I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
15
,
No
.
5
,
Octo
b
er
20
25
,
p
p
.
4
6
0
5
~
4
6
1
9
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
15
i
5
.
pp
4
6
0
5
-
4
6
1
9
4605
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
A
memo
ry
impro
v
ed propo
rtiona
t
e affine
proje
ctio
n alg
o
rithm
for spa
rse sy
stem
identif
i
ca
tion
Senthil
M
urug
a
n B
o
o
pa
la
n
1
,
Sa
ro
j
ini
Ra
j
u
2
,
K
rit
hig
a
Su
k
um
a
r
a
n
1
,
M
a
nim
eg
a
la
i Mu
nis
a
m
y
1
,
K
a
lph
a
na
I
la
ng
o
v
a
n
1
,
Su
dh
a
Ra
ma
cha
nd
ra
n
3
,
J
a
na
ni M
un
is
a
m
y
1
,
B
ha
ra
t
hira
j
a
R
a
m
a
m
o
o
r
t
hi
1
,
Sa
k
t
hiv
el
P
icha
ik
a
r
a
n
3
1
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
o
n
i
c
s a
n
d
C
o
m
mu
n
i
c
a
t
i
o
n
En
g
i
n
e
e
r
i
n
g
,
T
h
a
n
t
h
a
i
P
e
r
i
y
a
r
G
o
v
e
r
n
me
n
t
I
n
st
i
t
u
t
e
o
f
Te
c
h
n
o
l
o
g
y
,
V
e
l
l
o
r
e
,
I
n
d
i
a
2
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
o
n
i
c
s a
n
d
C
o
m
mu
n
i
c
a
t
i
o
n
En
g
i
n
e
e
r
i
n
g
,
G
o
v
e
r
n
m
e
n
t
C
o
l
l
e
g
e
o
f
E
n
g
i
n
e
e
r
i
n
g
,
B
o
d
i
n
a
y
a
k
a
n
u
r
,
I
n
d
i
a
3
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
i
c
a
l
a
n
d
El
e
c
t
r
o
n
i
c
s E
n
g
i
n
e
e
r
i
n
g
,
T
h
a
n
t
h
a
i
P
e
r
i
y
a
r
G
o
v
e
r
n
m
e
n
t
I
n
st
i
t
u
t
e
o
f
T
e
c
h
n
o
l
o
g
y
,
V
e
l
l
o
r
e
,
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
Dec
1
2
,
2
0
2
4
R
ev
is
ed
J
u
n
6
,
2
0
2
5
Acc
ep
ted
J
u
l 3
,
2
0
2
5
F
o
r
c
lu
ste
r
sp
a
rse
sy
ste
m
id
e
n
ti
fica
ti
o
n
,
it
is
k
n
o
w
n
th
a
t
th
e
c
lu
s
ter
sp
a
rse
imp
ro
v
e
d
p
ro
p
o
r
ti
o
n
a
te
a
ffin
e
p
ro
jec
ti
o
n
a
l
g
o
rit
h
m
(CS
-
IP
APA)
o
u
t
p
e
rfo
rm
s
t
h
e
sta
n
d
a
rd
IP
APA.
Ho
we
v
e
r,
sin
c
e
CS
-
I
P
APA
d
o
e
s
n
o
t
re
tai
n
p
a
st
p
r
o
p
o
rti
o
n
a
te
fa
c
to
rs,
it
s
p
e
rfo
rm
a
n
c
e
c
a
n
b
e
f
u
rth
e
r
imp
r
o
v
e
d
.
I
n
th
is
p
a
p
e
r,
a
m
o
d
ifi
c
a
ti
o
n
t
o
CS
-
I
P
A
P
A
is
p
r
o
p
o
se
d
b
y
u
ti
l
izin
g
th
e
p
a
st
in
sta
n
t
p
ro
p
o
rti
o
n
a
te
e
lem
e
n
ts
b
a
se
d
o
n
it
s
p
r
o
jec
ti
o
n
o
r
d
e
r.
S
tea
d
y
-
sta
te
p
e
rfo
rm
a
n
c
e
o
f
th
e
p
r
o
p
o
se
d
m
e
m
o
ry
c
l
u
ste
r
sp
a
rse
imp
r
o
v
e
d
p
ro
p
o
rti
o
n
a
te
a
ffin
e
p
r
o
jec
ti
o
n
a
lg
o
rit
h
m
(M
CS
-
IP
APA)
is
st
u
d
ied
b
y
d
e
ri
v
in
g
t
h
e
c
o
n
d
i
ti
o
n
f
o
r
m
e
a
n
sta
b
il
i
ty
.
D
iffere
n
t
sim
u
lati
o
n
se
tu
p
s
sh
o
w
th
a
t
th
e
p
ro
p
o
se
d
a
lg
o
rit
h
m
o
u
t
p
e
rfo
rm
s
d
iffere
n
t
v
e
rsio
n
s
o
f
I
P
APA
i
n
term
s
o
f
c
o
n
v
e
rg
e
n
c
e
ra
te,
n
o
rm
a
li
z
e
d
m
is
a
li
g
n
m
e
n
t
(NM)
a
n
d
trac
k
in
g
,
fo
r
d
iffere
n
t
ty
p
e
s
o
f
i
n
p
u
ts
li
k
e
c
o
lo
re
d
n
o
ise
,
wh
i
te
n
o
ise
,
a
n
d
s
p
e
e
c
h
sig
n
a
l.
By
in
c
o
rp
o
ra
ti
n
g
p
a
st
p
ro
p
o
rti
o
n
a
te
fa
c
to
rs,
th
e
p
ro
p
o
se
d
M
CS
-
IP
APA
sig
n
ifi
c
a
n
t
ly
re
d
u
c
e
s c
o
m
p
u
tatio
n
a
l
c
o
m
p
lex
it
y
fo
r
h
i
g
h
e
r
p
ro
jec
ti
o
n
o
r
d
e
rs
.
K
ey
w
o
r
d
s
:
Ad
ap
tiv
e
f
ilter
Af
f
in
e
p
r
o
jectio
n
alg
o
r
ith
m
C
lu
s
ter
s
p
ar
s
e
s
y
s
tem
E
ch
o
ca
n
ce
llatio
n
Sp
ar
s
e
s
y
s
tem
id
en
tific
atio
n
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
:
Sen
th
il M
u
r
u
g
an
B
o
o
p
alan
Dep
ar
tm
en
t o
f
E
lectr
o
n
ics an
d
C
o
m
m
u
n
icatio
n
E
n
g
in
ee
r
i
n
g
,
T
h
an
th
ai
Per
iy
ar
Go
v
er
n
m
en
t
I
n
s
titu
te
o
f
T
ec
h
n
o
lo
g
y
Vello
r
e,
T
am
il Na
d
u
,
I
n
d
ia
E
m
ail: b
s
en
th
il2
4
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
Sp
ar
s
e
s
y
s
tem
id
en
tific
atio
n
h
as
g
ain
ed
m
u
ch
im
p
o
r
ta
n
ce
in
th
e
d
o
m
ain
o
f
ad
ap
ti
v
e
s
ig
n
al
p
r
o
ce
s
s
in
g
.
T
h
e
s
y
s
tem
is
s
aid
to
h
av
e
a
s
p
ar
s
e
im
p
u
ls
e
r
esp
o
n
s
e
if
th
e
n
u
m
b
e
r
o
f
a
ctiv
e
o
r
lar
g
e
co
e
f
f
icien
ts
is
v
er
y
s
ig
n
if
ican
tly
less
er
th
an
th
at
o
f
in
ac
tiv
e
o
r
s
m
all
c
o
ef
f
icien
ts
[
1
]
,
[
2
]
.
I
n
a
b
lo
c
k
-
s
p
ar
s
e
o
r
clu
s
ter
-
s
p
ar
s
e
s
y
s
tem
,
th
e
lar
g
e
co
ef
f
i
cien
ts
ar
e
g
r
o
u
p
e
d
in
clu
s
ter
s
o
r
b
lo
ck
s
.
I
n
p
r
o
p
o
r
tio
n
ate
n
o
r
m
alize
d
least m
ea
n
s
q
u
ar
e
(
PNLM
S),
th
e
ch
o
s
en
s
tep
-
s
izes
o
r
ad
ap
tatio
n
g
ain
s
ar
e
p
r
o
p
o
r
tio
n
al
to
t
h
e
in
d
iv
id
u
al
f
ilter
co
ef
f
icien
ts
,
s
h
o
win
g
h
ig
h
e
r
p
er
f
o
r
m
an
ce
f
o
r
s
p
ar
s
e
s
y
s
tem
s
[
3
]
.
Ho
wev
e
r
,
th
e
PNLM
S,
s
h
o
ws
p
o
o
r
p
er
f
o
r
m
an
ce
f
o
r
Sp
ee
ch
a
n
d
co
lo
r
ed
n
o
is
e
in
p
u
ts
[
4
]
,
[
5
]
.
Ov
er
th
e
y
ea
r
s
,
s
ev
er
al
v
a
r
ian
ts
o
f
PNLM
S
h
av
e
b
ee
n
d
e
v
elo
p
ed
[
6
]
–
[
9
]
.
Af
f
in
e
p
r
o
jectio
n
al
g
o
r
ith
m
(
APA)
,
a
in
p
u
t
-
d
ata
r
eu
s
in
g
al
g
o
r
ith
m
,
ex
h
ib
its
g
o
o
d
p
er
f
o
r
m
an
ce
ev
en
f
o
r
th
e
co
lo
r
ed
n
o
is
e
an
d
s
p
e
ec
h
in
p
u
ts
[
1
0
]
,
[
1
1
]
.
Sev
er
al
v
ar
ian
ts
o
f
th
e
p
r
o
p
o
r
tio
n
ate
APA
(
PAPA)
h
av
e
b
ee
n
in
tr
o
d
u
ce
d
in
th
e
last
two
d
ec
ad
es.
I
m
p
r
o
v
ed
PAPA
s
h
o
ws
g
o
o
d
p
e
r
f
o
r
m
an
ce
e
v
en
f
o
r
d
is
p
er
s
iv
e
p
ath
s
[
1
2
]
–
[
1
5
]
.
I
n
m
em
o
r
y
I
PAPA
(
MI
PAPA),
th
e
p
r
o
p
o
r
tio
n
ate
elem
en
ts
tr
ac
k
r
ec
o
r
d
wa
s
ex
p
lo
ited
to
g
ain
p
er
f
o
r
m
an
ce
o
v
er
th
e
I
PAPA
[
1
2
]
.
I
n
im
p
r
o
v
ed
MI
PAPA
(
I
MI
PAPA)
[
1
3
]
,
l0
n
o
r
m
was
in
tr
o
d
u
ce
d
to
MI
PAPA
as
a
m
ea
s
u
r
e
o
f
s
p
ar
s
ity
,
s
h
o
win
g
im
p
r
o
v
e
d
co
n
v
e
r
g
en
ce
p
er
f
o
r
m
a
n
ce
o
v
er
th
e
MI
PAPA.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
5
,
Octo
b
e
r
20
25
:
4
6
0
5
-
4
6
1
9
4606
Ad
d
itio
n
ally
,
th
e
m
e
m
o
r
ized
p
r
o
p
o
r
ti
o
n
ate
elem
en
ts
b
en
e
f
it
co
m
p
u
tatio
n
al
c
o
m
p
lex
ity
.
