
m
i
n
at
i
on
.
Re
v
i
e
w o
f
S
c
i
ent
i
f
i
c
I
ns
t
r
um
e
nt
s
.
200
2;
7
3(
12)
:
43
75
-
438
0
.
[3
]
O
f
al
t
A
E
,
R
our
k
e
P
E
.
D
ev
el
o
pm
en
t
a
nd
P
er
f
or
m
an
c
e
of
O
n
-
l
i
ne
U
r
ani
um
A
na
l
y
z
er
s
.
W
i
l
m
i
ngt
on
.
1985
.
[4
]
D
o
m
in
ic
V
H
ar
per
,
J
as
o
n B
,
R
oger
W
F
l
uor
o
us
.
1,
2
,
3
-
T
r
i
az
ol
-
4
-
y
l
m
et
hy
l
A
m
i
ne
s
and
A
m
i
n
e
D
er
i
v
at
i
v
es
f
or
N
ov
el
S
u
r
f
ac
t
a
nt
.
A
us
t
r
a
li
a
n J
ou
r
n
al
o
f
C
hem
i
s
t
r
y
.
2015
;
68(
1)
:
57
-
68
.
[5
]
W
ibo
w
S
ant
o
s
o,
D
e
ng
H
ep
u.
M
ul
t
i
-
c
r
i
t
er
i
a
gr
o
up
d
ec
i
s
i
on
m
ak
i
n
g
f
or
ev
al
ua
t
i
ng
t
he
per
f
or
m
a
nc
e
of
e
-
w
as
t
e r
ec
y
c
l
i
n
g pr
o
gr
am
s
u
n
der
un
c
er
t
ai
nt
y
.
W
as
t
e M
anagem
ent
.
20
15;
40(
1)
:
127
-
135
.
[6
]
Ál
v
a
re
z
-
B
or
r
ego
J
.
P
os
i
t
i
on a
nd r
ot
at
i
on
-
i
nv
ar
i
a
nt
pa
t
t
er
n r
e
c
ogn
i
t
i
o
n s
y
s
t
em
by
b
i
nar
y
r
i
n
gs
m
as
k
s
.
J
our
n
al
o
f
M
oder
n O
pt
i
c
s
.
201
5;
62(
1
0)
:
8
51
-
86
4
.
[7
]
D
es
hm
uk
h A
m
i
t
,
Ra
y
K
.
A
nal
y
s
i
s
of
B
r
oa
dban
d V
ar
i
at
i
o
ns
of
U
r
anui
m
-
s
l
ot
c
ut
R
ec
t
ang
ul
ar
M
i
c
r
os
t
r
i
p
A
nt
enn
as
.
I
E
E
E
A
nt
e
nna
s
an
d
P
r
opag
at
i
o
n M
aga
z
i
ne.
2015;
41:
76
-
82
.
[8
]
F
er
gu
s
on
G
r
ant
.
S
c
r
e
eni
n
g
f
or
he
at
t
r
an
s
por
t
by
gr
o
un
dw
at
er
i
n
c
l
o
s
ed
ge
ot
her
m
a
l
s
y
s
t
e
m
s
.
G
r
oundw
at
er
.
20
15;
5
3(
3)
:
50
3
-
506
.
[9
]
P
us
pi
t
ani
ngr
u
m
D
i
ya
h
,
P
a
gu
ad
J
er
i
A
pr
i
an
s
y
ah.
T
he
A
nal
y
s
i
s
of
R
an
k
F
us
i
on
T
ec
h
ni
q
ues
t
o
I
m
pr
ov
e
Q
uer
y
R
el
ev
anc
e
.
T
el
k
o
m
ni
k
a
.
201
5
;
13(
4
)
:
1
495
-
1
504.
[
10]
S
huai
P
ang
,
W
e
n
bi
n
L
i
,
J
i
a
n
gm
i
n
g
K
a
n.
S
i
m
u
l
at
i
on
A
nal
y
s
i
s
of
I
nt
er
f
ac
e
C
i
r
c
ui
t
s
f
or
P
i
ez
oel
ec
t
r
i
c
E
ner
gy
H
ar
v
es
t
i
n
g w
i
t
h D
am
p
ed S
i
nu
s
oi
dal
S
i
gna
l
s
an
d R
a
ndom
S
i
gna
l
s
.
