Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
10
,
No.
4
,
A
ugus
t
2020
,
pp.
4270
~
427
8
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v10
i
4
.
pp
4270
-
42
78
4270
Journ
al h
om
e
page
:
http:
//
ij
ece.i
aesc
or
e.c
om/i
nd
ex
.ph
p/IJ
ECE
A
h
igh secu
rity an
d noise im
mu
nity
of
sp
eech based
on double
chaotic m
asking
Ehab Ab
dulR
az
z
aq
Hus
sei
n
1
,
M
urt
ad
h
a K. K
hash
an
2
, Ameer
K. J
awad
3
1,2
Depa
rtment
of
Elec
tr
ical Engi
n
ee
ring
,
Fa
cul
t
y
o
f
Engi
n
ee
ring
,
Univer
sit
y
of
B
ab
y
lon
,
Ir
aq
3
Depa
rtment of
Com
pute
r
Engi
n
ee
ring
,
Fa
cul
t
y
o
f
Engi
n
ee
ring
,
Univer
sit
y
of
Isl
a
m
ic
,
Ir
aq
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
J
ul
11
, 2
019
Re
vised
Feb
2
5
,
2020
Accepte
d
Ma
r
2
, 2
020
It
is
known
tha
t
inc
re
asing
the
sec
urity
of
the
informati
on
an
d
red
uci
ng
the
noise
eff
ect
through
publi
c
cha
nne
ls
are
t
wo
of
the
m
ai
n
priori
ties
in
deve
lop
ing
an
y
comm
unic
at
ion
s
y
stem.
In
thi
s
art
i
cl
e
,
an
eff
ici
ent
,
se
cur
e
comm
unic
at
ion
s
y
stem
with
two
le
vel
s
of
enc
r
y
pt
ion
has
bee
n
appl
ie
d
to
the
spee
ch
sign
al
.
Th
e
suggested
sec
urity
app
roa
ch
was
implemente
d
b
y
using
two
diffe
ren
t
stage
s
of
chaotic
m
asking
on
the
signal
;
one
m
asking
was
conduc
t
ed
b
y
us
ing
Lore
nz
s
y
s
tem
and
the
othe
r
m
asking
was
bui
lt
b
y
u
sin
g
Rӧss
le
r
cha
otic
flow
sy
st
em.
Th
e
m
ai
n
goal
of
deve
lop
ing
thi
s
two
-
cha
ot
ic
m
asking
appr
oa
ch
is
to
in
crease
the
k
e
y
s
pac
e
and
th
e
sec
urity
of
the
informati
on
.
Also,
an
immunit
y
technique
has
bee
n
implemente
d
in
the
suggested
a
pproa
ch
to
red
u
ce
the
noi
se
eff
ec
t
.
For
pra
ct
i
cal
appl
i
ca
t
ion
purposes,
thi
s
s
y
stem
was
te
sted
with
addi
ti
v
e
white
gaussian
nois
e
(AW
GN)
cha
nne
l.
The
si
m
ula
ti
on
resul
ts
show
tha
t
the
qu
al
ity
of
re
constr
uct
ed
spe
ec
h
signal
is
cha
ng
ea
bl
e
a
cc
ord
ing
to
the
used
si
gnal
to
noise
r
at
io
(SN
R);
the
ref
or
e,
a
pro
posed
te
chn
ique
base
d
on
dig
it
a
l
proc
essing
m
ethod
(DP
M)
was
appl
ie
d
to
t
he
first
m
aske
d
signal
b
y
conve
r
ti
ng
the
sam
ple
d
dat
a
from
the
analog
to
th
e
bina
r
y
form
at
.
The
sim
ula
ti
on
result
s
show
th
at
an
22
dB
(SN
R)
is
suffic
i
ent
to
re
c
over
t
he
spee
ch
signa
l
with
m
ini
m
um
noise
by
using t
he
suggested approach.
Ke
yw
or
d
s
:
Chaotic
f
l
ow s
yst
e
m
s
Dou
ble ch
a
otic
m
asking
com
m
un
ic
at
ion
sec
ur
it
y
Encr
y
ption
Sp
eec
h qu
al
it
y
Copyright
©
202
0
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
Eha
b Abd
ulRaz
zaq
Hussei
n,
Dep
a
rtm
ent o
f El
ect
rical
En
gi
neer
i
ng,
Faculty
of E
ngineerin
g,
Un
i
ver
sit
y o
f B
abyl
on, Baby
lon
,
Iraq
.
Em
a
il
: dr
.eh
a
b@i
tnet.u
obaby
lon
.e
du.iq
1.
INTROD
U
CTION
In
c
omm
un
ic
at
ion
syst
em
se
cur
it
y
is
the
m
ai
n
go
al
that
m
us
t
be
achieved
,
in
orde
r
to
kee
p
data
away
from
at
ta
cks
(an
eaves
droppe
r)
.
In
f
or
m
at
ion
m
us
t
be
encr
y
pted
over
var
io
us
te
chn
i
qu
e
s,
Send
e
ncr
y
pted
s
peec
h
or
an
y
form
of
data
need
to
us
e
a
ge
ner
al
key
that
sho
uld
be
known
by
t
he
rece
iver
of
com
m
un
ic
at
ion
syst
em
.
A
c
hao
ti
c
syst
em
has
m
any
pr
operti
es
li
ke,
ve
ry
sensiti
ve
to
init
ia
l
con
diti
on
a
nd
par
am
et
ers,
er
godic,
unpredi
ct
able sig
nal an
d determ
inist
i
c syst
e
m
. D
ue t
o
the im
po
rtan
t prop
e
rtie
s of
chao
t
i
c
that m
entioned, ch
a
otic ca
n b
e u
ti
li
zed in se
cur
e
co
m
m
un
ic
at
ion
.
Seve
ral
w
orks
and
stu
dy
ca
n
be
do
ne
in
the
ap
plica
ti
on
of
cha
otic
in
secu
re
c
om
m
un
ic
at
ion
.
Since
19
93
Oppe
nh
ei
m
a
nd
C
uo
m
o,
im
ple
m
ented
m
asking
sc
he
m
e
of
the
L
or
e
nz
syst
em
in
secure
com
m
un
ic
a
ti
on
[1
,
2
]
.
In
1996
Mi
la
no
vice
Zag
hloul
pu
t
an
im
pr
ove
d
te
ch
nique
to
im
ple
m
ent
chao
ti
c
m
asking
in
t
he
com
m
un
ic
at
i
on
fiel
d
[
3
,
4
]
.
A
fter
t
his
c
ha
otic
ci
rcu
it
of
sec
ur
e
com
m
un
ic
at
ion
ba
sed
on
Lor
e
nz
syst
em
is
i
m
ple
m
ented
in
disc
rete
el
ect
ro
nics
c
om
po
nen
t
[
5
,
6
]
.
In
20
09
a
s
ecur
e
c
omm
un
ic
at
ion
base
d
on
a
fr
ac
ti
on
al
cha
otic
m
et
ho
d
wa
s
pr
opos
e
d
[
7
]
.
In
2011,
a
ne
w
3
-
dim
ension
al
syst
e
m
based
on
chaos
theo
ry
was
pr
opose
d
by
Li
an
d
Ou
[
8
]
,
us
in
g
the
ch
aotic
fl
ow,
L
or
e
nz
syst
e
m
.