T
h
e
im
p
u
ls
e
r
esp
o
n
s
e
is
s
eg
m
en
ted
in
to
m
an
y
b
lo
c
k
s
,
m
o
d
if
y
in
g
PAPA
o
r
I
PAP
A
u
p
d
ate
eq
u
atio
n
with
l2
,
1
p
en
alty
[
1
4
]
to
g
iv
e
b
lo
ck
s
p
ar
s
e
PAPA
(
B
S
-
PA
P
A)
o
r
B
lo
ck
Sp
ar
s
e
I
PAPA
(
B
S
-
I
PAPA)
th
at
r
esu
lts
in
b
ette
r
co
n
v
e
r
g
en
ce
r
ate
,
tr
ac
k
in
g
,
r
ed
u
ce
d
m
is
alig
n
m
e
n
t.
I
n
clu
s
ter
s
p
ar
s
e
PA
PA
(
C
S
-
PAPA)
o
r
clu
s
ter
s
p
ar
s
e
C
S
-
I
PAPA
[
1
6
]
,
[
1
7
]
,
th
e
n
o
r
m
l2
,
0
p
en
alize
s
th
e
PAPA o
r
I
PAPA
as in
C
S
-
P
NL
MS
o
r
C
S
-
I
PNLM
S to
es
tim
at
e
s
p
ar
s
e
ch
an
n
els.
I
t
r
esu
lts
in
b
etter
co
n
v
er
g
en
ce
p
er
f
o
r
m
an
ce
a
n
d
tr
ac
k
in
g
,
a
n
d
less
m
is
alig
n
m
en
t.
T
h
e
im
p
r
o
v
is
atio
n
o
f
C
S
-
PAPA
b
y
ex
p
lo
itin
g
m
em
o
r
y
an
d
its
an
aly
s
es
was
ca
r
r
ied
o
u
t
in
[
1
7
]
.
T
h
e
l0
-
n
o
r
m
B
S
-
PAPA
(
l0
-
BS
-
PA
PA)
[
1
8
]
was
in
tr
o
d
u
ce
d
b
y
in
co
r
p
o
r
atin
g
th
e
l0
-
n
o
r
m
s
p
ar
s
ity
m
ea
s
u
r
e
in
to
B
S
-
PAPA.
T
h
e
ad
d
e
d
p
e
n
alty
l0
-
n
o
r
m
h
el
p
s
in
th
e
s
h
r
in
k
ag
e
o
f
in
ac
tiv
e
c
o
ef
f
icien
ts
th
e
r
eb
y
p
r
o
d
u
cin
g
h
ig
h
er
c
o
n
v
e
r
g
en
ce
r
ate
th
an
th
e
BS
-
PA
PA.
Sev
er
al
im
p
r
o
v
ed
v
er
s
io
n
s
o
f
th
e
PAPA a
n
d
t
h
e
I
PAPA we
r
e
p
r
esen
ted
o
v
er
th
e
y
ea
r
s
[
1
9
]
–
[
2
3
]
.
Me
m
o
r
y
C
S
-
PAPA
ex
ten
d
s
t
h
e
id
ea
o
f
m
em
o
r
y
c
o
n
ce
p
t
t
o
th
e
C
S
-
PAPA.
T
h
e
p
er
f
o
r
m
an
ce
o
f
th
e
MCS
-
PA
PA
i
s
h
ig
h
er
in
ter
m
s
o
f
tr
ac
k
in
g
,
co
n
v
er
g
en
ce
r
at
e
an
d
th
e
m
is
alig
n
m
en
t,
th
an
th
e
C
S
-
PAP
A.
T
h
is
r
esear
ch
wo
r
k
is
th
e
s
eq
u
el
o
f
[
1
7
]
f
o
r
f
u
r
th
e
r
im
p
r
o
v
em
en
t
o
f
C
S
-
I
PAPA.
I
n
th
is
p
ap
er
,
m
o
tiv
ated
b
y
[
1
2
]
,
[
1
7
]
,
we
p
r
o
p
o
s
e
th
e
Me
m
o
r
y
C
S
-
I
PAPA
(
MCS
-
I
PAPA)
b
y
ex
ten
d
in
g
th
e
id
ea
o
f
m
em
o
r
y
in
p
r
o
p
o
r
tio
n
ate
f
ac
to
r
s
to
t
h
e
C
S
-
I
PAP
A.
I
n
co
n
tr
ast
t
o
th
e
C
S
-
I
PAPA,
th
e
p
ast
h
is
to
r
y
o
f
p
r
o
p
o
r
tio
n
at
e
elem
en
ts
is
in
co
r
p
o
r
ated
in
C
S
-
I
PAPA to
en
h
an
ce
its
p
er
f
o
r
m
an
ce
.
T
h
e
m
ain
co
n
tr
ib
u
tio
n
s
o
f
th
is
r
esear
ch
p
ap
e
r
ar
e
:
a.
T
h
e
m
an
u
s
cr
ip
t
is
n
o
v
el
in
th
e
s
en
s
e,
f
o
r
th
e
f
ir
s
t
tim
e,
th
e
m
em
o
r
y
c
h
ar
ac
ter
is
tics
o
f
th
e
p
r
o
p
o
r
tio
n
ate
co
ef
f
icien
ts
ar
e
in
co
r
p
o
r
ate
d
i
n
an
I
m
p
r
o
v
ed
p
r
o
p
o
r
tio
n
ate
af
f
in
e
p
r
o
jectio
n
alg
o
r
ith
m
,
f
o
r
clu
s
ter
s
p
ar
s
e
ch
an
n
els.
T
h
e
m
ath
em
atica
l
a
n
aly
s
is
f
o
r
th
e
u
p
d
ate
eq
u
atio
n
o
f
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
MCS
-
I
PAP
A
i
s
f
u
lly
p
r
esen
ted
.
b.
Stead
y
-
s
tate
p
er
f
o
r
m
an
ce
s
tu
d
y
o
f
th
e
MCS
-
I
PAPA
is
d
er
iv
ed
.
T
h
e
co
n
d
itio
n
f
o
r
th
e
m
ea
n
s
tab
ilit
y
is
d
er
iv
ed
to
p
r
e
d
ict
th
e
s
tead
y
-
s
tate
p
er
f
o
r
m
an
ce
.
T
h
e
c
o
n
d
it
io
n
s
h
o
ws
th
at
th
e
m
ea
n
s
tab
ilit
y
d
ep
en
d
s
o
n
th
e
in
p
u
t
p
o
wer
lev
el,
c.
W
ith
d
if
f
er
en
t
in
p
u
ts
,
th
e
s
u
p
er
io
r
p
er
f
o
r
m
a
n
ce
o
f
th
e
p
r
o
p
o
s
ed
is
s
h
o
wn
o
v
er
th
e
co
m
p
etin
g
alg
o
r
ith
m
s
lik
e
th
e
B
S
-
I
PAPA a
n
d
C
S
-
I
P
APA.
d.
I
n
ter
m
s
o
f
n
u
m
b
er
o
f
th
e
m
u
l
tip
licatio
n
s
,
ad
d
itio
n
s
,
d
i
v
is
io
n
s
,
m
em
o
r
y
s
p
ac
es
an
d
co
m
p
a
r
is
o
n
s
,
th
e
tim
e
co
m
p
lex
ity
o
f
th
e
p
r
o
p
o
s
ed
MCS
-
I
PA
PA
is
co
m
p
ar
ed
ag
ain
s
t
ex
is
tin
g
alg
o
r
ith
m
s
.
T
h
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
s
h
o
ws s
ig
n
if
ican
t
r
e
d
u
ctio
n
in
n
u
m
b
er
o
f
m
u
ltip
lic
atio
n
s
f
o
r
h
i
g
h
er
p
r
o
jectio
n
o
r
d
er
.
I
n
th
is
p
a
p
er
,
lo
wer
ca
s
e
s
y
m
b
o
ls
in
b
o
ld
f
ac
e
an
d
u
p
p
er
ca
s
e
s
y
m
b
o
ls
in
b
o
ld
f
ac
e
ar
e
a
d
o
p
ted
f
o
r
co
lu
m
n
v
ec
t
o
r
s
an
d
m
at
r
ices,
i.e
.
an
d
X,
r
esp
ec
tiv
ely
.
Als
o
,
f
o
r
s
ca
lar
s
lik
e,
n
o
r
m
a
l
f
o
n
t
lo
wer
ca
s
e
s
y
m
b
o
ls
ar
e
u
s
ed
.
T
o
d
en
o
te
th
e
tim
e
d
ep
en
d
en
cy
o
f
s
ca
la
r
s
an
d
v
ec
to
r
s
lik
e
(
)
an
d
(
)
,
p
ar
en
th
eses
o
r
r
o
u
n
d
b
r
ac
k
ets
ar
e
em
p
l
o
y
ed
.
T
h
e
f
o
llo
win
g
n
o
tatio
n
s
ar
e
ta
k
en
u
p
i
n
h
is
r
esear
ch
ar
ticle
:
i)
(
,
)
:
T
r
an
s
p
o
s
e
o
f
a
v
ec
to
r
;
ii)
(
.
)
:
E
x
p
ec
tatio
n
o
r
s
tatis
tical
m
ea
n
;
iii)
|
|
.
|
|
:
E
u
clid
ea
n
n
o
r
m
o
f
a
v
ec
to
r
;
iv
)
|
|
|
|
A
2
:
Gen
er
alize
d
in
n
er
p
r
o
d
u
ct
; a
n
d
v
)
: I
d
en
tity
m
atr
i
x
o
f
d
im
en
s
io
n
p
x
p
.
T
h
e
r
e
s
t
o
f
t
h
e
p
a
p
e
r
i
s
o
r
g
a
n
iz
e
d
a
s
f
o
l
l
o
ws
:
s
e
c
ti
o
n
2
b
r
i
e
f
l
y
r
e
v
i
e
w
s
t
h
e
c
o
n
v
e
n
t
i
o
n
a
l
I
PA
P
A
a
n
d
CS
-
I
P
A
PA
.
T
h
e
p
r
o
p
o
s
e
d
a
l
g
o
r
i
t
h
m
M
C
S
-
I
P
AP
A
a
n
d
i
t
s
u
p
d
a
t
e
e
q
u
a
t
i
o
n
d
e
r
i
v
a
t
i
o
n
p
a
r
t
a
r
e
p
r
e
s
e
n
t
e
d
i
n
s
e
c
ti
o
n
3
.
I
n
s
e
c
ti
o
n
4
,
t
h
e
c
o
n
d
i
t
i
o
n
f
o
r
m
e
a
n
-
s
t
a
b
il
i
t
y
i
s
d
e
r
i
v
e
d
.
S
ec
t
i
o
n
5
p
r
e
s
e
n
ts
t
h
e
s
e
v
e
r
al
s
i
m
u
l
at
i
o
n
e
x
p
e
r
i
m
e
n
t
s
c
a
r
r
i
e
d
o
u
t
a
n
d
t
h
e
r
e
s
u
l
t
s
a
r
e
i
ll
u
s
t
r
a
t
e
d
.