T
el
k
om
ni
k
a.
2
015;
13(
3)
:
767
-
7
75
.
[
11]
D
eng
Z
h
ong
l
i
a
ng,
X
i
Y
ue,
Y
i
n
Lu.
U
na
m
b
i
g
uo
us
S
i
ne
-
P
ha
s
ed
B
O
C
(
k
n
,
n)
S
i
gna
l
A
c
qui
s
i
t
i
on
B
a
s
e
d
on C
om
bi
n
ed C
or
r
e
l
at
i
on F
un
c
t
i
on
s
.
T
el
k
o
m
ni
k
a
.
2
01
5
;
13
(
2
)
:
502
-
50
9
.
[
12]
S
ongha
o J
i
a,
C
ai
Y
ang.
R
e
c
e
i
v
ed S
i
gna
l
S
t
r
en
gt
h I
n
di
c
at
or
N
ode Lo
c
al
i
z
at
i
on A
l
gor
i
t
hm
B
as
ed o
n
C
ons
t
r
a
i
nt
P
ar
t
i
c
l
e S
w
ar
m
O
pt
i
m
i
z
at
i
on.
T
el
k
om
ni
k
a
.
20
1
5
;
13
(
1
)
:
22
1
-
2
29
.
[
13]
G
ar
c
és
,
J
or
ge
E
du
ar
do
.
A
t
om
i
s
t
i
c
S
i
m
ul
at
i
on
of
H
i
g
h
-
D
e
n
s
i
ty
U
r
ani
um
F
uel
s
.
S
c
i
en
c
e
&
T
ec
hn
ol
o
gy
of
N
uc
l
ear
I
ns
t
al
l
at
i
o
ns
.
2
01
1;
2011:
1
-
16
.
[
14]
G
ov
i
ndar
aj
an,
S
r
i
s
har
an
G
G
r
ay
bi
l
l
,
B
r
i
an
S
M
ak
ar
ew
i
c
z
,
P
hi
l
i
p
F
.
A
s
s
em
b
l
y
and
I
r
r
ad
i
at
i
o
n
M
odel
i
ng
of
R
es
i
d
ual
S
t
r
e
s
s
e
s
i
n Low
-
E
nr
i
c
h
ed U
r
ani
um
F
oi
l
-
B
a
s
e
d A
nnul
ar
T
ar
get
s
f
or
M
ol
y
b
denu
m
-
9
9
P
r
oduc
t
i
on
.
S
c
i
en
c
e &
T
ec
h
no
l
ogy
of
N
u
c
l
e
ar
I
n
s
t
a
l
l
at
i
on
s
.
2
013;
1(
9)
:
1
-
9
.
[
15]
Y
uan
C
enx
i
,
W
a
ng
X
um
i
n
g.
A
S
i
m
pl
e F
or
m
ul
a f
or
Lo
c
al
B
ur
nup an
d I
s
o
t
ope D
i
s
t
r
i
but
i
on
s
B
as
ed o
n
A
ppr
ox
i
m
at
el
y
C
ons
t
ant
R
el
a
t
i
v
e R
eac
t
i
on R
at
e.
S
c
i
en
c
e
&
T
ec
hno
l
og
y
of
N
uc
l
ear
I
n
s
t
al
l
a
t
i
on
s
.
2016;
1(
8)
:
1
-
8
.
[
16]
G
hor
annev
i
s
s
M
.
R
ev
i
ew
on
R
ec
en
t
D
ev
el
o
pm
ent
s
i
n
L
as
er
D
r
i
v
en I
n
er
t
i
a
l
F
us
i
on.
S
c
i
en
c
e
&
T
e
c
hn
o
l
og
y
o
f
N
uc
l
ea
r
I
ns
t
a
ll
at
i
o
ns
.
20
14;
1(
14)
:
1
-
14
.
Evaluation Warning : The document was created with Spire.PDF for Python.