Zhang
a
nd
Z
hao
ha
ve
stud
ie
d
a
two
-
st
age
dri
ver
syst
em
by
e
m
plo
yi
ng
the
pecora
an
d
ca
rroll
(P
C)
m
eth
od
on
the
L
orenz
c
haos
m
o
del
in
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
A h
i
gh secu
rit
y and n
oise i
mmun
it
y
of s
peech b
as
e
d
…
(
Eh
ab A
bdulR
az
z
aq Hussein
)
4271
the
fiel
d
of
se
cur
e
c
omm
un
ic
at
ion
s
[
9
]
.
I
n
2018,
pr
ese
nt
ed
a
secu
rity
of
s
peec
h
base
d
on
tw
o
e
ncry
ption
sta
ge,
th
e
fir
st
sta
ge
is
ti
m
e
scram
bling
to
co
ns
tr
uct
b
y
chao
ti
c
m
ap,
wh
il
e
the
seco
nd
sta
ge
is
c
ha
otic
m
asking establ
ished
b
y t
he
c
ha
otic flo
w [
10
].
Yam
ada
and
Fujisaka
[
11
]
,
condu
ct
e
d
on
the
first
rese
arch
on
sy
nchronizi
ng,
a
ch
aotic
syst
e
m
.
Wh
il
e
pecora
carr
oll
[
12
,
13
]
achieve
that
chao
ti
c
m
od
el
s
can
be
s
ync
hr
on
iz
e
d
wh
e
n
a
su
it
able
c
onne
ct
ion
desig
n
is
f
ound,
s
uc
h
as
the
chao
ti
c
ti
m
e
g
rowth
of
t
he
two
-
syst
em
s
beco
m
e
co
m
par
able.
Cha
otic
secu
r
e
com
m
un
ic
at
ion
is
pr
ese
nted
in
th
ree
m
ai
n
ty
pes:
Cha
otic
m
od
ulati
on
,
chao
ti
c
m
asking
an
d
c
ha
otic
sh
ift
keyi
ng [
14
,
15
]
. In
t
his wo
rk two l
evels
of ch
aotic
m
asking
are a
pp
li
ed
to
t
he
s
peec
h
sig
na
l.
The
outl
ine o
f
this
arti
c
le
is
a
s
go
a
hea
d.
I
n
Sect
ion
2
, w
e
s
et
the d
escri
ption
of
t
he
pro
pose
d
syst
em
;
the
ai
m
of
the
m
ul
ti
chao
ti
c
m
asking
;
the
de
sign
of
the
propose
d
syst
em
with
a
dd
it
ive
wh
it
e
gau
s
sia
n
no
is
e
(AWGN
)
c
hannel
ha
s
been
in
tro
du
ce
d.
I
n
S
ect
ion
3
,
we
e
xpos
e
the
qual
it
y
m
easur
em
ent
of
the
sp
eec
h
so
as
to
ex
plain
the
rob
us
tness
of
t
he
m
od
el
that
stud
ie
d
co
ncernin
g
the
resi
dual
cl
arit
y
of
the
ci
ph
ere
d
sig
nal
an
d
the
pro
per
ti
es
of
the
receive
d
data.
I
n
Sect
ion
4
the
sim
ul
at
ion
res
ults
ar
e
ob
ta
ine
d
incl
ud
e
t
he
m
aski
ng
/
de
-
m
asking
pr
oce
ss
an
d
the
res
ul
ts
of
us
i
ng
the
pro
posed
m
eth
od
t
o
re
du
ce
the
e
ff
ect
of
no
i
se.
T
he
c
oncl
usi
on
is
pr
ese
nted
in
t
he
last
secti
on.
2.
THE
PRO
PO
SED
M
ODEL
The
sec
ur
it
y
of
the
e
nc
rypte
d
sig
nal
inc
rea
ses
pro
portio
na
l
to
the
rand
om
ness
of
the
c
hao
ti
c
si
gn
al
and
the
siz
e
of
the
key
sp
ace
of
the
syst
em
that
be
us
e
d.
In
this
w
ork
double
m
asking
to
the
sp
eec
h
sign
al
base
d
on
c
hao
t
ic
sign
al
ha
s
be
en
intr
oduce
d.
Each
pa
rt
of
the
pr
opos
e
d
sy
stem
need
s
to
be
sync
hro
nize
d
wit
h
their
ide
ntica
l
par
ts
at
t
he
re
cei
ver
si
de.
A
s
a
res
ult,
sel
f
-
sync
hro
nizat
ion
(P
C
)
m
et
ho
d
that
pro
po
s
ed
by
Peco
ra
Ca
rroll
has
bee
n
us
e
d
in
this
m
od
el
,
in
wh
ic
h
two
identic
al
syst
em
s
,
on
e
m
ast
e
r
syst
e
m
(d
rive
n)
a
nd
oth
e
r
sla
ve
sys
tem
(r
esp
on
s
e)
can
be
sync
hron
iz
e
d,
by
ch
oosin
g
on
e
of
th
e
sta
te
s
of
th
e
diff
e
re
ntial
equ
at
ions
of
m
ast
er
syst
em
as
a
dr
ive
n
sign
al
t
o
the
ot
her
side
sla
ve
syst
e
m
[
12
]
.
L
or
e
nz
a
nd
Rӧ
s
sle
r
co
ns
ist
,
t
he
first
a
nd
the
sec
ond
sta
ges
of
the
pro
posed
m
od
el
,
resp
ect
ively
as
cl
arify
in
Figu
re
1.
The
n
the
no
ise
effe
ct
on
th
e
receive
d
data
ha
s
been
i
ntr
oduced
by
ap
plyi
ng
a
n
A
WGN
channel
to
the
syst
e
m
,
so
that
a
sp
eci
fic
m
e
t
hod
i
s
pro
po
se
d
t
o
r
edu
ce
no
ise
on
the
sig
nal
b
y
dig
it
al
proces
sin
g
m
eth
od
t
o
t
he
first
cha
otic
m
aski
ng
.
The pr
opose
d
m
od
el
co
m
po
ne
nts ca
n divide
d
as
the
fo
ll
ow
ing
:
Fig
ure
1
.
Ge
ne
ral d
ia
gr
am
o
f pro
po
se
d
m
od
e
l
2.1.
Double
m
as
ki
ng
se
cure c
om
munica
tion
Ther
e
are
se
ve
ral
ap
plica
ti
ons
to
the
c
ha
otic
sign
al
in
secu
r
e
com
m
un
ic
at
i
on
s
su
c
h
as
the
par
am
et
er
chao
ti
c
m
od
ul
at
ion
,
the
c
haot
ic
sh
ift
keyi
ng
,
the
on
-
off
keyi
ng,
an
d
th
e
chao
ti
c
m
asking
[
16
]
.
The
chao
ti
c
m
asking
ca
n
be
con
si
der
e
d
the
sim
plest
ap
plica
ti
on
f
or
t
he
cha
otic
sig
nal
in
the
sec
ur
e
c
omm
un
icati
on
s
.
Also
,
it
can
be
easi
ly
execu
te
d
in
the
el
ect
r
onic
ci
rcu
it
s,
s
o
it
is
us
ed
in
th
is
propose
d
sy
stem
.
The
sche
m
at
ic
of the
double
m
asking for
th
e pro
posed
sys
tem
u
se
d
i
n
thi
s stu
dy is s
how
n
in
Fig
ure
2.