T
h
e
c
o
m
p
u
t
a
t
i
o
n
a
l
c
o
m
p
l
e
x
i
ty
a
n
d
t
h
e
t
r
a
n
s
i
e
n
t
p
e
r
f
o
r
m
a
n
c
e
o
f
t
h
e
p
r
o
p
o
s
e
d
a
l
g
o
r
i
t
h
m
a
r
e
s
t
u
d
i
e
d
i
n
s
e
ct
i
o
n
6
.
F
i
n
a
l
l
y
,
s
e
c
ti
o
n
7
c
o
n
c
l
u
d
e
s
t
h
e
r
e
s
e
a
r
c
h
p
a
p
e
r
.
2.
B
RI
E
F
RE
V
I
E
W
O
F
I
P
AP
A
AND
CS
-
I
P
AP
A
T
h
e
r
o
ad
m
ap
to
t
h
e
r
esear
ch
wo
r
k
is
g
iv
en
b
y
th
e
t
h
eo
r
eti
ca
l
f
r
am
ewo
r
k
b
y
p
r
esen
tin
g
th
e
ex
itin
g
r
elev
an
t
th
eo
r
ies
in
th
e
liter
atu
r
e
.
T
h
e
ec
h
o
ca
n
ce
llatio
n
is
a
ch
allen
g
in
g
s
p
ar
s
e
s
y
s
tem
id
en
tific
atio
n
p
r
o
b
lem
in
wh
ich
th
e
ca
n
ce
ller
m
o
d
els
th
e
im
p
u
ls
e
r
esp
o
n
s
e
o
f
th
e
ec
h
o
p
ath
.
A
ty
p
ical
m
o
d
ellin
g
o
f
ec
h
o
ca
n
ce
ller
is
s
h
o
wn
in
Fig
u
r
e
1
.
Her
e,
th
e
im
p
u
ls
e
r
esp
o
n
s
e
o
f
th
e
u
n
k
n
o
wn
ec
h
o
p
ath
is
g
iv
en
b
y
h
o
f
len
g
th
K.
T
h
e
s
ig
n
al
at
th
e
f
ar
-
en
d
is
g
iv
e
n
b
y
its
s
n
ap
s
h
o
t
tak
en
at
t
im
e
in
s
tan
t
l
as
x
(
l
)
=
[
x
(
l
)
,
x
(
l
−
1
)
,
x
(
l
−
2
)
,
…
,
x
(
l
−
K
+
1
)
]
T
.
T
h
en
th
e
d
esire
d
s
ig
n
al
is
ex
p
r
ess
ed
b
elo
w,
as
th
e
s
u
m
o
f
t
h
e
o
u
tp
u
t
o
f
th
e
u
n
k
n
o
wn
ec
h
o
p
ath
an
d
th
e
n
ea
r
-
en
d
o
r
a
d
d
itiv
e
n
o
is
e
a(
l
)
[
2
]
,
(
)
=
x
(
)
ℎ
+
a
(
)
(
1
)
T
h
e
ec
h
o
p
ath
is
esti
m
ated
b
y
th
e
ad
a
p
tiv
e
f
ilter
as
i
(
−
1
)
=
[
i
0
(
−
1
)
,
i
1
(
−
1
)
,
…
,
i
K
−
1
(
−
1
)
]
.
T
h
e
o
u
tp
u
t
o
f
th
e
a
d
ap
tiv
e
f
ilter
is
g
iv
en
b
y
(
2
)
(
)
=
(
)
(
−
1
)
.
(
2
)
T
h
en
,
th
e
e
r
r
o
r
in
th
e
esti
m
ati
o
n
o
f
th
e
ec
h
o
p
ath
is
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
A
mem
o
r
y
imp
r
o
ve
d
p
r
o
p
o
r
ti
o
n
a
te
a
ffin
e
p
r
o
jectio
n
a
lg
o
r
ith
m
fo
r
…
(
S
en
th
il Mu
r
u
g
a
n
B
o
o
p
a
la
n
)
4607
(
)
=
(
)
−
(
)
(
3
)
Fig
u
r
e
1
.
An
E
ch
o
ca
n
ce
ller
m
o
d
el
2
.
1
.
I
P
AP
A
T
o
f
u
r
t
h
er
ex
p
lo
it
th
e
s
p
ar
s
it
y
an
d
im
p
r
o
v
e
th
e
s
p
ee
d
o
f
co
n
v
er
g
en
ce
,
th
e
PAPA
o
r
th
e
I
PAPA
r
ec
y
cles th
e
in
p
u
t sig
n
al.
T
h
e
i
n
p
u
t d
ata
m
atr
ix
is
g
iv
en
as in
[
1
1
]
.
(
)
=
[
(
)
,
(
−
1
)
,
…
,
(
−
p
+
1
)
]
(
4
)
wh
er
e
‘
p
’
d
e
n
o
tes
th
e
p
r
o
jecti
o
n
o
r
d
e
r
o
f
I
PAPA
an
d
p
<K
.
T
h
e
o
u
tp
u
t
v
ec
t
o
r
,
th
e
d
esire
d
s
ig
n
al
v
ec
to
r
,
an
d
th
e
er
r
o
r
v
ec
to
r
o
f
I
PAPA a
r
e
ex
p
r
ess
ed
as
(
5
)
-
(
7
)
:
z
(
)
=
X
(
)
i
(
−
1
)
,
(
5
)
d
(
)
=
[
(
)
,
(
−
1
)
,
…
,
(
−
p
+
1
)
]
(
6
)
e
(
)
=
d
(
)
−
z
(
)
(
7
)
wh
er
e
d
(
l
)
,
y
(
l
)
a
n
d
e
(
l
)
r
ep
r
esen
t
t
h
e
d
esire
d
s
ig
n
al
v
ec
to
r
,
th
e
o
u
t
p
u
t
v
ec
to
r
,
an
d
th
e
er
r
o
r
v
ec
to
r
,
r
esp
ec
tiv
ely
.
T
h
en
th
e
u
p
d
ated
eq
u
atio
n
o
f
th
e
I
PAPA is
ex
p
r
ess
ed
b
ased
o
n
[
2
4
]
as
(
8
)
:
i
(
)
=
i
(
−
1
)
+
U
(
−
1
)
X
(
)
(
X
(
)
U
(
−
1
)
X
(
)
+
δ
I
P
A
P
A
I
p
)
−
1
e
(
l
)
(
8
)
wh
er
e
U
(
)
,
δ
I
PAPA
,
μ
d
en
o
tes
r
esp
ec
tiv
ely
,
th
e
p
r
o
p
o
r
tio
n
ate
m
atr
ix
,
r
e
g
u
lar
izatio
n
p
ar
am
eter
,
a
n
d
th
e
s
tep
s
ize
o
f
I
PAPA.
T
h
en
U(
l
)
is
ex
p
r
ess
ed
b
ased
o
n
[
2
4
]
as
(
9
)
:
U
(
l
)
=
dia
g
{
u
0
(
l
)
,
u
1
(
l
)
,
…
,
u
K
-
1
(
l
)
}
(
9
)
wh
er
e
th
e
elem
en
ts
ar
e
g
iv
en
b
y
(
1
0
)
,
u
f
(
l)
=
(1
-
α)
2
+
(
1+α
)
‖
i
f
(
l)
‖
2
2
∑
‖
i
n
(
)
‖
2
K
-
1
n=
0
,
0
≤
f
≤
K
-
1
(
1
0
)
wh
er
e
α
is
a
co
n
s
tan
t
a
n
d
is
o
f
ten
s
elec
ted
b
etwe
en
-
1
an
d
1
.
T
h
e
I
PAPA
ac
ts
lik
e
PAPA
f
o
r
α
clo
s
e
to
1
.
I
f
α
=
-
1
,
I
PAPA is
s
am
e
as APA.
A
g
o
o
d
s
elec
tio
n
f
o
r
α
is
eith
er
0
o
r
-
0
.
5
[
4
]
.
2
.
2
.
CS
-
I
P
AP
A
B
y
in
co
r
p
o
r
atin
g
a
m
ix
ed
-
n
o
r
m
l
2,
0
p
en
alty
in
to
th
e
I
PAPA,
th
e
C
S
-
I
PAPA
is
p
r
esen
ted
,
f
av
o
r
i
n
g
th
e
clu
s
ter
-
s
p
ar
s
e
ch
ar
ac
ter
is
tic
o
f
th
e
ec
h
o
p
ath
c
h
an
n
el
[
1
6
]
.
T
h
is
alg
o
r
ith
m
is
b
ased
o
n
th
e
f
ac
t
th
at
th
e
l
0
-
n
o
r
m
is
a
b
etter
ch
o
ice
f
o
r
ex
p
lo
itin
g
th
e
s
p
ar
s
e
ch
ar
ac
ter
is
tic
th
an
th
e
l
1
-
n
o
r
m
.
Her
e,
th
e
l
2
-
n
o
r
m
is
u
s
ed
f
o
r
s
eg
r
eg
atin
g
th
e
ch
a
n
n
el
in
to
c
lu
s
ter
s
o
f
eq
u
al
s
ize.
T
h
e
u
p
d
ate
eq
u
atio
n
o
f
th
e
al
g
o
r
ith
m
C
S
-
I
PAP
A
is
s
am
e
as th
e
C
S
-
PAP
A,
ex
ce
p
t f
o
r
t
h
e
d
ef
in
itio
n
o
f
th
e
g
ain
d
is
tr
ib
u
tio
n
m
atr
ix
.
)
‖
2,
0
=
‖
‖
[
‖
i
[
0]
‖
2
‖
i
[
1]
‖
2
⋮
‖
i
[M
-
1]
‖
2
]
‖
‖
0
(
1
1
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
5
,
Octo
b
e
r
20
25
:
4
6
0
5
-
4
6
1
9
4608
I
n
(
1
1
)
,
‘
M
’
d
en
o
tes
th
e
n
u
m
b
er
o
f
cl
u
s
ter
s
in
th
e
ec
h
o
p
at
h
ch
an
n
el,
i.e
,
M
=
K/Q
an
d
‘
Q’
is
th
e
n
u
m
b
e
r
o
f
co
ef
f
icien
ts
p
er
clu
s
ter
o
r
clu
s
ter
s
ize.
T
h
en
th
e
l
0
-
n
o
r
m
ap
p
r
o
x
im
ati
o
n
[
2
5
]
o
f
th
e
ec
h
o
p
at
h
o
r
th
e
weig
h
t e
s
tim
ate
v
ec
to
r
is
|
|
i(
l
)|
|
0
≈
∑
(1
-
e
-
β||
i
[
t
]
||
)
M
−
1
t
=0
,
(
1
2
)
an
d
‘
β’
s
h
o
u
ld
b
e
alwa
y
s
g
r
ea
ter
th
an
ze
r
o
.
T
h
en
,
t
h
e
l
2,
0
-
n
o
r
m
ap
p
r
o
x
im
atio
n
[
1
6
]
o
f
th
e
ec
h
o
p
ath
o
r
weig
h
t
esti
m
ate
v
ec
to
r
is
|
|
i(
l
)|
|
2,
0
≈
∑
(1
-
e
-
β||
i
[
t
]
|
|
2
)
M
-
1
t
=0
,
(
1
3
)
T
h
e
u
p
d
ate
eq
u
atio
n
o
f
th
e
C
S
-
I
PAPA is
as
(
1
4
)
[
1
6
]
.
i(
l
)
=
i(
l
-
1
)
+
μ
U
CI
(
l
-
1
)
X(
)(
X
T
(
)
U
CI
(
-
1
)
X(
)+
δ
CI
I
p
)
-
1
e(
).
