T
E
L
KO
MNIK
A
, V
ol
.
17
,
No.
6,
Dec
e
mb
er
20
1
9,
p
p.2
78
2~
27
89
IS
S
N: 1
69
3
-
6
93
0
,
accr
ed
ited
F
irst
Gr
ad
e b
y K
em
en
r
istekdikti,
Decr
ee
No: 2
1/E/
K
P
T
/20
18
DOI:
10.12928/TE
LK
OM
N
IK
A
.v
1
7
i
6
.
12802
◼
27
82
Rec
ei
v
ed
Dec
e
mb
er
20
,
20
1
8
; R
ev
i
s
ed
Ma
y
2
5
, 2
01
9
;
A
c
c
ep
ted
J
ul
y
2
,
20
1
9
A w
irele
ss prec
od
in
g t
e
ch
niq
u
e fo
r
millime
tre
-
w
ave
MIM
O
sys
te
m b
ased o
n SIC
-
M
MS
E
Ro
u
n
ak
u
l Is
l
am Bo
b
y
*
1
, K
h
aiz
u
r
an Ab
d
u
lla
h
2
,
A.
Z
.
Juso
h
3
,
Nagm
a
P
ar
ve
en
4
,
A.
L
.
As
n
aw
i
5
El
e
c
tr
i
c
a
l
a
n
d
Co
m
p
u
t
e
r En
g
i
n
e
e
ri
n
g
,
Ku
l
l
i
y
y
a
h
o
f
En
g
i
n
e
e
r
i
n
g
a
n
d
I
n
te
rn
a
t
i
o
n
a
l
I
s
l
a
m
i
c
Un
i
v
e
r
s
i
t
y
M
a
l
a
y
s
i
a
,
G
o
m
b
a
k
,
Se
l
a
n
g
o
r
,
M
a
l
a
y
s
i
a
,
te
l
p
:
(
+6
0
3
)
6
1
9
6
4
0
0
0
.
*C
o
rre
s
p
o
n
d
i
n
g
a
u
th
o
r,
e
-
m
a
i
l
:
ro
u
n
a
q
u
l
2
0
2
0
@
g
m
a
i
l
.
c
o
m
1
,
k
h
a
i
z
u
ra
n
@i
i
u
m
.e
d
u
.m
y
2
,
a
z
a
m
a
n
i
@
i
i
u
m
.e
d
u
.
m
y
3
Ab
strac
t
A
c
o
m
m
u
n
i
c
a
t
i
o
n
m
e
th
o
d
i
s
p
ro
p
o
s
e
d
u
s
i
n
g
M
i
n
i
m
u
m
M
e
a
n
Sq
u
a
re
Erro
r
(M
M
SE)
p
re
c
o
d
i
n
g
a
n
d
Su
c
c
e
s
s
i
v
e
I
n
te
rf
e
re
n
c
e
Can
c
e
l
l
a
ti
o
n
(SIC)
te
c
h
n
i
q
u
e
fo
r
m
i
l
l
i
m
e
tre
-
wa
v
e
m
u
l
ti
p
l
e
-
i
n
p
u
t
m
u
l
ti
p
l
e
-
o
u
t
p
u
t
(m
m
-
Wa
v
e
M
IM
O
)
b
a
s
e
d
w
i
re
l
e
s
s
c
o
m
m
u
n
i
c
a
ti
o
n
s
y
s
te
m
.
Th
e
m
m
-
Wa
v
e
M
IM
O
te
c
h
n
o
l
o
g
y
fo
r
wir
e
l
e
s
s
c
o
m
m
u
n
i
c
a
ti
o
n
s
y
s
te
m
i
s
t
h
e
b
a
s
e
p
o
te
n
t
i
a
l
t
e
c
h
n
o
l
o
g
y
fo
r
i
ts
h
i
g
h
d
a
ta
tra
n
s
f
e
r
ra
te
fo
l
l
o
we
d
b
y
d
a
ta
i
n
s
tr
u
c
t
i
o
n
a
n
d
l
o
w
p
o
wer
c
o
n
s
u
m
p
ti
o
n
c
o
m
p
a
re
d
t
o
L
o
n
g
-
T
e
rm
Ev
o
l
u
ti
o
n
(L
TE)
.