At
transm
it
te
r
side.
I
n
the f
irs
t
blo
ck
,
the
L
orenz f
lo
w
syst
em
is
us
ed
[
17
]
,
since
we
hav
e
three
sta
te
s
in
the
syst
em
(
x
̇
,
y
̇
,
z
̇
)
an
d
acc
ording
to
sel
f
-
sync
hro
nizat
ion
,
by
cho
ic
e
x
m1
as
a
chao
ti
c
m
ask
to
the
dat
a
,
as foll
ow.
T
he Lore
nz
m
ast
er
syst
e
m
is g
ive
n by:
x
̇
m1
=
σ
(
y
m1
−
x
m1
)
y
̇
m1
=
r
x
m1
−
y
m1
−
x
m1
z
m1
(1)
z
̇
m1
=
x
m1
y
m1
−
b
z
m1
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
4
,
A
ugus
t
2020
:
4270
-
4278
4272
So
,
the
first m
aske
d
sig
nal
c
1
(
t
)
is
giv
e
n by:
c
1
(
t
)
=
m
(
t
)
+
x
m1
(
t
)
(2)
Wh
e
re:
m
(
t
)
is t
he speec
h
sig
nal.
m
&
s su
bsc
rip
t:
is the m
ast
er an
d sl
ave
syst
em
s r
especti
vely
.
In
the
seco
nd
bloc
k,
the
R
ӧs
sle
r
cha
otic
syst
e
m
is
use
d
a
nd
acc
ord
ing
t
o
sel
f
-
sy
nchr
on
iz
at
io
n,
by cho
os
in
g
y
m2
a
s the c
ha
otic m
ask
t
o
c
1
(
t
)
as foll
ows,
Fig
ure
2
.
Sc
he
m
at
ic
o
f
do
ub
l
e m
asking
pro
po
s
ed
syst
em
The
R
ӧs
sle
r
m
ast
er m
od
el
is
giv
e
n by:
x
̇
m2
=
−
(
y
m2
+
z
m2
)
y
̇
m2
=
x
m2
+
a
y
m2
(3)
z
̇
m2
=
b
+
z
m2
(
x
m2
−
c
)
So
,
the
seco
nd
m
asked
si
gn
al
c
2
(
t
)
w
hich se
nt t
o
t
he
c
hannel is
gi
ven
by:
c
2
(
t
)
=
c
1
(
t
)
+
y
m2
(
t
)
(4)
At
recei
ver
si
de
.
Th
e
recei
ve
d
sig
nal
c
̂
2
(
t
)
is
subtract
ed
f
ro
m
the
fi
rst
bl
ock
(
Rӧssler
sla
ve
s
yst
e
m
)
by the c
ha
otic
sign
al
y
s2
to
giv
e
the r
ec
overe
d
s
ign
al
c
̂
1
(
t
)
as foll
ows.
The
Rӧssle
r
sla
ve
syst
em
is give
n by:
x
̇
s2
=
−
(
y
s2
+
z
s2
)
y
̇
s2
=
x
s2
+
a
y
s2
(5)
z
̇
s2
=
b
+
z
s2
(
x
s2
−
c
)
So
,
c
̂
1
(
t
)
is ex
pr
ess
ed by:
c
̂
1
(
t
)
=
c
2
(
t
)
−
y
s2
c
̂
1
(
t
)
=
(
c
1
(
t
)
+
y
m2
)
−
y
s2
(6)
=
c
1
(
t
)
+
e
r2
wh
e
re:
e
r2
=
y
m2
−
y
s2
Finall
y
the
rec
ov
e
re
d
i
nfor
m
at
ion
m
̂
(
t
)
is
obta
in
ed
by
s
ub
tract
i
ng
c
̂
1
(
t
)
from
the
c
hao
ti
c
sig
nal
x
s1
of
the sec
ond blo
ck (Lo
re
nz
sla
ve
syst
em
)
as foll
ows.
T
he Lo
ren
z
slave
syst
e
m
is g
ive
n
by:
x
̇
s1
=
σ
(
y
s1
−
x
s1
)
y
̇
s1
=
r
x
s1
−
y
s1
−
x
s1
z
s1
(7)
z
̇
s1
=
x
s1
y
s1
−
b
z
s1
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
A h
i
gh secu
rit
y and n
oise i
mmun
it
y
of s
peech b
as
e
d
…
(
Eh
ab A
bdulR
az
z
aq Hussein
)
4273
So
,
m
̂
(
t
)
is give
n b
y:
m
̂
(
t
)
=
c
̂
1
(
t
)
−
x
s1
=
(
c
1
(
t
)
+
e
r2
)
−
x
s1
(8)
=
m
(
t
)
+
e
l1
+
e
r2
wh
e
re:
e
l1
=
x
m1
−
x
s1
The
sy
nchr
on
i
zat
ion
e
rror bet
ween L
or
e
nz d
rive
a
nd r
e
spo
ns
e syst
em
s is g
ive
n by:
e
l1
=
x
m1
−
x
s1
e
l2
=
y
m1
−
y
s1
(9)
e
l3
=
z
m1
−
z
s1
The
sy
nchr
on
i
zat
ion
e
rror bet
ween R
ӧs
sle
r d
rive
a
nd r
e
spo
ns
e syst
em
s is g
ive
n by:
e
r1
=
x
m2
−
x
s2
e
r2
=
y
m2
−
y
s2
(10)
e
r3
=
z
m2
−
z
s2
2.2.
Double
m
as
ki
ng
se
cure c
om
munica
tion s
ystem u
nder
t
h
e
AW
GN
ch
annel
Fo
r
pract
ic
al
app
li
cat
io
ns
th
e
pro
po
se
d
sy
stem
was
te
ste
d
a
nd
sim
ulate
d
with
A
WG
N
c
ha
nn
el
,
so
as
t
o
st
ud
y
t
he
ef
fect
of
no
ise
in
the
rec
onstr
ucted
i
nform
at
ion
an
d
to
est
i
m
at
e
the
am
ou
nt
of
po
w
er
m
us
t
be
ad
de
d
to
th
e
sign
al
in
ord
er
to
recei
ve
it
at
good
qual
i
ty.
Fig
ur
e
3
il
lus
trat
es
the
pro
pose
d
double
m
askin
g
secur
e
co
m
m
un
ic
at
ion
syst
em
u
nd
er
the
A
WGN c
ha
nn
el
.
Figure
3.
D
oubl
e
m
asking propo
s
ed
syst
em
u
nd
e
r
t
he
A
WGN c
hannel
The reco
ve
red
sp
eec
h
sig
nal i
s affect
ed by
noise
as
f
ollow
:
(
)
=
2
(
)
+
(
)
(11)
w
he
re
(
)
is
A
ddit
ive
Wh
it
e Ga
ussi
an
No
ise
(A
WGN)
.
And refe
rr
e
d
t
o
(
7
)
then
̂
(
)
=
(
)
+
1
+
2
+
(
)
(12)
2.3.
No
ise
reduc
tio
n by di
git
al pr
ocessin
g
me
thod
(DPM
)
In
this
m
et
ho
d,
the
first
m
asked
si
gn
al
c
1
(
t
)
is
transfo
rm
ed
fro
m
the
analo
g
to
t
he
di
gital
ut
il
iz
ing
(ADC)
c
onve
rter,
befor
e
m
ake
m
asking
with
the
c
ha
os
si
gn
al
of
t
he
sec
ond
blo
c
k
of
t
he
pr
opos
e
d
s
yst
e
m
.