(
1
4
)
wh
er
e
δ
CI
is
th
e
r
eg
u
lar
izatio
n
p
ar
am
eter
f
o
r
C
S
-
I
PAPA.
T
h
e
d
iag
o
n
al
m
atr
i
x
U
CI
(
l
-
1
)
is
U
CI
(
l
-
1)
=
d
iag
[
u
0
(
l
-
1)
1
Q
,
u
1
(
l
-
1)
1
Q
,..,
u
M
-
1
(
l
-
1)
1
Q
],
(
1
5
)
wh
er
ein
1
Q
is
a
r
o
w
v
ec
to
r
o
f
Q
o
n
es.
T
h
e
t
th
clu
s
ter
h
as
th
e
g
ain
elem
en
t
u
t
(
l
-
1
)
an
d
is
g
iv
en
b
y
(
1
0
)
.
Alth
o
u
g
h
C
S
-
I
PAPA
d
em
o
n
s
tr
ates
b
etter
tr
ac
k
in
g
an
d
co
n
v
er
g
en
ce
r
ates
[
1
6
]
co
m
p
ar
ed
to
MI
PAPA,
I
PAPA,
an
d
B
S
-
I
PA
PA
,
it
d
o
es
n
o
t
co
n
s
id
er
p
ast
p
r
o
p
o
r
tio
n
ate
elem
en
ts
wh
en
u
p
d
atin
g
ea
ch
ad
ap
tiv
e
f
ilter
co
ef
f
icien
t,
r
ely
i
n
g
s
o
lely
o
n
t
h
e
cu
r
r
e
n
t tim
e
s
tep
.
3.
M
E
T
H
O
D
T
h
e
m
em
o
r
y
C
S
-
I
PAPA
is
in
tr
o
d
u
ce
d
in
th
is
s
ec
tio
n
,
b
y
e
x
ten
d
in
g
th
e
co
n
ce
p
t
o
f
m
em
o
r
y
o
f
th
e
p
r
o
p
o
r
tio
n
ate
ele
m
en
ts
to
t
h
e
C
S
-
I
PAPA.
B
ec
au
s
e
th
e
alg
o
r
ith
m
APA
co
n
s
id
er
s
th
e
h
is
to
r
y
o
f
p
ast
‘
p
’
m
o
m
en
ts
o
f
p
r
o
p
o
r
tio
n
ate
ele
m
en
ts
,
it
ca
n
b
e
co
n
s
id
er
ed
as
an
ad
a
p
tiv
e
alg
o
r
ith
m
with
m
em
o
r
y
,
th
e
h
is
to
r
y
o
f
th
e
last
‘
p
’
p
r
o
p
o
r
tio
n
ate
elem
en
ts
is
tak
en
in
to
ac
co
u
n
t
f
o
r
u
p
d
atin
g
ea
ch
f
ilter
co
ef
f
icien
t.
R
ec
u
r
s
iv
e
im
p
lem
en
tatio
n
o
f
th
e
p
r
o
p
o
r
t
io
n
ate
elem
en
ts
ca
n
b
e
ac
h
iev
ed
u
s
in
g
th
is
ap
p
r
o
ac
h
.
T
h
is
ap
p
r
o
ac
h
lead
s
to
a
s
ig
n
if
ican
t
r
ed
u
ctio
n
in
th
e
co
m
p
u
tatio
n
al
co
m
p
lex
ity
in
ter
m
s
o
f
m
u
ltip
licatio
n
s
f
o
r
h
ig
h
er
v
alu
es
o
f
p
r
o
jectio
n
o
r
d
e
r
[
1
2
]
.
T
h
e
tech
n
iq
u
e
e
m
p
lo
y
ed
in
t
h
is
r
esear
ch
wo
r
k
is
in
co
r
p
o
r
atin
g
m
e
m
o
r
y
to
p
r
o
p
o
r
tio
n
ate
elem
e
n
ts
o
f
C
S
-
I
PAPA th
at
ca
n
im
p
r
o
v
e
its
co
n
v
er
g
e
n
ce
p
er
f
o
r
m
a
n
ce
.
T
h
u
s
,
th
e
p
r
o
p
o
s
ed
MCS
-
I
P
APA
is
d
er
iv
ed
b
y
f
ir
s
t
s
tar
tin
g
with
th
e
o
p
tim
izatio
n
p
r
o
b
lem
,
th
en
d
er
iv
in
g
th
e
f
ilter
u
p
d
atin
g
e
q
u
atio
n
o
f
th
e
C
S
-
I
PAPA
f
r
o
m
th
e
b
asis
-
p
u
r
s
u
it
an
d
th
e
m
eth
o
d
o
f
L
ag
r
an
g
e
m
u
ltip
lier
s
.
T
h
en
,
i
n
tr
o
d
u
cin
g
th
e
co
n
ce
p
t
o
f
m
em
o
r
y
in
t
o
th
e
clu
s
ter
-
s
p
ar
s
e
ch
a
n
n
el
f
av
o
r
s
p
e
r
f
o
r
m
an
ce
ch
ar
ac
ter
is
tics
im
p
r
o
v
em
en
t a
n
d
a
r
e
d
u
ctio
n
i
n
n
u
m
b
er
o
f
m
u
ltip
licatio
n
s
.
T
h
e
C
S
-
I
PAPA seek
s
an
o
p
tim
u
m
s
o
lu
tio
n
f
o
r
t
h
e
f
o
llo
win
g
o
p
tim
izatio
n
p
r
o
b
lem
.
m
in
i(
)
1
2
‖
i(
)
-
i(
-
1
)
‖
U
C
I
-
1
(
l
-
1)
2
s
.
t
d(
l
)
-
X
T
(
l
)
i(
l
)
=0
(
1
6
)
wh
er
e
U
CI
-
1
(
l
-
1
)
is
d
ef
in
ed
b
y
(
1
5
)
.
T
h
e
ab
o
v
e
o
p
tim
izatio
n
(
1
6
)
is
b
ased
o
n
th
e
c
o
n
ce
p
t
p
r
o
p
o
s
ed
in
[
2
6
]
,
[
2
7
]
.
A
c
o
n
s
tr
ain
ed
o
p
tim
iza
tio
n
(
1
6
)
is
tr
an
s
f
o
r
m
e
d
in
to
an
u
n
c
o
n
s
tr
ain
ed
o
p
tim
izatio
n
p
r
o
b
lem
b
y
th
e
m
eth
o
d
o
f
L
ag
r
an
g
e
m
u
ltip
lie
r
s
[
2
6
]
,
[
2
7
]
with
m
an
y
co
n
s
tr
ain
ts
.
T
h
e
co
s
t f
u
n
ctio
n
o
f
th
e
C
S
-
I
PAPA,
i.e
.
J
(
l
)
is
J(
l
)
=
1
2
[
i(
l
)
-
i(
l
-
1)
]
T
U
CI
-
1
(
l
-
1
)
[
i(
l
)
-
i(
l
-
1
)
]
+[
d
(
l
)
-
X
T
(
l
)
i(
l
)
]
T
λ(
l
),
(
1
7
)
wh
er
e
λ(
l
)
=
[
λ
0
(
l
),
λ
1
(
l
)
,
.
.
.
λ
p
-
1
(
l
)
]
T
is
th
e
L
ag
r
an
g
e
m
u
ltip
lier
v
ec
to
r
an
d
[
i(
l
)
-
i(
l
-
1)
]
T
U
CI
-
1
(
l
-
1
)
[
i(
l
)
-
i(
l
-
1
)
]
d
en
o
tes
th
e
R
iem
an
n
ian
d
is
tan
ce
b
etw
ee
n
i(
l
)
an
d
i(
l
-
1
)
.
E
q
u
atin
g
th
e
f
ir
s
t d
er
iv
ativ
es
o
f
th
e
co
s
t f
u
n
ctio
n
t
o
ze
r
o
to
g
iv
e
,
∂
J
(
l
)
∂
i(
l
)
=
0,
(
1
8
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
A
mem
o
r
y
imp
r
o
ve
d
p
r
o
p
o
r
ti
o
n
a
te
a
ffin
e
p
r
o
jectio
n
a
lg
o
r
ith
m
fo
r
…
(
S
en
th
il Mu
r
u
g
a
n
B
o
o
p
a
la
n
)
4609
∂
J
(
l
)
∂
λ(
l
)
=
0,
(
1
9
)
Fro
m
(
1
8
)
,
th
e
u
p
d
atin
g
eq
u
atio
n
o
f
th
e
C
S
-
I
PAPA b
ec
o
m
es
,
i(
l)
=
i(
l
-
1
)
+
U
CI
(l
-
1
)
X
(
l)
λ(
l)
.
(
2
0
)
T
h
e
d
er
iv
ativ
e
in
(
1
9
)
g
iv
es
,
d(
l
)
-
X
T
(
l
)
i(
l
)
=
0.
(
2
1
)
As
in
[
1
7
]
,
to
f
u
r
th
er
r
ed
u
ce
th
e
co
m
p
u
tatio
n
al
co
m
p
lex
it
y
,
th
e
clu
s
ter
-
s
p
ar
s
e
f
ea
tu
r
e
o
f
th
e
ch
an
n
el
an
d
s
lid
in
g
win
d
o
w
tech
n
iq
u
e
ca
n
b
e
in
clu
d
ed
i
n
(
2
0
)
.
Fo
r
i
n
clu
s
io
n
o
f
clu
s
ter
-
s
p
ar
s
e
f
ea
tu
r
e
,
let
x
Q
(
l
-
tQ)
=
[
x
(
l
-
tQ,
x
(
l
-
tQ
-
1
)
,
.
.
,
x
(
l
-
tQ
-
Q+
1
)
]
T
,
with
t=
0
,
1
,
.
.
,
M
-
1
.
T
h
e
tQ
ter
m
in
d
icate
s
th
e
p
r
o
d
u
ct
o
f
t a
n
d
Q.
T
h
en
(
2
0
)
b
ec
o
m
es
i(
l
)
=
i(
l
-
1
)
+
P
CI
(
l
)
λ(
l
)
.
(
2
2
)
wh
er
e
P
CI
(
l
)
is
a
K
×
p
m
atr
ix
,
wh
o
s
e
elem
en
ts
ar
e
d
eter
m
in
ed
b
y
th
e
clu
s
ter
wis
e
p
r
o
d
u
ct
o
f
U
an
d
X.
T
h
e
s
u
b
s
cr
ip
t CI
o
n
P
CI
(
l
)
im
p
lies
C
S
-
I
PAP
A.
Pre
-
m
u
ltip
ly
in
g
(
2
2
)
b
y
X
T
(
l
)
to
g
et
X
T
(
l
)
i(
l
)
=
X
T
(
l
)
i(
l
-
1
)
+
X
T
(
l
)
P
CI
(
l
)
λ(
l
).