Th
e
m
m
-
Wa
v
e
s
y
s
te
m
i
s
a
l
re
a
d
y
a
v
a
i
l
a
b
l
e
i
n
i
n
d
o
o
r
h
o
t
s
p
o
t
a
n
d
W
i
-
Fi
b
a
c
k
h
a
u
l
fo
r
i
ts
h
i
g
h
b
a
n
d
wi
d
t
h
a
v
a
i
l
a
b
i
l
i
t
y
a
n
d
p
o
te
n
t
i
a
l
l
e
a
d
to
ra
te
o
f
n
u
m
e
r
o
u
s
G
b
p
s
/
u
s
e
r.
B
u
t,
i
n
m
o
b
i
l
e
wir
e
l
e
s
s
c
o
m
m
u
n
i
c
a
ti
o
n
s
y
s
te
m
th
i
s
te
c
h
n
i
q
u
e
i
s
l
a
g
g
i
n
g
b
e
c
a
u
s
e
th
e
c
h
a
n
n
e
l
fa
c
e
s
re
l
a
ti
v
e
o
r
th
o
g
o
n
a
l
c
o
o
rd
i
n
a
ti
o
n
a
n
d
m
u
l
t
i
p
l
e
n
o
d
e
d
e
te
c
ti
o
n
p
ro
b
l
e
m
s
whi
l
e
ra
p
i
d
m
o
v
e
m
e
n
t
o
f
n
o
d
e
s
(tra
n
s
m
i
tt
e
r
a
n
d
re
c
e
i
v
e
r)
o
c
c
u
r.
To
i
m
p
ro
v
e
th
e
c
o
n
v
e
n
ti
o
n
a
l
m
m
-
wav
e
M
IM
O
n
o
d
a
l
d
e
te
c
ti
o
n
a
n
d
c
o
o
r
d
i
n
a
t
i
o
n
p
e
r
fo
rm
a
n
c
e
,
th
e
s
y
s
te
m
p
ro
c
e
s
s
e
s
d
a
ta
u
s
i
n
g
s
y
m
b
o
l
i
z
e
d
e
rro
r
v
e
c
to
r
te
c
h
n
i
q
u
e
fo
r
l
i
n
e
a
ri
z
a
ti
o
n
.
T
h
e
n
t
h
e
M
M
SE
p
re
c
o
d
i
n
g
d
e
te
c
ti
o
n
te
c
h
n
i
q
u
e
i
m
p
ro
v
e
s
t
h
e
l
i
n
k
s
tre
n
g
th
b
y
c
o
n
s
ta
n
t
l
y
fi
tt
i
n
g
t
h
e
c
h
a
n
n
e
l
c
o
e
ff
i
c
i
e
n
ts
b
a
s
e
d
o
n
n
u
m
b
e
r
o
f
i
n
d
e
p
e
n
d
e
n
t
s
e
rv
i
c
e
a
n
te
n
n
a
s
(M
),
Si
g
n
a
l
to
No
i
s
e
Ra
ti
o
(S
NR
),
Cha
n
n
e
l
M
a
tri
x
(CM
)
a
n
d
m
e
a
n
s
q
u
a
re
e
rro
rs
(M
S
E).
To
m
a
i
n
ta
i
n
s
e
q
u
e
n
t
i
a
l
l
y
e
n
c
o
d
e
d
u
s
e
r
d
a
ta
c
o
n
n
e
c
t
i
v
i
ty
a
n
d
to
o
v
e
r
c
o
m
e
d
a
ta
l
o
s
s
,
SIC
m
e
th
o
d
i
s
u
s
e
d
i
n
c
o
m
b
i
n
a
ti
o
n
wi
th
M
M
SE.