At
the
receive
r
side,
t
he
first
recovere
d
sig
na
l
c
̂
1
(
t
)
w
hich
is
bi
nar
y
data
f
orm
will
be
c
ha
nged
ov
e
r
bac
k
t
o
the
analo
g
f
orm
by
us
ing
(
DA
C
)
trans
f
orm
er.
This
co
nversi
on
of
fi
rs
t
m
asked
signa
l
to
bin
ary
f
orm
wil
l
reduce
the
im
pact
of
noise
on
the
sign
al
.
I
n
f
act
,
the
sam
ple
s
that
conve
rted
f
ro
m
the
analog
t
o
the
bin
ar
y
ca
n
be
act
e
d
by
(
1,
2,
3
or
4...
.
N
b
)
bits
f
or
ea
ch
sam
ple
[
18
,
19
]
.
Cl
ear
s
peech
is
obta
ined
,
by
i
ncr
ea
se
the num
ber
of
bits that
us
ed
.
The pr
opose
d
m
et
ho
d
is
d
e
pi
ct
ed
by t
he
s
ub
seq
uen
t
points:
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
4
,
A
ugus
t
2020
:
4270
-
4278
4274
-
Ma
ke
c
onve
rting
to
t
he
sam
ples
of
t
he
fir
st
m
asked
si
gnal
1
(
)
from
analog
to
t
he
dig
i
ta
l
form
us
in
g (ADC)
.
-
Ma
sk
in
g
the
c
hange
d ov
e
r da
ta
w
it
h
the
sec
ond bloc
k
c
ha
ot
ic
sign
al
.
-
Add A
WGN c
hannel to
the
m
asked
si
gn
al
.
-
On
the
receiv
er
side,
an
d
af
te
r
synch
roniz
at
ion
of
the
fi
rst
blo
c
k
sla
ve
syst
e
m
,
the
bin
ary
data
was
recovere
d, an
d t
hen
by
us
in
g (
DA
C
)
the
sam
pled data
was ob
ta
ine
d.
3.
QUALIT
Y
M
EASUR
MEN
T
OF
SPEE
C
H
A
num
eral
of
qu
a
ntit
ie
s,
m
easur
em
ents
can
be
util
iz
ed
to
e
valuate
the
ef
f
ect
iveness
of
t
he
pro
pose
d
syst
e
m
con
cer
ning
sec
ur
it
y
of
enc
rypte
d
sig
nal
an
d
the
pr
op
e
rtie
s
of
the
recove
red
s
pe
ech
sig
nal.
T
he
se
are
sign
al
-
t
o
noisy
rati
o
(SNR)
,
(
LPC)
li
near
predic
ti
ve
co
de
m
easur
e
,
ce
ps
t
ral
distance
m
e
asur
e
(CD
)
an
d
m
ean
sq
ua
re e
rro
r
(
MSE)
[
20
,
21
]
.
Th
ese
m
easur
e
m
ents are defi
ned as
fo
ll
ows:
3.1.
Segme
ntal
spe
ctra
l si
gnal
-
to
no
ise
ra
tio
Segm
ental
sp
ect
ral
sign
al
-
to
no
ise
rati
o
(SSSNR
)
is
a
quantifi
cat
ion
of
no
ise
in
a
spe
ci
fic
sign
al
.
It
is
a
colle
c
ti
ve
m
easur
em
ent
of
the
resid
ual
cl
arit
y
of
the
encry
pted
sp
eec
h
and
the
fin
e
ness
of
the r
ec
onstr
ucted
s
peec
h.
It is
sp
eci
fie
d by th
e subse
qu
e
nt e
qu
at
io
n
[
22
]
:
SSS
N
R
(
s
(
i
)
,
sn
(
i
)
)
=
1
0
log
(
∑
s
(
i
)
2
i
=
l
i
=
1
∑
(
s
(
i
)
−
sn
(
i
)
)
2
i
=
l
i
=
1
)
(13)
w
h
e
r
e
l
i
s
“
n
um
b
e
r
o
f
s
a
m
p
l
e
s
”
,
s
(
i
)
i
s
o
r
i
g
i
n
a
l
s
i
g
n
a
l
a
m
pl
i
t
u
d
e
a
n
d
sn
(
i
)
i
s
“
t
h
e
a
m
p
l
i
t
u
d
e
”
o
f
t
h
e
e
n
c
r
y
p
t
e
d
or d
ec
rypted
si
gn
al
.
3.2.
Li
near
pre
-
di
cat
i
ve
c
od
e
(L
PC)
d
lpc
=
ln
(
AV
A
T
BV
B
T
)
(14)
w
he
re:
V
is
the
auto
-
co
rr
el
at
ion
m
at
rix
of
t
he
act
ual
sp
eec
h
bl
oc
k,
v
ect
ors A
&
B
co
ntain
the
LPC
coe
ffi
ci
ent
for
the
pu
re sp
eech
blo
c
k
a
nd
the
recover
ed
or enc
rypted
s
peech bl
ocks
[
23
]
.
3.3.
Cp
es
tra
l
dist
ance
me
as
ure
(
CD)
CD
=
10
log
10
[
2
∑
{
C
x
(
n
)
−
C
y
(
n
)
}
2
p
n
=
1
]
1
2
(15)
w
he
re:
c
y
(
n
)
a
nd
c
x
(
n
)
are
th
e
ce
pestral
c
oe
ff
ic
ie
nts
of
the
ori
gin
al
s
pe
ech
an
d
the
rec
onstr
ucted
or
encr
y
pted
s
pee
ch.
3.4.
Mean
sq
uare
error
(MSE)
MSE
=
∑
(
y
−
x
)
2
n
(16)
Is
the m
easur
e o
f
the
recove
re
d
sp
eec
h qu
al
it
y, wh
e
re th
e s
m
al
le
st value o
f
MSE m
eans g
oo
d
qual
it
y
of
the
rec
onstr
ucted
s
peech
si
gn
al
w
he
re
(x)
is
“
the
or
igina
l
sign
al
”,
(y)
is
“t
he
reconstr
uc
te
d
sign
al
”
an
d
(
n)
is “t
he
le
ngth
vec
tor of t
he si
gnal
”
[
24
,
25
]
.
4.
SIMULATI
O
N
RES
ULTS
In
this
par
t,
t
he
the
or
et
ic
al
analy
ses
and
resu
lt
s
are
presente
d
to
ex
plain
the
ef
fe
ct
iveness
of
the
pro
po
se
d
s
yst
e
m
.
The
sp
eech
file
enter
ed
thr
ough
the
coo
l
edit9
6
de
vice,
the
file
with
8
kHz,
8bit
s
per
sam
ples,
on
e
channel,
42
960
sam
ples
fo
r
5.370
0
seco
nd.
The
sim
ulati
on
resu
lt
s
will
be
pr
ese
nted
as
f
ollows;
double
m
asking
im
ple
m
entation
of
the
s
pe
ech
sig
nal
fro
m
so
ur
ce
to
de
sti
nation,
the
i
m
pact
of
A
WGN
channel
noise
on
s
peec
h
da
ta
at
receiver
side
inclu
des
the
resu
lt
s
of
noise
re
du
ct
ion
by
us
in
g
dig
it
al
processi
ng
m
eth
od
(
DP
M
),
th
e
key
sp
ace
a
nd
pe
rfo
rm
ance
of
the
pro
pos
ed
m
od
el
as
com
par
ed
with
so
m
e
encr
y
ption sy
stem
s ar
e prese
nt
ed.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
A h
i
gh secu
rit
y and n
oise i
mmun
it
y
of s
peech b
as
e
d
…
(
Eh
ab A
bdulR
az
z
aq Hussein
)
4275
4.1.