(
2
3
)
T
h
e
er
r
o
r
v
ec
to
r
e(
l
)
is
e(
l
)
=
d(
l
)
-
X
T
(
l
)
i(
l
-
1)
(
2
4
)
Usi
n
g
(
2
1
)
a
n
d
(
2
4
)
in
to
(
2
3
)
t
o
o
b
tain
λ
(
l
)
as
(
2
5
)
.
λ(
l
)
=
(
X
T
(
l
)P
CI
(
l
)
)
-
1
e(
l
)
.
(
2
5
)
Su
b
s
titu
tin
g
(
2
5
)
in
to
(
2
2
)
,
th
e
C
S
-
I
PAPA u
p
d
ates f
ilter
co
ef
f
icien
ts
as
(
2
6
)
.
i(
l)
=
i(
l
-
1
)
+
P
C
I
(
l)
(
X
T
(
l)
P
C
I
(
l)
)
-
1
e(
l)
.
(
2
6
)
I
n
tr
o
d
u
cin
g
th
e
p
ar
am
eter
s
lik
e
r
e
g
u
lar
izatio
n
p
ar
a
m
eter
δ
CI
an
d
c
o
n
v
e
r
g
en
ce
r
ate
µ
in
(
3
1
)
g
iv
es
co
n
tr
o
l
o
v
e
r
th
e
av
o
id
an
ce
o
f
n
u
m
e
r
ical
d
i
f
f
icu
lty
an
d
th
e
weig
h
t
v
ec
to
r
in
c
r
em
en
t,
r
esp
ec
tiv
ely
.
T
h
e
n
,
th
e
u
p
d
ate
eq
u
atio
n
o
f
C
S
-
I
PAPA
b
ec
o
m
es
i(
l
)
=
i(
l
-
1
)
+
μ
P
CI
(
l
)(
X
T
(
l
)
P
CI
(
l
)+
δ
CI
I
p
)
-
1
e(
l
)
(
2
7
)
Dir
ec
t
co
m
p
u
tatio
n
o
f
(
−
1
)
(
)
n
ee
d
s
p
K
m
u
ltip
licatio
n
s
.
B
y
m
ak
in
g
u
s
e
o
f
th
e
clu
s
ter
-
s
p
ar
s
e
f
ea
tu
r
e
o
f
th
e
ch
a
n
n
el,
we
ca
n
m
in
i
m
is
e
th
e
n
u
m
b
er
o
f
m
u
ltip
licatio
n
s
r
eq
u
ir
ed
,
esp
ec
ially
f
o
r
h
ig
h
er
v
alu
es
o
f
p
r
o
jectio
n
o
r
d
er
.
P
CI
(
l
)
ju
s
t r
e
q
u
ir
es (
Q+
p
-
1)
M
m
u
ltip
licati
o
n
s
.
As
in
[
1
2
]
,
[
1
7
]
,
th
e
h
is
to
r
y
o
f
p
r
o
p
o
r
tio
n
ate
ele
m
en
ts
c
an
b
e
in
clu
d
ed
to
f
u
r
th
e
r
d
e
cr
ea
s
e
th
e
co
m
p
u
tatio
n
al
c
o
m
p
lex
ity
o
f
C
S
-
I
PAP
A.
T
h
e
m
atr
ix
P
CI
(
l
)
is
s
elec
ted
in
ter
m
s
o
f
th
e
d
ia
g
o
n
al
m
a
t
r
i
x
U
t
(
l
-
1
)
w
i
t
h
t
=
0
,
1
,
⋯
,
M
-
1
.
C
h
o
o
s
i
n
g
U
_
t
(
l
-
1)
=
u
t
(
l
-
1
)
I
p
,
w
h
e
r
e
u
t
(
l
-
1)
is
th
e
g
ain
ter
m
f
o
r
th
e
t
th
clu
s
ter
f
r
o
m
(
1
5
)
.
E
x
p
an
d
i
n
g
P
CI
(
l
)
g
iv
es
(
3
.
3
2
)
,
wh
er
e
u
(
l
-
1
)
h
as
th
e
g
ain
ter
m
s
f
o
r
M
clu
s
ter
s
.
T
h
e
⊙
o
p
er
ato
r
d
e
n
o
tes
Had
am
ar
d
p
r
o
d
u
ct
o
r
elem
en
t
-
wis
e
m
u
ltip
licatio
n
,
i.e
,
b
⊙
c
=
[
b
(
1
)
c(
1
)
,
b
(
2
)
c(
2
)
,
⋯
,
b
(
K)
c(
K)
]
T
,
wh
er
e
b
an
d
c
ar
e
v
ec
to
r
s
o
f
len
g
t
h
K.
B
y
u
s
in
g
a
m
o
d
i
f
ied
g
ain
m
atr
ix
with
t
=
0
,
1
,
⋯
, M
-
1,
U
t
(
l
-
1
)
=
d
ia
g
[
u
t
(
l
-
1
)
,
u
t
(
l
-
2
)
,
…,
u
t
(
l
-
p
)
]
,
(
2
8
)
th
e
h
is
to
r
y
o
r
m
em
o
r
y
o
f
p
r
o
p
o
r
tio
n
ate
g
ain
elem
e
n
ts
o
f
M
clu
s
ter
s
ca
n
b
e
in
co
r
p
o
r
ate
d
in
th
e
C
S
-
I
PAP
A.
B
y
th
is
way
,
to
iter
ate
i(
l
)
to
i
(
l
+1
)
,
t
h
e
last
‘
p
’
p
r
o
p
o
r
tio
n
at
e
elem
en
ts
ar
e
co
n
s
id
er
e
d
.
T
h
u
s
,
th
e
m
atr
ix
P
CI
(
l
)
b
ec
o
m
es
P
CI
'
(
l
)
as sh
o
wn
in
th
e
(
3
0
)
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
5
,
Octo
b
e
r
20
25
:
4
6
0
5
-
4
6
1
9
4610
P
CI
(
l
)
=[
u
(
l
-
1)
⊙
x
Q
(
l
-
tQ
)
u(
l
-
1)
⊙
x
Q
(
l
-
tQ
-
1
)
⋯⋯
u(
l
-
1)
⊙
x
Q
(
l
-
tQ
-
p
+1
)
]
(
2
9
)
P
CI
′
(
l
)
=[
u
(
l
-
1
)
⊙
x
Q
(
l
-
tQ
)
u
(
l
-
2
)
⊙
x
Q
(
l
-
tQ
-
1
)
⋯⋯
u(
l
-
p)
⊙
x
Q
(
l
-
tQ
-
p
+1
)
]
(
3
0
)
3
.
1
.
P
r
o
po
s
ed
a
lg
o
rit
hm
T
h
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
MCS
-
I
PAPA
h
as
two
ad
v
an
ta
g
es
b
ec
au
s
e
o
f
th
is
m
o
d
if
icatio
n
.
Firstl
y
,
as
th
e
p
r
o
p
o
s
ed
MCS
-
I
PAPA
tak
es
in
to
ac
co
u
n
t
th
e
last
‘
p
’
p
r
o
p
o
r
tio
n
ate
elem
e
n
ts
,
th
e
c
o
n
v
er
g
en
ce
r
ate
an
d
tr
ac
k
in
g
im
p
r
o
v
e
as
‘
p
’
in
cr
ea
s
es.
An
o
th
er
ad
v
a
n
tag
e
is
th
at
th
e
co
m
p
u
tatio
n
al
c
o
m
p
lex
ity
o
f
th
e
p
r
o
p
o
s
ed
is
lo
wer
th
an
th
at
o
f
th
e
C
S
-
I
PAPA.
T
h
is
ad
v
an
tag
e
in
ter
m
s
o
f
co
m
p
u
tatio
n
al
co
m
p
lex
ity
is
s
h
o
wn
in
T
ab
le
1
.
T
h
e
eq
u
atio
n
o
f
P
CI
'
(
l
)
is
ex
p
r
ess
ed
in
r
ec
u
r
s
iv
e
r
ea
lizatio
n
as
(
3
1
)
,
P
CI
'
(
l
)
=
[
u
(
-
1
)
⊙
x
Q
(
l
-
t
Q
)
P
-
1
'
(
l
-
1)
]
(
3
1
)
w
h
er
e
in
th
e
m
at
r
ix
P
-
1
'
(
l
-
1)
is
,
P
-
1
'
(
l
-
1)
=
[
u
(
l
-
2)
⊙
x
Q
(
l
-
t
Q
-
1)
⋯
u(
l
-
p)
⊙
x
Q
(
l
-
t
Q
-
p
+1
)
]
.
B
y
s
u
b
s
titu
tin
g
P
CI
(
l
)
with
P
CI
'
(
l
)
in
(
2
7
)
,
t
h
e
u
p
d
ate
eq
u
atio
n
o
f
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
is
g
iv
en
b
y
i
(
l
)
=
i
(
l
-
1
)
+
μ
P
CI
'
(
l
)
(
X
T
(
l
)
P
CI
'
(
l
)
+
δ
CI
I
p
)
-
1
e(
l
)
(
3
2
)
T
h
e
m
atr
ix
P
-
1
'
(
l
-
1)
h
as
th
e
f
ir
s
t
p
-
1
co
lu
m
n
s
o
f
P
CI
'
(
l
-
1)
.
T
h
e
f
ir
s
t
p
-
1
co
lu
m
n
s
o
f
P
CI
'
(
l
-
1)
ca
n
b
e
u
tili
ze
d
d
ir
ec
tly
f
o
r
c
o
m
p
u
tin
g
th
e
m
atr
ix
P
CI
'
(
l
)
,
th
er
e
b
y
s
av
i
n
g
co
m
p
u
tatio
n
s
.
Fo
r
P
CI
'
(
l
-
1)
,
th
e
s
lid
in
g
win
d
o
w
tech
n
iq
u
e
ca
n
n
o
t
b
e
ap
p
lied
.
Fro
m
(
3
4
)
,
it
r
eq
u
ir
es
p
K
m
u
l
tip
licatio
n
s
,
to
f
in
d
P
CI
(
l
)
.
H
o
wev
er
,
f
r
o
m
(
3
0
)
,
t
o
co
m
p
u
te
P
CI
'
(
l
)
,
it
r
eq
u
ir
es
‘
K’
m
u
l
tip
licatio
n
s
o
n
ly
.
T
h
is
ad
v
an
t
ag
e
in
co
m
p
u
tatio
n
b
ec
o
m
es
s
ig
n
if
ican
t
f
o
r
h
ig
h
er
p
r
o
jectio
n
o
r
d
er
‘
p
’
.
T
h
u
s
,
f
r
o
m
t
h
e
co
s
t
f
u
n
ctio
n
o
f
th
e
C
S
-
I
PAPA,
th
e
u
p
d
ate
eq
u
atio
n
o
f
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
is
d
er
iv
ed
.
T
h
e
co
n
d
itio
n
f
o
r
m
ea
n
-
s
tab
ilit
y
is
d
er
iv
ed
in
t
h
e
n
ex
t sectio
n
.
4.
ST
A
B
I
L
I
T
Y
O
F
M
C
S
-
I
P
AP
A
T
h
e
s
tab
ilit
y
o
f
th
e
MCS
-
I
PAPA
is
s
tu
d
ied
in
th
is
s
ec
tio
n
.
s
tep
-
s
ize
µ
p
lay
s
a
m
ajo
r
r
o
le
to
en
s
u
r
e
th
e
co
n
v
e
r
g
en
ce
o
f
a
n
y
k
in
d
o
f
ad
a
p
tiv
e
alg
o
r
ith
m
.