M
ATL
AB
was
u
s
e
d
to
v
a
l
i
d
a
te
th
e
p
ro
p
o
s
e
d
s
y
s
te
m
p
e
rf
o
rm
a
n
c
e
.
Key
w
ords
:
c
h
a
n
n
e
l
m
a
tri
x
,
m
i
l
l
i
m
e
tr
e
-
wav
e
,
m
i
n
i
m
u
m
m
e
a
n
s
q
u
a
re
e
rro
r
,
q
u
a
n
t
i
z
e
d
s
y
s
t
e
m
,
s
u
c
c
e
s
s
i
v
e
i
n
te
rf
e
re
n
c
e
-
c
a
n
c
e
l
l
a
ti
o
n
Copy
righ
t
©
2
0
1
9
Uni
v
e
rsi
t
a
s
Ahm
a
d
D
a
hl
a
n.
All
rig
ht
s
r
e
s
e
rve
d
.
1.
Int
r
o
d
u
ctio
n
T
he
c
om
m
un
i
c
ati
on
s
i
n
th
e
mi
l
l
i
m
etre
wav
e
ba
n
d
s
uff
ers
fr
om
i
nc
r
ea
s
ed
pa
t
h
l
os
s
ex
po
ne
nts
,
hi
gh
er
s
ha
do
w
fad
i
ng
,
b
l
oc
k
ag
e
an
d
pe
ne
tr
at
i
on
l
os
s
es
,
etc
.,
wh
e
r
e
s
ub
-
6
G
Hz
s
y
s
tem
s
l
ea
di
ng
to
a
po
orer
l
i
nk
m
argi
n
t
ha
n
l
e
ga
c
y
s
y
s
tem
s
[1
-
3].
S
pa
t
i
al
s
pa
r
s
i
ty
of
the
c
h
an
n
el
a
l
on
g
wi
t
h
t
he
us
e
of
l
arg
e
an
t
en
n
a
arr
ay
s
mo
ti
v
a
tes
a
s
ub
s
e
t
o
f
ph
y
s
i
c
al
l
ay
er
be
am
fo
r
m
i
ng
s
c
he
m
es
ba
s
ed
o
n
di
r
ec
t
i
on
al
tr
an
s
m
i
s
s
i
on
s
f
or
s
i
gn
a
l
l
i
n
g.
I
n
t
hi
s
c
on
tex
t,
t
he
r
e
ha
v
e
b
ee
n
a
f
ew
s
tud
i
es
o
n
the
de
s
i
gn
an
d
pe
r
f
orma
nc
e
an
al
y
s
i
s
of
di
r
ec
t
i
o
na
l
be
am
form
i
ng
/
prec
od
i
ng
s
tr
uc
tures
f
or
s
i
n
gl
e
-
us
er
mu
l
ti
-
i
np
u
t
m
ul
t
i
-
o
utp
ut
(
MIM
O
)
s
y
s
tem
s
[4
-
7]
.
Howev
er,
by
r
es
tr
i
c
ti
ng
at
ten
ti
o
n
to
s
m
al
l
c
e
l
l
c
ov
erag
e
an
d
by
r
ea
pi
ng
th
e
i
nc
r
ea
s
e
d
arr
ay
ga
i
ns
fr
om
the
us
e
of
l
arge
an
t
en
n
a
arr
ay
s
at
bo
th
the
ba
s
e
-
s
tat
i
on
an
d
us
er
en
ds
,
s
i
gn
i
fi
c
a
nt
r
at
e
i
mp
r
ov
e
me
nts
c
an
b
e
r
ea
l
i
z
ed
i
n
prac
ti
c
e.