Simul
at
i
on
re
sults
of d
ou
bl
e chao
tic ma
s
king
Re
sp
ect
to
the
process
of
e
ncry
ption
as e
xp
la
ined
in secti
on
(II.A),
th
e o
ri
gi
nal sp
eech si
gnal
en
te
re
d
with
4296
0
sa
m
ples,
5.
37
00
seco
nd
durati
on
sho
wn
in
Fi
gure
4(a)
.
The
n
the
m
asked
sign
al
f
ro
m
the
first
sta
ge
of
enc
ryption
is
giv
e
n
i
n
Fi
gure
4(
b
)
.
Finall
y
the
e
nc
rypted
sig
nal
t
hat
pr
oduce
d
f
ro
m
the
tw
o
st
ages
o
f
m
ast
er s
yst
e
m
is give
n by Fi
gure
4(
c
)
.
(a)
(b)
(c)
Figure
4.
(a
) O
rigin
al
s
peec
h si
gn
al
,
(b) M
as
ked sig
nal afte
r
the
f
irst
sta
ge
of e
ncr
ypti
on,
(c)
The
en
c
ryp
te
d
s
peech
Fr
om
the
ab
ove
encr
y
pted
si
gn
al
,
it
is
cl
ear
that
the
entire
sp
eec
h
data
is
hidden
inside
the
cha
otic
sign
al
a
nd
thi
s
is
one
of
t
he
m
ai
n
require
m
ents
to
obta
in
the
ef
fici
ent
secu
rity
.
On
t
he
sla
ve
side,
after
a
pply
ing
the
process
of
the
first
a
nd
th
e
sec
ond
de
-
m
askin
g
bl
ock
s
t
he
final
rec
ov
e
red
s
peech
m
̂
(
t
)
wa
s
ob
ta
ine
d w
hich
is n
ea
rly
ex
act
to
the
origi
nal
sp
eec
h data wi
th m
ean squar
e
erro
r
(MS
E)
=
5
.
2692
∗
10
−
6
.
4.2.
Prop
os
ed
s
ys
t
em unde
r
AW
GN
c
hannel
Dou
ble
m
ask
secur
e
c
omm
u
nicat
ion
syst
e
m
is
su
bj
ect
e
d
an
d
te
ste
d
unde
r
the
A
WG
N
c
ha
nn
el
,
Table
1
il
lustr
at
es
the
ef
fect
of
noise
on
r
ecov
e
re
d
s
pee
ch
sig
nal
for
diff
e
re
nt
S
NRs.
It
is
noti
ceable
that
the
syst
em
has
poor
pe
rfo
rm
a
nce
at
32
an
d
35
dB,
wh
il
e
w
hen
the
SN
R
is
reac
hing
ove
r
37
dB
t
he
recovere
d
sp
eec
h
sig
nal
beco
m
es
cl
ear
er
(
SSS
NR
ha
ve
posit
ive
val
ue).
T
o
cl
ari
fy
this,
diff
e
re
nt
SN
Rs
a
re
sim
ulate
d
as
in Figu
re
5.
By
us
in
g
the
pro
po
se
d
m
et
ho
d
DP
M
,
the
first
m
asked
s
ign
al
was
c
onver
te
d
f
ro
m
the
analo
g
t
o
the
di
gital
for
m
so
that
the
i
m
pu
nity
of
th
e
cha
ng
e
in
t
he
m
asked
sig
na
l
values
due
to
the
noise
w
ou
l
d
increase.
Table
2
s
hows
the
r
esults
that
obta
ined
by
us
i
ng
t
he
pr
opos
e
d
sc
hem
e
at
diff
er
en
t
SN
Rs
.
It
is
seen
from
this
ta
bl
e,
in
case
of
15
a
nd
16
dB
,
the
syst
e
m
has
po
or
e
xe
cution
util
iz
ing
t
he
three
sp
eec
h
unde
rstan
dab
il
it
y
m
od
el
s,
fro
m
18
dB
an
d
go
up,
the
syst
e
m
execu
ti
on
is
enh
a
nce
d
(i.e
SSS
NR
be
gin
t
o
ha
ve
po
sit
ive
estee
m
).
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
4
,
A
ugus
t
2020
:
4270
-
4278
4276
(a)
(b)
Figure
5
.
The
re
cov
e
re
d
s
peec
h
at
: (a
)
S
NR=
35 d
B
(b) SNR
=39 d
B
Table
1.
T
est
the
qu
al
it
y
of
re
cov
e
re
d
s
peech f
or
diff
e
re
nt S
NRs
SNR [
d
B]
LPC
SSSNR
[
d
B]
MSE
29
1
.29
8
2
-
7
.62
9
8
0
.03
8
8
32
1
.24
2
9
-
4
.41
8
1
0
.01
9
4
35
1
.16
4
7
-
1
.14
5
5
0
.00
9
7
37
1
.10
8
1
1
.07
7
6
0
.00
6
1
39
1
.01
9
4
3
.16
7
4
0
.00
3
9
40
0
.99
8
0
4
.36
2
5
0
.00
3
0
Table
2.
T
est
the
qu
al
it
y
of
re
cov
e
re
d
s
peech f
or
diff
e
re
nt S
NRs
b
y
us
in
g (DP
M)
SNR [
d
B]
LPC
SSSNR
[
d
B]
MSE
15
3
.49
0
8
-
2
4
.55
4
2
2
.56
5
7
16
3
.18
6
5
-
1
3
.01
2
9
0
.87
2
3
18
2
.53
6
4
1
9
.06
0
8
0
.00
8
0
20
1
.02
4
6
3
1
.63
7
6
0
.00
2
5
21
0
.25
3
1
3
4
.29
5
2
3
.71
3
5
∗
10
−
6
22
0
.09
9
3
3
4
.29
3
1
6
.00
4
1
∗
10
−
8
4.3.
Key are
a
Key
area
is
the
overall
n
um
ber
o
f
“
keys”
t
ha
t
can b
e u
ti
li
zed f
or
the
e
ncr
y
ption
strat
e
gy. I
t
sho
uld
b
e
gr
eat
e
nough,
to
fig
ht
the
bru
te
fo
rce
at
ta
c
ke
r.
I
n
our
pr
opos
e
d
syst
em
a
high
sec
ur
it
y
can
be
obta
ined
from
the secret
key i
s d
i
vid
e
d on tw
o
syst
em
Lo
re
nz
a
nd Rӧs
sle
r sy
stem
s as f
oll
ow
s:
a)
Fo
r
L
or
e
nz
sys
tem
The
sec
ret
ke
y
is
(
σ
,
r
,
b
l
)
a
floati
ng
point
pr
eci
si
on
of
∆
=
10
−
2
is
util
iz
ed
f
or
t
he
sec
ret
keys,
therefo
re th
e
key
area is
(
10
2
)
3
=
10
6
.
b)
Fo
r
Rӧs
sle
r
sy
stem
The
sec
ret
ke
y
is
(
a
,
b
r
,
c
)
a
floati
ng
point
pr
eci
si
on
of
∆
=
10
−
3
is
util
iz
ed
f
or
t
he
sec
ret
keys,
therefo
re
t
he
ke
y
area
is
(
10
3
)
3
=
10
9
.