T
h
e
h
ig
h
er
t
h
e
v
alu
e
o
f
t
h
e
s
tep
s
ize,
th
e
h
ig
h
er
th
e
p
o
s
s
ib
ilit
y
o
f
th
e
ad
ap
tiv
e
alg
o
r
ith
m
to
d
iv
er
g
e
f
r
o
m
th
e
o
p
tim
u
m
s
o
lu
tio
n
.
T
h
er
ef
o
r
e,
it
is
h
ig
h
ly
im
p
o
r
tan
t
to
s
tu
d
y
th
e
r
an
g
e
o
f
µ
t
h
at
c
o
n
f
ir
m
s
th
e
co
n
v
er
g
e
n
ce
in
t
h
e
m
ea
n
.
i.e
.
,
E
(
ĩ
(
)
)
→0
as
l
→
∞
.
T
h
is
s
ec
tio
n
f
in
d
s
th
e
r
a
n
g
e
o
f
s
tep
-
s
ize
th
at
en
s
u
r
es
s
tab
ilit
y
in
th
e
m
ea
n
.
T
h
e
weig
h
t
-
er
r
o
r
v
ec
t
o
r
is
s
tated
as
ĩ(
l
)
=
h
-
i(
l
)
.
I
n
ter
m
s
o
f
t
h
e
ĩ(
l
)
,
t
h
e
u
p
d
atin
g
eq
u
atio
n
o
f
th
e
MCS
-
I
PAPA,
i.e
.
,
(
3
2
)
b
ec
o
m
es
.
ĩ(
l
)
=ĩ(
l
-
1)
-
µ
P
CI
'
(
l
)(X
T
(
l
)
P
CI
'
(
l
)
+δ
CI
I
p
)
-
1
e(
l
).
(
3
3
)
Su
b
s
titu
tin
g
e
(
l
)
=
a(
l
)+
X
T
(
l
)
ĩ
(
l
-
1)
in
(
3
3
)
.
y
ield
s
ĩ
(
l
)
=
ĩ
(
l
-
1)
-
μ
P
CI
'
(
l
)(
X
T
(
l
)
P
CI
'
(
l
)+
δ
CI
I
p
)
-
1
×(
a(
l
)+
X
T
(
l
)
ĩ
(
l
-
1
)
)
(
3
4
)
T
h
e
f
o
llo
win
g
v
alid
ass
u
m
p
tio
n
s
ar
e
m
ad
e
to
m
ak
e
c
o
n
v
er
g
e
n
ce
an
aly
s
is
tr
ac
tab
le
[
2
8
]
–
[
3
0
]
.
Ass
u
m
p
tio
n
1
.
T
h
e
n
o
is
e
a
(
l
)
is
ass
u
m
ed
to
b
e
a
ze
r
o
-
m
ea
n
W
GN.
Ass
u
m
p
tio
n
2
.
T
h
e
weig
h
t
-
e
r
r
o
r
v
ec
to
r
ĩ
(
l
)
,
th
e
in
p
u
t
v
ec
to
r
x
(
l
)
,
an
d
th
e
n
o
is
e
a
(
l
)
ar
e
s
tatis
tically
in
d
ep
en
d
en
t o
f
ea
ch
o
t
h
er
.
Ass
u
m
p
tio
n
3
.
T
h
e
alg
o
r
ith
m
co
n
v
er
g
es to
o
p
tim
u
m
in
th
e
m
ea
n
-
s
q
u
ar
e
s
en
s
e.
B
y
tak
in
g
s
tatis
tical
m
ea
n
o
r
e
x
p
ec
tatio
n
o
n
b
o
th
s
id
es o
f
(
3
4
)
to
g
i
v
e
E(
ĩ
(
l
)
)
=
E
(
ĩ
(
l
-
1
)
)
-
μ
E
(
(
P
CI
'
(
l
)(
X
T
(
l
)
P
CI
'
(
l
)+
δ
CI
I
p
)
-
1
a(
l
))
-
μ
E
(
(
P
CI
'
(
l
)(
X
T
(
l
)
P
CI
'
(
l
)+
δ
CI
I
p
)
-
1
X
T
(
l
)
ĩ
(
l
-
1
)
)
(
3
5
)
Usi
n
g
Ass
u
m
p
tio
n
2
in
(
3
5
)
,
t
h
e
ev
o
lu
tio
n
o
f
th
e
m
ea
n
o
f
t
h
e
weig
h
t
-
er
r
o
r
v
ec
to
r
is
g
iv
en
b
y
(
3
6
)
.
E(
ĩ
(
l
)
)
=(
I
-
μ
B
)
E
(
(
ĩ
(
l
-
1
)
)
,
(
3
6
)
wh
er
e
B=
E(P
CI
'
(
l
)(
X
T
(
l
)
P
CI
'
(
l
)+
δ
CI
I
p
)
-
1
X
T
(
l
)
)
.
E((
ĩ
(
l
-
1
)
)
ca
n
co
n
v
er
g
e
if
φ
(
I
-
μ
B)
<1
wh
er
ein
φ
(
.
)
r
ep
r
esen
ts
th
e
s
p
ec
tr
al
r
ad
iu
s
o
f
th
e
B
.
B
y
E
ig
en
v
alu
e
d
ec
o
m
p
o
s
itio
n
,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
A
mem
o
r
y
imp
r
o
ve
d
p
r
o
p
o
r
ti
o
n
a
te
a
ffin
e
p
r
o
jectio
n
a
lg
o
r
ith
m
fo
r
…
(
S
en
th
il Mu
r
u
g
a
n
B
o
o
p
a
la
n
)
4611
B=
W
Λ
W
T
,
(
3
7
)
wh
ich
lead
s
to
−
µ
=
(
−
µ
)
(
3
8
)
T
h
e
ab
o
v
e
eq
u
atio
n
b
ec
o
m
es
Φ(I
K
-
µB)=
Φ
(
I
K
-
µΛ)
<1
(
3
9
)
wh
er
e
W
is
th
e
eig
en
v
ec
to
r
a
n
d
Λ
is
a
d
iag
o
n
al
m
atr
i
x
h
av
i
n
g
th
e
eig
en
v
alu
es
o
f
R
.
Fro
m
(
3
9
)
,
it
ca
n
b
e
s
ee
n
th
at
th
e
co
n
v
er
g
en
ce
in
th
e
m
e
an
s
en
s
e
is
en
s
u
r
ed
o
r
g
u
ar
a
n
t
ee
d
f
o
r
th
e
f
o
llo
win
g
r
a
n
g
e
o
f
µ
.
0
<μ
<
2
φ
(
B
)
(
4
0
)
T
h
e
ab
o
v
e
co
n
d
itio
n
o
n
µ
in
(
4
0
)
g
iv
es
th
e
n
ec
ess
ar
y
an
d
s
u
f
f
icien
t
co
n
d
itio
n
f
o
r
th
e
p
r
o
p
o
s
ed
MCS
-
I
PAPA
to
b
e
s
tab
le.
5.
SI
M
UL
A
T
I
O
N
R
E
S
UL
T
S
AND
DIS
CUSS
I
O
N
Sev
er
al
s
im
u
latio
n
ex
p
er
im
e
n
ts
ar
e
ca
r
r
ied
o
u
t
to
e
v
alu
ate
th
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
MCS
-
I
PAPA.
W
h
il
e
p
er
f
o
r
m
in
g
s
im
u
latio
n
s
,
th
e
len
g
th
o
f
t
h
e
ad
ap
tiv
e
f
ilter
N
is
s
et
to
1
0
2
4
,
an
d
th
e
u
n
k
n
o
wn
s
y
s
tem
is
ass
u
m
ed
to
h
av
e
th
e
s
am
e
len
g
th
.
T
h
e
p
er
f
o
r
m
an
ce
o
f
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
is
co
m
p
ar
ed
with
th
at
o
f
ex
is
tin
g
alg
o
r
ith
m
s
lik
e
th
e
I
PAPA,
MI
PAPA,
B
S
-
I
PAPA,
an
d
th
e
C
S
-
I
PAPA
with
r
esp
ec
t
to
co
n
v
er
g
e
n
ce
r
ate,
NM
,
an
d
tr
ac
k
in
g
.
T
wo
d
if
f
e
r
en
t
ty
p
es
o
f
clu
s
ter
s
,
as
s
h
o
wn
in
Fig
u
r
e
2
,
ar
e
u
s
ed
.
Fig
u
r
e
2
(
a
)
is
a
s
in
g
le
c
lu
s
ter
wh
ich
h
as
3
2
n
o
n
-
ze
r
o
ac
tiv
e
tap
s
in
th
e
r
an
g
e
[
2
8
1
,
3
1
2
]
an
d
a
d
o
u
b
le
clu
s
ter
is
s
h
o
wn
in
Fig
u
r
e
2
(
b
)
,
wh
ich
h
as
6
4
n
o
n
-
ze
r
o
tap
s
,
3
2
ea
ch
in
th
e
r
a
n
g
es
[
2
8
1
,
3
1
2
]
an
d
[
7
9
3
,
8
2
4
]
.
Fo
r
th
e
n
etwo
r
k
ec
h
o
p
ath
a
n
d
s
atellite
ec
h
o
p
ath
,
th
e
im
p
u
ls
e
r
esp
o
n
s
e
is
r
eg
ar
d
ed
as
a
s
in
g
le
clu
s
ter
an
d
a
d
o
u
b
le
clu
s
ter
s
p
ar
s
e
s
y
s
tem
,
r
esp
ec
tiv
ely
.
I
n
all
th
e
e
x
p
e
r
im
en
ts
,
to
s
tu
d
y
th
e
tr
ac
k
in
g
b
eh
a
v
io
r
o
f
th
e
p
r
o
p
o
s
ed
MCS
-
I
PAPA,
f
ir
s
t,
a
s
in
g
le
clu
s
ter
i
s
u
tili
ze
d
a
n
d
th
en
ab
r
u
p
tly
,
th
e
s
im
u
latio
n
en
v
ir
o
n
m
en
t
is
s
h
if
ted
to
a
d
o
u
b
le
clu
s
ter
.
C
o
lo
r
ed
n
o
is
e,
wh
ite
g
au
s
s
ian
n
o
is
e
(
W
GN)
,
an
d
s
p
ee
ch
a
r
e
th
e
th
r
ee
d
if
f
e
r
en
t
in
p
u
t
s
ig
n
a
ls
th
at
ar
e
u
s
ed
in
th
e
s
im
u
latio
n
s
.
B
y
f
ilt
er
in
g
th
e
W
GN
th
r
o
u
g
h
a
s
y
s
t
em
o
f
f
ir
s
t
-
o
r
d
er
with
a
p
o
le
at
0
.
8
,
c
o
lo
r
ed
n
o
is
e
is
o
b
tain
ed
.
A
s
p
ee
ch
s
ig
n
al
i
n
r
ea
l
-
tim
e
s
am
p
led
at
8
k
Hz
is
u
tili
ze
d
.