T
he
s
e
w
ork
s
s
h
ow
tha
t
di
r
ec
ti
on
al
s
c
he
me
s
are
no
t
on
l
y
g
oo
d
fr
om
a
n
i
mp
l
e
me
nt
ati
on
s
ta
nd
p
oi
n
t
bu
t
are
al
s
o
r
o
bu
s
t
t
o
ph
as
e
c
ha
ng
es
ac
r
os
s
c
l
us
ters
an
d
a
l
l
ow
a
s
mo
ot
h
tr
ad
e
-
off
be
tw
ee
n
p
ea
k
b
ea
mf
ormi
ng
ga
i
n
an
d
i
ni
t
i
a
l
us
er
d
i
s
c
ov
ery
l
ate
nc
y
.
T
he
r
e
ha
s
a
l
s
o
be
en
progr
es
s
i
n
ge
n
eral
i
z
i
ng
s
uc
h
d
i
r
ec
ti
on
a
l
c
on
s
tr
uc
ti
on
s
for
mu
l
t
i
-
us
er
MIM
O
tr
an
s
m
i
s
s
i
on
s
[8
-
1
1].
S
ev
era
l
r
ec
en
t
w
ork
s
ha
v
e
ad
dr
es
s
ed
hy
bri
d
be
am
form
i
ng
for
m
i
l
l
i
me
tr
e
wav
e
s
y
s
tem
s
.
T
he
prob
l
em
of
fi
nd
i
ng
the
op
t
i
m
al
prec
od
er
a
nd
c
om
bi
ne
r
w
i
th
a
hy
brid
arc
hi
tec
t
u
r
e
i
s
po
s
ed
as
a
s
pa
r
s
e
r
ec
on
s
tr
uc
ti
on
pro
bl
e
m
i
n
[
12
],
l
e
ad
i
ng
to
al
go
r
i
thm
s
a
nd
s
o
l
ut
i
on
s
ba
s
ed
on
b
as
i
s
pu
r
s
u
i
t
m
eth
o
ds
.
W
h
i
l
e
th
e
s
ol
uti
on
s
ac
hi
ev
e
go
o
d
pe
r
forma
nc
e
i
n
c
erta
i
n
c
as
es
,
to
a
dd
r
es
s
t
he
pe
r
for
m
an
c
e
ga
p
b
etwe
en
th
e
s
o
l
uti
o
n
propos
ed
i
n
[1
2]
an
d
th
e
un
c
o
ns
tr
ai
n
ed
be
am
f
ormer
s
tr
uc
ture,
a
n
i
tera
ti
v
e
s
c
he
m
e
i
s
pro
po
s
ed
i
n
[
13
,
1
4]
r
el
y
i
n
g
o
n
a
hi
er
arc
hi
c
al
tr
a
i
ni
ng
c
od
e
bo
ok
for
ad
ap
t
i
v
e
es
t
i
ma
ti
o
n
of
mi
l
l
i
m
etre
w
av
e
c
ha
nn
e
l
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
MNIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
A
wi
r
el
es
s
prec
o
di
ng
t
ec
hn
i
qu
e f
or m
i
l
l
i
me
tr
e
-
wav
e
...
(
Roun
ak
ul
Is
l
am
B
o
by
)
2783
T
he
a
uth
ors
i
n
[
13
,
14
]
s
ho
w
t
ha
t
a
f
ew
i
terat
i
on
s
of
t
he
s
c
he
me
are
e
no
u
gh
to
ac
h
i
ev
e
near
-
op
t
i
m
al
pe
r
for
ma
nc
e.
In
[1
5],
i
t
i
s
es
ta
bl
i
s
he
d
tha
t
a
hy
bri
d
arc
h
i
tec
ture
c
an
ap
proac
h
the
p
erfor
m
an
c
e
o
f
a
di
gi
ta
l
arc
h
i
tec
ture
i
f
the
nu
mb
er
of
RF
c
ha
i
ns
i
s
twi
c
e
t
ha
t
of
the
d
ata
-
s
tr
ea
ms
.
A
he
uris
t
i
c
al
go
r
i
t
hm
wi
t
h
g
oo
d
pe
r
f
ormanc
e
i
s
d
ev
el
o
pe
d
wh
e
n
th
i
s
c
on
d
i
ti
on
i
s
no
t
s
at
i
s
fi
e
d.