Th
eref
or
e
,
th
e
tot
al
key
s
pace
is
(
10
6
∗
10
9
)
wh
ic
h
is
a
hu
ge
key
siz
e to r
e
sist
th
e comm
on
att
ack.
4.4.
Perfo
r
ma
nce
comp
ariso
n
of t
he
so
m
e encr
yp
ti
on
sy
s
tem
s a
n
d dou
ble c
haot
ic
m
aski
n
g
s
ys
te
m
Fo
r
hi
gh
secu
rity
the
resid
ua
l
intel
li
gib
il
ity
of
e
ncr
y
pted
sig
nal
s
houl
d
be
sm
al
l
as
possible
t
o
pr
e
ve
nt
any
a
tt
acker
from
reach
t
o
e
ncr
y
pted
data.
F
or
m
or
e
ex
plain
ed
a
bout
m
eth
ods
that
us
e
d
for
encr
y
pting
t
he
sel
f
-
s
peech
cl
ip
c
on
si
der
e
d
i
n
this
resear
ch
see
[
26
]
.
The
stud
ie
d
w
orks
us
e
d
are:
“F
re
qu
e
ncy
do
m
ai
n
scram
bling
”
,
“Ti
m
e
dom
ai
n
scram
bl
ing
”
a
ls
o
“Tw
o
dim
ensi
on
al
scram
bling
”
res
pecti
ve
ly
as
in
Table
3.
Table
3.
Propo
sed
m
od
el
r
e
sul
ts as com
par
e
d wit
h
s
om
e encr
ypti
on syst
e
m
s
Clas
sical Scr
a
m
b
li
n
g
LPC
SSSNR
[
d
B]
CD
Ti
m
e
do
m
ain
scr
a
m
b
lin
g
0
.65
3
2
0
.97
5
4
2
.42
7
3
Frequ
en
cy
do
m
ain
scra
m
b
lin
g
0
.58
2
3
-
0
.27
3
5
2
.50
9
5
Two d
i
m
en
sio
n
s scra
m
b
lin
g
0
.61
3
2
-
1
.95
4
3
3
.23
6
9
Prop
o
sed
sy
ste
m
0
.97
5
1
-
2
1
.56
2
0
3
.96
6
1
It
is
ob
vious
f
ro
m
Table
3
that
the
stu
died
m
od
el
is
m
uc
h
bette
r
tha
n
oth
e
r
e
ncr
ypti
on
m
et
ho
ds.
In
public,
an
d
after
lo
okin
g
t
o
the
gai
ned
re
su
lt
s
we
obse
r
ve
that
t
he
acc
um
ulate
d
exec
ution
by
us
in
g
doubly
chao
ti
c
m
asking
util
iz
ed,
in
t
erm
s
of
“SSS
NR”
is
m
uch
r
edu
ce
d
(
from
0.975
4
to
-
21.
5620)
c
om
par
ed
wit
h
the tim
e d
om
ain
m
et
ho
d
a
nd t
his in
dicat
e of
decr
ease
the
r
e
sidu
al
i
ntell
igibil
it
y of
the
enc
rypt
ed
d
at
a.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
A h
i
gh secu
rit
y and n
oise i
mmun
it
y
of s
peech b
as
e
d
…
(
Eh
ab A
bdulR
az
z
aq Hussein
)
4277
5.
CONCL
US
I
O
N
In
this p
a
per, an
acti
ve
secur
e
co
m
m
un
ic
ation
syst
e
m
w
as a
pp
li
ed
to the info
rm
ation
signal
b
ased
on
the
double
chao
ti
c
m
asking
te
ch
nique.
The
f
ollo
wing
con
cl
us
io
ns
can
be
dr
a
w
n
from
this
stud
y:
(a)
T
he
sug
ges
te
d
secu
rity
app
r
oac
h
was
im
plem
ented
by
us
in
g
tw
o
dif
f
eren
t
sta
ges
of
chao
ti
c
m
aski
ng
on
the
sig
nal;
on
e
m
asking
was
cond
ucted
by
us
in
g
L
or
e
nz
s
yst
e
m
and
the
oth
e
r
m
asking
was
bu
il
t
by
us
ing
Rӧssler
cha
otic
flo
w
syst
em
;
(b)
A
la
rg
e
key
s
pace
of
(
10
6
∗
10
9
)
was
obta
ined
by
usi
ng
t
his
pro
po
se
d
m
et
ho
dolo
gy.
The
la
r
ge
key
sp
ace
of
this
m
od
el
m
akes
the
syst
e
m
with
a
hi
gh
r
andom
ness
an
d
m
or
e
i
m
m
un
it
y
agai
ns
t
at
ta
cks;
(c
)
By
us
ing
this
double
cha
otic
m
asking
te
c
hniq
ue,
t
he
resi
du
al
i
ntell
igibil
it
y
of
the
e
ncr
ypte
d
data
decr
ease
d
w
her
e
the
S
S
SN
R
is
m
uch
reduce
d
from
0.975
4
to
-
21.
5620
wh
ic
h
is
m
uch
lowe
r
as
com
par
ed
with
so
m
e
of
the
pre
vious
stu
dies;
(d)
Fo
r
us
a
ge
ap
plica
ti
on
s,
the
pr
opos
e
d
syst
em
was
te
ste
d
ov
e
r
an
A
WGN
cha
nnel
and
the
res
ul
ts
sh
ow
e
d
that
the
qu
al
it
y
of
the
recovere
d
inf
or
m
a
ti
on
be
gin
s
to
be
cl
ear
at
a
m
ini
m
u
m
SN
R
va
lue
of
37
dB
with
an
MS
E
e
qu
al
to
0.0
061;
(e)
T
he
dig
it
al
processi
ng
m
et
hod
(D
PM
)
has
be
en
c
om
bin
ed
with
this
a
ppr
oach
to
r
ed
uce
the
ef
fect
of
no
ise
on
the
r
ecov
e
re
d
in
for
m
at
ion
.
This
c
om
bin
at
ion
le
d
t
o
m
ake
the
pro
po
se
d
ap
proac
h
wor
k
properly
at
SN
R
of
21
dB
with
a
m
ean
sq
ua
re
error
(
MSE
)
=
3
.
7135
∗
10
−
6
.
Althou
gh
t
his
com
bin
at
ion
is
ver
y
ben
e
fici
al
to
i
m
pr
ove
the
qu
al
it
y
of
the
rec
ov
e
re
d
sign
al
,
t
he
c
om
pl
ic
at
ion
of
the
syst
em
was
increase
d
a
nd
this
will
requi
re
so
m
e
incre
m
ents
in
the b
a
ndwi
dth
of the c
ha
nn
el
du
e
to
t
he
c
ha
nge in
the
data t
ype f
ro
m
the a
nalo
g
to
the
b
i
nar
y
form
.
REF
E
RE
NCE
S
[1]
Cuom
o
K.M,
Oppenhe
im
A.
V,
and
Strogatz
S.