An
i
n
d
ep
e
n
d
en
t
W
GN
with
a
s
ig
n
al
-
to
-
n
o
is
e
r
atio
,
SNR
=
3
0
d
B
i
s
ad
d
ed
to
th
e
b
ac
k
g
r
o
u
n
d
o
f
th
e
u
n
k
n
o
wn
s
y
s
tem
.
A
co
m
m
o
n
p
er
f
o
r
m
an
ce
m
etr
ic
f
o
r
a
d
ap
tiv
e
alg
o
r
ith
m
s
is
N
M
wh
ich
is
d
ef
in
e
d
as
1
0
lg
{(
‖
h
-
i‖
)
2
/
(
‖
h
‖
)
2
}
.
I
n
all
th
e
s
im
u
latio
n
s
,
th
e
en
tire
p
r
o
ce
s
s
is
d
iv
id
ed
in
to
two
h
a
lv
es,
th
e
f
ir
s
t
h
alf
allo
tted
f
o
r
th
e
p
er
f
o
r
m
an
ce
s
tu
d
y
o
f
a
s
i
n
g
le
clu
s
ter
an
d
th
e
s
ec
o
n
d
h
alf
f
o
r
th
e
d
o
u
b
le
cl
u
s
ter
.
I
n
th
e
f
o
llo
win
g
s
ec
tio
n
s
,
th
e
im
p
ac
t
o
f
v
ar
y
in
g
clu
s
ter
s
ize
‘
Q’
an
d
th
e
p
ar
am
eter
‘
β’
o
n
th
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
p
r
o
p
o
s
ed
is
s
tu
d
ied
.
T
h
en
th
e
p
er
f
o
r
m
an
ce
c
o
m
p
a
r
is
o
n
s
tu
d
y
is
d
o
n
e
b
y
co
m
p
ar
in
g
th
e
p
r
o
p
o
s
ed
MCS
-
I
PAP
A
with
th
e
ex
is
tin
g
alg
o
r
ith
m
s
in
th
e
liter
atu
r
e.
(
a)
(
b
)
Fig
u
r
e
2
.
C
lu
s
ter
-
s
p
ar
s
e
ch
an
n
el
with
its
ty
p
es (
a)
a
s
in
g
le
clu
s
ter
an
d
(
b
)
a
d
o
u
b
le
clu
s
ter
5
.
1
.
P
er
f
o
r
m
a
nce
curv
es
o
f
M
CS
-
I
P
AP
A
wit
h
diff
er
ent
β
W
ith
in
p
u
ts
lik
e
co
lo
r
ed
n
o
is
e,
W
GN
an
d
s
p
ee
ch
s
ig
n
al,
th
e
ef
f
ec
t o
f
v
a
r
y
in
g
‘
β’
o
n
th
e
p
e
r
f
o
r
m
a
n
ce
o
f
th
e
p
r
o
p
o
s
ed
MCS
-
I
PAPA
is
s
tu
d
ied
,
with
clu
s
ter
s
ize
Q,
s
et
to
2
.
L
o
w,
m
ed
iu
m
an
d
h
i
g
h
v
alu
es
o
f
β
ar
e
ch
o
s
en
as
i
n
[
1
6
]
.
Fig
u
r
e
3
s
h
o
ws
th
e
ef
f
ec
t
o
f
β
o
n
MCS
-
I
PAPA.
Fo
r
d
if
f
er
e
n
t
v
alu
es
o
f
β
lik
e
2
,
5
,
1
0
,
an
d
2
0
,
th
e
s
im
u
latio
n
r
esu
lts
ar
e
s
h
o
wn
in
Fig
u
r
e
s
3
(
a
)
a
n
d
3
(
b
)
f
o
r
co
l
o
r
ed
n
o
is
e
an
d
W
GN,
r
esp
ec
tiv
ely
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
5
,
Octo
b
e
r
20
25
:
4
6
0
5
-
4
6
1
9
4612
T
h
e
p
r
o
p
o
s
ed
MCS
-
I
PAPA
s
h
o
ws
a
r
e
d
u
ctio
n
in
NM
with
an
in
c
r
ea
s
e
in
β
f
o
r
s
in
g
le
a
n
d
d
o
u
b
l
e
clu
s
ter
s
y
s
tem
s
,
f
o
r
c
o
lo
r
ed
n
o
is
e
an
d
W
GN
in
p
u
ts
.
I
f
β
=
1
0
,
th
e
MCS
-
I
PAPA
s
h
o
ws
th
e
b
est
NM
f
o
r
b
o
th
th
e
in
p
u
ts
.
T
h
e
MCS
-
I
PAPA
s
h
o
ws
h
ig
h
er
m
is
alig
n
m
en
t
f
o
r
o
th
er
v
alu
es
o
f
β.
B
u
t,
f
o
r
s
p
ee
ch
s
ig
n
al,
th
e
MCS
-
I
PAP
A
s
h
o
ws
a
d
ec
r
ea
s
e
in
NM
with
a
d
ec
r
ea
s
e
in
β
f
o
r
b
o
th
s
in
g
le
a
n
d
d
o
u
b
l
e
clu
s
ter
s
y
s
tem
s
,
attain
in
g
th
e
m
in
im
u
m
NM
f
o
r
β
=
2
.
So
,
f
o
r
th
e
p
e
r
f
o
r
m
an
ce
co
m
p
a
r
is
o
n
o
f
th
e
p
r
o
p
o
s
ed
with
th
e
alg
o
r
ith
m
s
lik
e
th
e
I
PAPA,
MI
PAPA,
B
S
-
I
PAPA,
an
d
th
e
C
S
-
I
PAPA,
t
h
e
p
ar
am
eter
β is
s
elec
ted
as 1
0
f
o
r
c
o
lo
r
ed
n
o
is
e
an
d
W
GN,
β as 2
f
o
r
th
e
s
p
ee
ch
s
ig
n
al.
(
a)
(
b
)
Fig
u
r
e
3
.
E
f
f
ec
t o
f
β
with
d
if
f
er
en
t v
alu
es o
n
MCS
-
I
PAPA (
μ
=
0
.
0
1
; SNR
=
3
0
d
B
; p
=
2)
(
a)
co
lo
r
ed
i
n
p
u
t
a
n
d
(
b
)
W
GN
in
p
u
t
5
.
2
.
P
er
f
o
rma
nce
curv
es o
f
M
CS
-
I
P
AP
A
wit
h diff
er
ent
clus
t
er
s
ize
Q
Var
y
in
g
th
e
clu
s
ter
s
ize
Q
ca
n
af
f
ec
t
NM
.
L
o
w,
m
e
d
iu
m
a
n
d
h
ig
h
v
alu
es
o
f
Q
ar
e
c
h
o
s
en
as
in
[
1
6
]
.
C
lu
s
ter
s
o
f
d
if
f
er
en
t
s
izes
o
f
Q
=
2
,
4
,
8
,
1
6
ar
e
s
elec
ted
to
s
tu
d
y
an
d
ev
alu
ate
th
e
e
f
f
ec
t
o
f
Q
o
n
th
e
p
er
f
o
r
m
an
ce
ch
a
r
ac
ter
is
tics
o
f
th
e
MCS
-
I
PAPA.
Fo
r
th
ese
s
tu
d
ies,
th
e
p
ar
am
eter
β
is
s
et
to
a
v
alu
e
o
f
1
0
f
o
r
b
o
th
c
o
lo
r
ed
n
o
is
e
an
d
W
GN,
2
f
o
r
s
p
ee
c
h
in
p
u
t.
Fig
u
r
e
4
s
h
o
ws
th
e
ef
f
ec
t
o
f
v
ar
y
i
n
g
Q
o
n
MCS
-
I
PAPA.
T
h
e
r
esu
lts
ar
e
s
h
o
wn
in
Fig
u
r
es
4
(
a
)
an
d
4
(
b
)
f
o
r
co
lo
r
ed
n
o
is
e
an
d
W
GN
in
p
u
t,
r
esp
ec
tiv
ely
.
(
a)
(
b
)
Fig
u
r
e
4
.
E
f
f
ec
t o
f
Q
with
d
if
f
er
en
t v
alu
es o
n
MCS
-
I
PAPA (
μ
=
0
.
0
1
; SNR
=
3
0
d
B
; p
=
2)
(
a
)
co
lo
r
ed
i
n
p
u
t
(
p
o
le
at
0
.
8
)
a
n
d
(
b
)
W
GN
in
p
u
t
Fig
u
r
e
5
s
h
o
ws
th
e
p
ar
am
eter
v
ar
iatio
n
o
n
MCS
-
PAPA
w
ith
s
p
ee
ch
in
p
u
t.
Fig
u
r
e
s
5
(
a
)
an
d
5
(
b
)
d
ep
ict
th
e
ef
f
ec
t
o
f
β
an
d
Q
o
n
MCS
-
I
PAPA,
r
esp
ec
tiv
ely
.
T
h
e
NM
o
f
th
e
p
r
o
p
o
s
ed
MCS
-
I
PAPA
d
ec
r
ea
s
es
with
clu
s
ter
s
ize,
f
o
r
all
th
r
ee
in
p
u
ts
,
f
o
r
b
o
th
s
in
g
le
an
d
d
o
u
b
le
clu
s
ter
s
.
T
h
e
MCS
-
I
PAPA
s
h
o
ws
th
e
least
NM
f
o
r
Q
=
2.
T
h
u
s
,
th
e
two
p
ar
am
eter
s
‘
β’
a
n
d
‘
Q’
r
esu
lt
in
th
e
p
er
f
o
r
m
an
ce
im
p
r
o
v
e
m
en
t
o
f
th
e
MCS
-
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
A
mem
o
r
y
imp
r
o
ve
d
p
r
o
p
o
r
ti
o
n
a
te
a
ffin
e
p
r
o
jectio
n
a
lg
o
r
ith
m
fo
r
…
(
S
en
th
il Mu
r
u
g
a
n
B
o
o
p
a
la
n
)
4613
I
PAPA
in
ter
m
s
o
f
NM
.
B
y
s
e
lectin
g
s
u
itab
le
v
alu
es
to
β
a
n
d
th
e
clu
s
ter
s
ize
Q,
th
e
MCS
-
I
PAPA
is
s
h
o
wn
to
o
u
tp
er
f
o
r
m
t
h
e
co
m
p
etin
g
a
lg
o
r
ith
m
s
lik
e
th
e
I
PAPA,
MI
PAPA,
B
S
-
I
PA
PA,
an
d
t
h
e
C
S
-
I
PAPA
f
o
r
id
en
tify
in
g
t
h
e
s
in
g
le
o
r
d
o
u
b
l
e
clu
s
ter
ch
an
n
els.
(
a)
(
b
)
Fig
u
r
e
5.
E
f
f
ec
t o
f
d
if
f
e
r
en
t β
an
d
Q
o
n
MCS
-
I
PAPA with
s
p
ee
ch
in
p
u
t
(μ
=
0
.
0
2
; SNR
=
3
0
d
B
; p
=
2)
:
(
a)
v
ar
y
in
g
β
an
d
(
b
)
v
ar
y
in
g
Q
5
.
3
.
P
er
f
o
r
m
a
nce
curv
es
o
f
M
CS
-
I
P
AP
A
a
g
a
ins
t
ex
is
t
ing
a
lg
o
rit
h
m
s
Fig
u
r
e
6
s
h
o
ws
th
e
p
er
f
o
r
m
a
n
ce
cu
r
v
es
o
f
MCS
-
I
PAPA
an
d
o
th
er
al
g
o
r
ith
m
s
.