S
ev
era
l
ot
he
r
wor
k
s
s
uc
h
as
[
16
,
17
]
ha
v
e
a
l
s
o
ex
p
l
ore
d
i
terat
i
v
e/a
l
g
orit
hm
i
c
s
ol
ut
i
on
s
f
or hy
brid
be
am
f
o
r
mi
ng
.
A
c
o
mm
o
n
t
he
me
tha
t
u
nd
erli
es
m
os
t
of
t
he
s
e
wor
k
s
i
s
the
as
s
um
pti
on
of
ph
as
e
-
on
l
y
c
on
tr
ol
i
n
th
e
RF
/
an
a
l
o
g
do
ma
i
n
for
the
hy
bri
d
b
e
am
form
i
n
g
arc
h
i
tec
ture
.
T
hi
s
as
s
um
pti
on
ma
k
es
s
en
s
e
at
th
e
us
er
en
d
w
i
th
a
s
m
al
l
er
n
um
b
er
of
a
nte
n
na
s
(
r
el
ati
v
e
to
t
h
e
ba
s
e
-
s
tat
i
o
n
en
d),
w
he
r
e
o
pe
r
at
i
n
g
th
e
P
A
s
be
l
ow
t
he
i
r
pe
ak
r
ati
n
g
ac
r
os
s
RF
c
ha
i
n
s
c
an
l
ea
d
to
a
s
ub
s
ta
nt
i
al
l
y
p
oo
r
up
l
i
nk
pe
r
f
ormanc
e
.
O
n
th
e
oth
er
ha
nd
,
a
mp
l
i
tu
de
c
on
tr
o
l
(
de
n
ote
d
as
am
p
l
i
t
ud
e
t
ap
er
i
ng
i
n
the
an
te
nn
a
th
eo
r
y
l
i
t
erature)
i
s
ne
c
es
s
ary
at
the
ba
s
e
-
s
tat
i
o
n
en
d
w
i
th
ma
ny
a
nte
n
na
s
for
s
i
de
-
l
ob
e
ma
na
g
em
e
nt
a
nd
m
i
t
i
ga
t
i
ng
ou
t
-
of
-
ba
nd
e
mi
s
s
i
on
s
.
F
urther,
g
i
v
en
tha
t
t
he
b
as
e
-
s
tat
i
on
i
s
a
ne
twork
r
es
ou
r
c
e,
s
i
mu
l
t
an
eo
us
am
p
l
i
tud
e
a
nd
p
h
as
e
c
on
tr
o
l
of
the
i
n
di
v
i
du
al
an
ten
na
s
a
c
r
os
s
RF
c
ha
i
ns
i
s
fea
s
i
bl
e
at
mi
l
l
i
me
tr
e
wav
e
b
as
e
-
s
tat
i
o
ns
at
a
l
ow
-
c
om
pl
ex
i
ty
an
d
c
os
t
[18
].
T
he
mi
l
l
i
m
etre
wav
e
ex
pe
r
i
m
en
ta
l
prot
oty
pe
d
em
on
s
tr
a
ted
i
n
al
l
ows
s
i
m
ul
t
an
eo
us
am
p
l
i
t
ud
e
an
d
ph
as
e
c
on
tr
o
l.
T
a
bl
e
1
s
ho
ws
the
s
um
ma
r
y
of
th
e
r
e
l
ate
d
r
ev
i
ew p
ap
ers
.
T
ab
l
e
1
.
S
u
mm
ar
i
z
ati
on
of
no
ta
bl
e
r
ev
i
ew
pa
pe
r
s
.
Met
h
o
d
s
Y
e
a
r
s
A
d
v
a
n
t
a
g
e
s
D
is
a
d
v
a
n
t
a
g
e
s
C
o
n
v
e
n
t
ion
a
l
mm
W
a
v
e
2016
-
17
1
.