H,
“
S
y
nchr
oniz
a
ti
on
of
lo
ren
z
-
base
d
chaotic
c
irc
u
it
s
wit
h
appl
i
ca
t
ions
to
comm
unic
at
ions
,
”
I
EE
E
Tr
ansacti
ons
on
c
ircu
it
s
and
systems
II
:
analog
an
d
digi
tal
signal
proce
ss
ing
,
vol
.
40,
no
.
10
,
pp
.
6
26
-
633,
1993
.
[2]
S.
Sridhara
n
,
E
.
Daws
on,
B.
Goldburg
,
“
Fast
Fourier
Tra
nsf
orm
Based
Speec
h
En
cr
y
pti
on
S
y
stem
,
”
I
E
E
Proce
ed
ings I
-
Comm
unic
ati
ons,
Spe
ec
h
and
Vi
s
ion
,
v
o
l. 138,
n
o.
3,
pp.
215
-
223
,
Jun.
1991.
[3]
Mila
nović
V,
an
d
Za
ghlou
l
M.
E
,
“
Im
prove
d
m
as
king
al
gor
it
hm
for
cha
ot
ic
communic
a
ti
ons
s
y
st
e
m
s
,”
El
e
ct
ronic
s
Lett
ers
,
vo
l. 32,
no.
1
,
pp
.
11
–
12
,
1996.
[4]
Rupak
K.,
“
De
si
gn
and
Im
ple
m
ent
at
ion
o
f
Secur
e
Chaotic
Com
m
unic
at
ion
S
y
s
t
ems
,”
Ph.D.
Th
e
sis,
Univer
sit
y
o
f
Northum
bria
at
Newca
stle,
2011
.
[5]
Yu
-
Min
L,
Yu
-
Hong
Z,
and
Ji
n
-
Quan
Y,
“
Cir
cui
t
implementation
of
sec
ur
e
c
om
m
unic
at
ion
s
y
stem
b
ase
d
o
n
improved
chaoti
c
m
asking al
go
ri
thm
,”
Journal
of
Circui
ts
and
S
yste
ms
,
vol
.
14
,
no
.
1
,
pp
.
116
–
118
,
2009.
[6]
J.
C.
Feng,
C.
K.
Tse,
“
Re
construc
ti
on
of
Chao
ti
c
Sign
al
s
with
Applic
a
ti
ons
to
Chaos
-
Based
C
om
m
unic
at
ions
,”
Tsinghua
Univ
er
sit
y
Press
,
2008
.
[7]
Arm
an
Kiani
-
B,
et
al
.
,
“
A
ch
a
oti
c
sec
ure
com
m
unic
at
ion
sche
m
e
using
fra
ct
i
onal
cha
ot
ic
s
y
stems
base
d
on
an
extende
d
fra
ct
ion
al
k
al
m
an
f
il
te
r
,”
Comm
unic
ati
ons
in
Nonli
near
Scienc
e
an
d
Numerical
S
i
mulati
on
,
v
o
l.
1
4,
no.
3
,
pp
.
863
-
8
79,
2009
.
[8]
X.
Li
,
and
Q.
O
u,
“
Dy
n
amical
prope
rties
and
sim
ula
ti
on
of
a
n
ew
lore
nz
-
l
ike
c
haot
i
c
s
y
stem
,”
Nonli
near
Dyn
.,
v
ol.
65
,
no
.
3
,
pp
.
255
-
270
,
2011
.
[9]
Y
.
Zha
ng
,
Y
,
H.
Zha
o
,
and
L.
Zha
ng
,
“
S
y
nchr
onizati
on
o
f
cha
o
ti
c
s
y
st
e
m
and
it
s
application
in
se
cur
e
comm
unic
at
ion
,
“
In
Inte
rnationa
l
Confe
ren
ce
on
Computer
Appl
i
cat
ion
and
System
Mode
li
ng
(
ICCASM)
,
vol.
14
,
pp.
94
-
97
,
2010
.
[10]
Saad
S.
Hreshee,
Hikm
at
N.
Ab
dull
ah
,
and
Am
ee
r
K.
Jawad
,
“
A
high
sec
urity
comm
unic
at
ion
s
y
stem
base
d
o
n
cha
ot
ic
sc
rambli
ng
and
cha
o
ti
c
m
asking
,”
In
te
r
nati
onal
Journa
l
on
Comm
unications
Antenna
and
Propagati
o
n
(
I.
Re
.
C.
A.P.
)
,
vo
l.
8
,
no
.
3
,
2018
.
[11]
Fujisaka
,
H.
,
and
Yam
ada
,
T,
“
Stabi
lit
y
the
or
y
of s
y
nchr
oni
ze
d
m
oti
on
in
coupled
-
oscil
l
at
or
s
y
st
e
m
s
,”
Prog.
Theor.
Phy
s.,
v
ol
.
69
,
n
o.
1
,
pp
.
32
-
47
,
1983
[12]
Pecor
a,
L.
M
.
,
and
Carr
ol
l,
T.L,
“
S
y
nch
ronizat
ion
in
ch
aot
i
c
s
y
st
ems
,”
Ph
ys.
Rev.
Lett.,
v
ol.
64
,
no,
8
,
pp
.
821
-
824
,
19
90
.
[13]
Sivape
rum
al
,
S,
“
Hy
brid
S
y
n
chr
oniz
a
ti
on
of
Id
e
nti
c
al
Chao
ti
c
Sy
stems
Via
Nov
el
Slidi
ng
cont
ro
l
with
Appli
ca
t
i
on
to
H
y
per
ch
aotic
Vaid
y
anatha
n
—
Volos s
y
stem
,”
I
nt.
J. Cont
rol
Theory
App
l
.
,
vo
l.
9,
pp
.
261
–
278
,
2016.
[14]
Li
S,
Alvar
ez
G
,
and
Chen
G,
“
Brea
king
a
cha
o
s
-
base
d
sec
ure
c
om
m
unic
at
ion
sche
m
e
designe
d
b
y
an
improved
m
odula
ti
on
m
et
h
od
,
”
Chaos
,
Sol
i
tons a
nd
Frac
tals
, v
ol. 25
,
no.
1,
pp.
109
-
120
,
20
05.
[15]
Huan
Z.
,
Shaofa
ng
H.,
Zuo
C.
,
a
nd
Xixia
ng
Z.
,
“
Dual
Ke
y
Spee
c
h
Enc
r
y
pti
on
Al
gorit
hm
Based
Underde
te
rm
ined
BS
S
,”
The
Sc
ie
n
ti
fic
World
Jour
nal
,
pp
.
1
-
7
,
201
4.
[16]
Nichol
as
R
.
,
“
Int
roduc
ti
on
to Lor
enz'
s S
y
st
em of Equa
ti
ons
,”
Mat
h
6100,
Dec
.
20
03.
[17]
B.
Sklar,
“
Digi
tal
Com
m
unic
at
io
n:
Fundam
ent
al
s
and
Applicati
on
s
,”
P
rent
ice
hall
publi
cation
PT
R
,
Th
ird
Edition
,
2011.
[18]
Ta
riq
,
M.K,
“
Objective
te
sts
of
spee
ch
sign
al
,”
M.Sc.
The
sis,
C
oll
eg
e
of
Engi
n
e
eri
ng
Al
-
Mus
ta
n
siri
y
a
Univer
sit
y
,
Depa
rtment
o
f
E
le
c
tri
c
al E
ngin
eering
,
2001
.
[19]
C.
Chin
Yi,
and
X.