T
h
e
p
e
r
f
o
r
m
a
n
ce
ev
alu
atio
n
s
ar
e
illu
s
tr
ated
in
Fig
u
r
es
6
(
a
)
an
d
6
(
b
)
,
with
c
o
lo
r
ed
n
o
is
e
in
p
u
t
an
d
W
GN
in
p
u
t,
r
esp
ec
tiv
ely
.
W
ith
s
p
ee
ch
in
p
u
t,
Fig
u
r
e
7
d
ep
icts
th
e
p
er
f
o
r
m
an
ce
c
u
r
v
es
o
f
MCS
-
I
PAPA
an
d
o
th
er
alg
o
r
ith
m
s
.
Fo
r
SNR
s
3
0
d
B
an
d
1
5
d
B
,
th
e
p
er
f
o
r
m
an
ce
ev
alu
atio
n
p
l
o
ts
ar
e
s
h
o
wn
in
Fig
u
r
es
7
(
a
)
an
d
7
(
b
)
r
esp
ec
tiv
ely
.
T
h
e
r
esu
lts
o
f
s
im
u
latio
n
tr
ials
th
at
ar
e
in
d
ep
en
d
en
tly
r
e
p
ea
ted
1
5
tim
es a
r
e
en
s
em
b
le
av
er
ag
e
d
to
o
b
tain
th
e
p
lo
ts
.
Fro
m
th
e
o
b
tain
e
d
p
l
o
ts
,
it
ca
n
b
e
s
ee
n
th
at
th
e
MCS
-
I
PAPA
is
s
h
o
win
g
h
ig
h
er
p
e
r
f
o
r
m
a
n
ce
th
a
n
th
e
alg
o
r
ith
m
s
lik
e
th
e
M
I
PAPA,
I
PAPA,
B
S
-
I
PAP
A,
an
d
t
h
e
C
S
-
I
PAPA.
Fo
r
th
e
c
o
lo
r
ed
n
o
i
s
e
an
d
W
GN
in
p
u
t,
th
e
p
ar
am
eter
β
is
m
ain
tain
ed
at
1
0
,
f
o
r
th
e
MCS
-
I
PAPA
a
n
d
β
is
s
et
to
2
f
o
r
th
e
s
p
ee
c
h
in
p
u
t.
Fo
r
all
th
e
th
r
ee
clu
s
ter
o
r
b
lo
c
k
alg
o
r
ith
m
s
n
am
ely
th
e
MCS
-
I
PAPA,
C
S
-
I
PAP
A,
an
d
th
e
B
S
-
I
PAP
A,
th
e
s
am
e
clu
s
ter
s
ize
o
f
Q
=
2
is
ass
ig
n
ed
.
T
h
e
s
tep
-
s
ize
f
o
r
ea
ch
alg
o
r
ith
m
is
s
et
to
µ
=
0
.
0
1
f
o
r
c
o
lo
r
ed
n
o
is
e
an
d
W
GN
in
p
u
t,
b
u
t
µ
is
s
et
to
0
.
0
2
f
o
r
s
p
ee
ch
s
ig
n
al
in
p
u
t.
T
h
e
r
e
g
u
lar
izatio
n
p
ar
am
ete
r
f
o
r
I
PAPA
is
ch
o
s
en
as
δ
=
0
.
0
1
.
T
h
e
r
eg
u
lar
izatio
n
p
ar
a
m
eter
is
s
et
to
0
.
0
1
f
o
r
th
e
o
th
er
f
o
u
r
al
g
o
r
ith
m
s
as
well.
T
h
e
p
ar
am
et
er
α
is
ch
o
s
en
as
0
.
W
ith
β
=
1
0
an
d
Q
=
2
,
th
e
p
r
o
p
o
s
ed
MCS
-
I
PAPA,
f
o
r
co
lo
r
e
d
n
o
is
e
an
d
W
GN
in
p
u
ts
,
s
h
o
ws
b
etter
tr
ac
k
in
g
,
less
er
NM
th
an
th
e
ex
is
tin
g
alg
o
r
ith
m
s
.
Fig
u
r
es
6
(
a
)
an
d
6
(
b
)
s
h
o
w
th
at
th
e
MCS
-
I
PAPA
s
h
o
ws
h
ig
h
er
co
n
v
er
g
en
ce
r
ate,
b
etter
tr
ac
k
i
n
g
,
less
er
NM
,
th
an
th
e
e
x
is
tin
g
alg
o
r
ith
m
s
.
(
a)
(
b
)
Fig
u
r
e
6
.
Per
f
o
r
m
an
c
e
cu
r
v
es o
f
MCS
-
I
PAPA a
n
d
o
th
er
alg
o
r
ith
m
s
(
μ
=
0
.
0
1
; SNR
=
3
0
d
B
;
p
=
2)
:
(
a)
co
lo
r
e
d
in
p
u
t (
p
o
le
at
0
.
8
)
(
b
)
W
GN
in
p
u
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
5
,
Octo
b
e
r
20
25
:
4
6
0
5
-
4
6
1
9
4614
Fro
m
Fig
u
r
e
s
7
(
a
)
an
d
7
(
b
)
,
it
is
s
h
o
wn
th
at
th
e
M
C
S
-
I
PA
P
A
s
h
o
ws
h
ig
h
er
co
n
v
er
g
e
n
ce
r
ate,
b
etter
tr
ac
k
in
g
,
a
n
d
less
er
NM
,
th
a
n
th
e
C
S
-
I
PAPA
f
o
r
s
p
ee
ch
s
ig
n
al
in
p
u
t
as
well.
I
n
Fig
u
r
e
7
(
b
)
,
im
p
r
o
v
em
e
n
t
in
NM
is
2
0
%
an
d
2
9
%,
s
h
o
wn
b
y
MCS
-
I
PAPA
o
v
er
th
e
C
S
-
I
PAPA,
f
o
r
s
in
g
le
clu
s
ter
in
[
1
0
0
0
,
6
0
0
0
]
an
d
d
o
u
b
le
clu
s
ter
in
th
e
iter
atio
n
r
an
g
e
[
4
1
0
0
0
,
4
7
0
0
0
]
,
r
esp
e
ctiv
ely
.
Fro
m
th
e
ab
o
v
e
d
is
c
u
s
s
ed
an
aly
s
es,
th
e
ch
o
s
en
p
ar
am
eter
s
o
f
β
an
d
Q
s
h
o
w
h
ig
h
er
p
er
f
o
r
m
a
n
ce
f
o
r
th
e
MCS
-
I
PAPA
in
ter
m
s
o
f
co
n
v
er
g
en
ce
r
ate,
NM
,
an
d
tr
ac
k
in
g
.
Hen
ce
,
th
e
s
im
u
latio
n
p
lo
ts
p
r
o
v
e
th
at
th
e
p
r
o
p
o
s
ed
MCS
-
I
PAPA
o
u
tp
e
r
f
o
r
m
s
th
e
ex
is
tin
g
alg
o
r
ith
m
s
lik
e
I
PAPA,
MI
PAPA,
B
S
-
I
PA
PA,
an
d
C
S
-
I
PAP
A
f
o
r
s
in
g
le
an
d
d
o
u
b
le
clu
s
te
r
ch
an
n
els.
(
a)
(
b
)
Fig
u
r
e
7
.
Per
f
o
r
m
an
c
e
cu
r
v
es o
f
MCS
-
I
PAPA a
n
d
o
th
er
alg
o
r
ith
m
s
f
o
r
s
p
ee
c
h
in
p
u
t (
μ
=
0
.
0
2
;
p
=
2)
:
(
a)
SNR
=3
0
d
B
an
d
(
b
)
SNR
=1
5
d
B
5
.
4
.
P
er
f
o
r
m
a
nce
curv
es f
o
r
hig
her
pro
j
ec
t
io
n o
rder
a
nd
dis
per
s
iv
e
s
y
s
t
em
s
I
n
th
is
s
ec
tio
n
,
a
s
tu
d
y
is
ex
te
n
d
ed
f
o
r
h
ig
h
e
r
p
r
o
jectio
n
o
r
d
er
an
d
d
is
p
er
s
iv
e
s
y
s
tem
.
Fig
u
r
e
8
s
h
o
ws
th
e
p
er
f
o
r
m
a
n
ce
cu
r
v
es
o
f
al
g
o
r
ith
m
s
f
o
r
h
ig
h
er
p
r
o
jectio
n
o
r
d
er
a
n
d
d
is
p
er
s
iv
e
s
y
s
tem
.
W
ith
s
p
ee
ch
as
in
p
u
t,
Fig
u
r
e
s
8
(
a
)
an
d
8
(
b
)
s
h
o
w
th
e
p
e
r
f
o
r
m
an
ce
c
o
m
p
ar
is
o
n
o
f
th
e
MCS
-
I
PAPA,
o
v
er
t
h
e
C
S
-
I
PAPA,
an
d
th
e
MI
PAPA
f
o
r
h
ig
h
er
p
r
o
je
ctio
n
o
r
d
er
a
n
d
d
is
p
er
s
iv
e
s
y
s
tem
,
r
esp
ec
tiv
ely
.
B
y
en
s
em
b
le
av
er
ag
in
g
o
f
1
0
in
d
ep
en
d
en
t
tr
ials
,
th
e
s
im
u
latio
n
p
lo
ts
ar
e
o
b
tain
ed
.
T
h
e
I
P
APA
b
eh
av
es
alm
o
s
t
lik
e
an
NL
MS
alg
o
r
ith
m
f
o
r
=
2
.
So
,
it is
h
ig
h
ly
n
ec
ess
ar
y
to
ch
ec
k
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
p
r
o
p
o
s
ed
MCS
-
I
PAPA f
o
r
h
ig
h
er
p
r
o
jectio
n
o
r
d
er
.
(
a)
(
b
)
Fig
u
r
e
8.
Per
f
o
r
m
an
c
e
cu
r
v
es o
f
MCS
-
I
PAPA,
C
S
-
I
P
APA
a
n
d
MI
PAPA f
o
r
s
p
ee
ch
i
n
p
u
t
:
(
a)
h
ig
h
er
p
r
o
jectio
n
o
r
d
er
(
p
=
6
,
ξ=0
.
8
7
5
,
μ
=0
.
0
2
)
an
d
(
b
)
d
is
p
er
s
iv
e
s
y
s
tem
(
ξ=0
.
4
8
4
,
μ
=
0
.
0
5
,
p
=2
)
6.
CO
M
P
UT
AT
I
O
NAL
CO
M
P
L
E
X
I
T
Y
I
n
T
ab
le
1
,
th
e
co
m
p
u
tatio
n
al
co
m
p
lex
ity
o
f
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
is
co
m
p
ar
e
d
ag
ain
s
t
d
if
f
er
en
t
alg
o
r
ith
m
s
n
am
el
y
th
e
I
PAPA,
MI
PAPA,
B
S
-
I
PAPA,
C
S
-
I
PAPA
in
ter
m
s
o
f
th
e
to
tal
n
u
m
b
er
o
f
ad
d
itio
n
s
Evaluation Warning : The document was created with Spire.PDF for Python.