H
igh
f
r
e
q
u
e
n
c
y
6
GH
z
.
1
.
H
igh
p
a
t
h
lo
s
s
e
x
p
o
n
e
n
t
s
,
2
.
h
igh
e
r
s
h
a
d
o
w
f
a
d
ing
,
3
.
b
lo
c
k
a
g
e
a
n
d
p
e
n
e
t
r
a
t
ion
los
s
e
s
,
e
t
c
.
S
ing
le
u
s
e
r
M
I
MO
2013
-
16
1
.
R
o
b
u
s
t
t
o
p
h
a
s
e
c
h
a
n
g
e
s
a
c
r
o
s
s
c
lus
t
e
r
s
a
n
d
a
ll
o
w
a
s
moo
t
h
t
r
a
d
e
-
o
f
f
b
e
t
w
e
e
n
p
e
a
k
b
e
a
mf
o
r
mi
n
g
g
a
ins
.
I
n
it
ial
u
s
e
r
d
is
c
o
v
e
r
y
lat
e
n
c
y
.
1
.
L
a
r
g
e
a
n
t
e
n
n
a
a
r
r
a
y
s
mot
iv
a
t
e
a
s
u
b
s
e
t
o
f
p
h
y
s
i
c
a
l
lay
e
r
b
e
a
mf
o
r
m
ing
.
Mult
i
-
u
s
e
r
M
I
MO
2014
-
17
1
.
Ge
n
e
r
a
li
z
ing
s
u
c
h
d
ire
c
t
ion
a
l
c
o
n
s
t
r
u
c
t
ion
s
f
o
r
mult
i
-
u
s
e
r
.
1
.
S
w
i
t
c
h
ing
mod
e
d
e
c
r
e
a
s
e
e
f
f
i
c
i
e
n
c
y
.
2
.
C
e
r
t
a
in
d
a
t
a
lo
s
s
.
P
u
r
s
u
it
me
t
h
o
d
s
-
b
a
s
e
d
H
y
b
r
id
a
r
c
h
it
e
c
t
u
r
e
.
2014
1
.
I
n
c
r
e
a
s
e
d
p
e
r
f
o
r
man
c
e
b
y
a
d
d
r
e
s
s
ing
t
h
e
p
e
r
f
o
r
man
c
e
g
a
p
b
e
t
w
e
e
n
t
h
e
c
h
a
n
n
e
l
s
w
it
c
h
ing
.
1
.
A
s
s
u
med
p
h
a
s
e
c
o
n
t
r
o
l
in
t
h
e
R
F/
a
n
a
log
d
o
main
,
o
n
l
y
p
o
s
s
ible
i
n
s
mall
n
u
mbe
r
o
f
a
n
t
e
n
n
a
e
.
D
igit
a
l
h
y
b
r
id
a
r
c
h
it
e
c
t
u
r
e
.
2016
-
17
1
.
H
y
b
r
id
b
e
a
mf
o
r
mi
n
g
.
2
.
A
h
e
u
r
is
t
i
c
a
lgo
r
i
t
h
m
u
s
e
d
f
o
r
b
e
t
t
e
r
p
e
r
f
o
r
man
c
e
.
1
.
N
u
mbe
r
o
f
R
F
c
h
a
in
s
is
t
w
i
c
e
t
h
a
t
o
f
t
h
e
d
a
t
a
-
s
t
r
e
a
m
s
.
2
.
S
u
b
s
t
a
n
t
iall
y
p
o
o
r
u
p
l
ink
p
e
r
f
o
r
man
c
e
.
S
A
P
C
m
mWa
v
e
2017
1
.
S
im
u
l
t
a
n
e
o
u
s
a
mpli
t
u
d
e
a
n
d
p
h
a
s
e
c
o
n
t
r
o
l
o
f
t
h
e
ind
i
v
idu
a
l
a
n
t
e
n
n
a
s
a
c
r
o
s
s
R
F
c
h
a
ins
.