Daol
in
“
Sec
ure
dig
it
a
l
com
m
unic
at
ion
usin
g
cont
ro
ll
ed
pro
je
c
ti
ve
s
y
n
chr
on
isat
ion
of
ch
aos
,”
Chaos,
Solitons
and
Fract
a
ls
,
vo
l.
23
,
no
.
3
,
pp
.
1
063
-
1070,
2005
.
[20]
Y.
Zha
ng,
Y,
H.
Zha
o,
L
.
Zha
ng
,
“
Sy
nchr
on
izati
o
n
of
cha
oti
c
s
y
s
t
em
and
it
s
appl
i
ca
t
ion
in
sec
ure
comm
unic
at
ion,”
In:
Int
ernati
onal
Confe
ren
ce on Com
pute
r A
ppl
i
cat
ion
and
S
ystem
Mode
li
ng
(
ICCASM)
,
vol
.
14,
pp.
94
-
97
,
2010
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
4
,
A
ugus
t
2020
:
4270
-
4278
4278
[21]
Ramin
V.,
“
Se
quenc
e
S
y
n
chr
o
niz
a
ti
on
in
Chaos
-
Based
Dire
ct
Sequence
Sprea
d
Spect
ru
m
Comm
unic
at
io
n
S
y
stems
,”
PhD
.
The
sis,
Univ
er
si
t
y
of
Auck
la
nd
,
New Z
e
al
and
,
2
012.
[22]
A.
V.
Prabu,
S.
Sriniva
sara
o
,
T
.
Appara
o,
M.
J.
Rao
and
K.
B.
R
ao,
“
Audio
Enc
r
y
pt
ion
in
Handset
s
,”
In
te
rnation
al
Journal
of
Computer
App
li
ca
ti
on
s
,
v
ol. 40,
n
o.
6,
pp.
40
-
45
,
2012
.
[23]
X.
Chen,
et
a
l
.
,
“
Adapti
ve
Control
of
Multi
pl
e
Chaot
ic
S
y
s
tem
s
W
it
h
Unkno
wn
Para
m
et
ers
In
Two
Diffe
re
nt
S
y
nchr
oni
zat
ion M
odes
,”
Ad
v. Di
ff
er.
Equ
.
,
vol. 2
31,
pp
.
1
-
17
,
20
16.
[24]
Rubak
K.,
K.
B
usawon,
and
Z
.
Ghass
emloogy
,
“
Novel
Casca
de
Chaot
i
c
Maski
ng
For
Secur
e
Com
m
unic
at
ion
,”
In
Confe
renc
e
T
he
9th
annual
P
ostgraduate
Symposium
on
the
conv
erge
nc
e
of
Telecomm
unic
ations
,
Net
working
&
Broadcasti
ng
(
P
GNET 2008)
,
At:
L
ive
rpool
,
UK
2008.
[25]
A.
Abel
and
W
.
Schwarz
,
“
Chaos
Com
m
unic
at
ions
-
Princi
ple
s,
Schemes
and
Sy
s
te
m
Anal
y
sis
,”
Proc
ee
dings
o
f
The
IEEE
,
vol
.
9
0,
no
.
5
,
pp
.
691
–
710,
2002
.
[26]
N.
Abdulla
h,
et
al
.
,
“
Design
of
eff
icient
no
ise
red
uction
sche
m
e
for
sec
ure
sp
e
ec
h
m
aske
d
b
y
cha
ot
ic
signa
l
,”
Journal
of
Ame
r
ic
an
S
ci
en
ce
,
vo
l
.
11
,
no
.
7
,
pp
.
49
-
55,
2015
.
BIOGR
AP
H
I
ES
OF
A
UTH
ORS
Eh
ab
Abdu
lRa
z
z
a
q
Hus
sein
,
PhD
.
MS
c.
Ele
ct
ri
ca
l
Engi
n
ee
r
ing
was
born
i
n
Bab
y
lon
on
Janua
r
y
1
,
1976.
He
obta
ine
d
h
is
BS
c
degr
ee
(19
97)
in
El
e
ct
ri
cal
Engi
neering
at
the
Facult
y
of
Engi
ne
eri
ng,
Univer
sit
y
of
B
a
b
y
lon
and
MS
c
degr
ee
(2000),
in
elec
tri
c
al
e
ngine
er
ing
at
the
Depa
r
tment
of
El
e
ct
ri
ca
l
En
gine
er
ing,
Unive
rsi
t
y
of
Technol
og
y
and
h
is
PhD
.
Degre
e
from
the
Depa
r
tment
of
El
e
ct
ri
ca
l
En
gine
er
ing
at
the
Facul
t
y
of
Eng
ine
er
ing,
Univ
er
sit
y
of
B
asra
h,
Curre
ntly
h
e
works
as
assistant
profe
ss
or
at
the
E
le
c
trica
l
Depa
rtment
at
the
Facu
lty
o
f
Engi
ne
eri
ng,
Un
ive
rsit
y
of
Bab
y
l
on.
His
m
ai
n
int
ere
st
is
signal
pr
oce
ss
ing,
anal
y
s
i
s,
informati
on
tra
nsiti
on
,
senso
rs a
nd
cont
rol
s
y
stem a
naly
sis.
Murt
adh
a
Kha
fie
f
Khas
han
,
El
e
ct
ri
ca
l
Engi
n
ee
r
was
born
in
Thi
Qar,
1991.
He
obta
ine
d
his
BCs
degr
ee
(2014)
in
El
ec
tr
ical
Engi
nee
ring
a
t
the
Facul
t
y
of
E
ngine
er
ing,
Uni
ver
sit
y
of
Kufa
.
Curre
ntly
he
is
stud
y
ing
for
a
m
aste
r'
s
degr
e
e
in
Elec
tr
ic
a
l
Engi
ne
eri
ng
a
t
the
Facult
y
o
f
Engi
ne
eri
ng,
Un
ive
rsit
y
of
Bab
ylon.
His
m
ai
n
in
te
rest
is
comm
u
nic
a
ti
on
s
y
stems
,
digi
tal
signal
proc
essing,
m
e
a
surem
e
nt
and
co
ntrol
s
y
s
te
m
an
a
l
y
sis.
Amee
r
K
.
Ja
wad
was
born
in
Hill
ah
,
Ira
q
,
in
1988.
He
re
ceiv
ed
his
BS
c
degr
ee
in
Elec
tron
ic
and
Com
m
unic
at
ion
Eng
ine
e
ri
ng
from
the
El
e
ct
roni
c
and
Comm
unic
at
io
n
Engi
nee
r
ing
Depa
rtment,
Al
-
Hos
sain
Univer
sit
y
Coll
ege,
Kar
bal
a
,
Ira
q
,
in
2
012.
He
obt
ai
ne
d
his
MS
c
in
El
e
ct
roni
c
and
Com
m
unic
at
ion
Engi
ne
eri
ng
from
the
El
e
ctrical
Engi
n
ee
rin
g
Depa
rtment
,
Al
-
Mos
ta
nsiria
Univer
sit
y
,
Ira
q
,
in
2015.
Since
2015,
he
h
as
bee
n
with
the
staff
of
the
Dep
art
m
ent
of
Com
pute
r En
gine
er
ing, Isla
m
i
c
Univer
si
t
y
Col
le
ge
,
Na
ja
f
–
Ira
q
.
Evaluation Warning : The document was created with Spire.PDF for Python.