2
.
L
o
w
-
c
o
mp
lex
i
t
y
a
n
d
c
o
s
t
.
1
.
S
t
a
n
d
a
r
d
c
a
p
a
c
i
t
y
o
f
ma
x
im
u
m
1
2
7
p
o
int
s
.
H
y
b
r
id
p
r
e
c
o
d
ing
s
ing
le
-
u
s
e
r
mm
W
a
v
e
2017
1
.
H
y
b
r
id
p
r
e
c
o
d
ing
/
c
o
mbin
ing
i
s
c
a
p
a
b
le.
2
.
S
a
me
p
e
r
f
o
r
man
c
e
o
f
t
h
e
f
u
ll
y
d
igit
a
l.
1
.
Failur
e
o
f
d
e
d
i
c
a
t
e
d
c
o
mpu
t
e
r
o
r
c
o
n
n
e
c
t
ion
p
r
o
b
lem
c
a
n
f
a
il
t
h
e
s
y
s
t
e
m.
2
.
R
e
q
u
ire
d
main
t
e
n
a
n
c
e
.
H
y
b
r
id
p
r
e
c
o
d
ing
f
o
r
mult
i
-
u
s
e
r
m
mWa
v
e
2015
1
.
C
o
mbina
t
ion
o
f
R
F
c
o
mbine
r
a
n
d
R
F
b
e
a
mf
o
r
mer
t
o
ma
x
im
i
z
e
t
h
e
c
h
a
n
n
e
l
g
a
in.
2
.
D
e
r
iv
e
d
a
s
a
z
e
r
o
-
f
o
r
c
ing
(
ZF)
p
r
e
c
o
d
e
r
.
1
.
For
a
s
mall
p
lan
t
.
2
.
E
x
t
e
n
s
ion
n
o
t
p
o
s
s
ible
.
Mea
n
-
s
q
u
a
r
e
d
e
r
r
o
r
(
MS
E
)
h
y
b
r
id
p
r
e
c
o
d
e
r
2011
1
.
Ma
x
i
mum
li
k
e
li
h
o
o
d
(
ML)
d
e
c
o
d
e
r
a
n
d
a
mi
n
i
mum
mea
n
s
q
u
a
r
e
e
r
r
o
r
(
MM
S
E
)
d
e
c
o
d
e
r
.
2
.
W
ind
o
w
c
o
e
f
f
i
c
ien
t
s
u
s
e
d
t
o
g
e
n
e
r
a
t
e
t
h
e
q
u
a
n
t
i
z
e
d
v
a
lue
s
.
1
.
The
p
e
r
f
o
r
man
c
e
d
e
p
e
n
d
s
o
n
d
e
t
e
c
t
ion
e
n
g
ine
.
2.
Re
se
a
r
ch
Me
t
h
o
d
T
he
pro
po
s
ed
s
y
s
tem
i
s
a
c
om
b
i
na
t
i
o
n
of
s
uc
c
es
s
i
v
e
i
nte
r
fere
nc
e
c
an
c
e
l
l
ati
on
(
S
IC)
an
d
M
i
ni
mu
m
Me
a
n
S
q
ua
r
e
E
r
r
or
(
MM
S
E
)
or
c
a
n
be
wr
i
tt
en
as
S
IC
-
M
MS
E
.
In
th
i
s
proc
es
s
,
i
ni
ti
a
l
l
y
,
th
e
r
a
w
d
ata
i
s
s
a
mp
l
ed
an
d
prepare
d
for
s
u
b
-
ba
n
d
p
ac
k
ag
i
ng
ac
c
ordi
n
g
to
us
er
s
’
d
ata
s
y
mb
ol
.
T
he
c
od
er
i
s
j
o
i
n
ed
al
on
g
w
i
th
M
MS
E
de
tec
t
i
o
n
s
y
s
tem
,
wh
i
c
h
w
i
l
l
de
p
en
d
u
po
n
us
er
or
op
erat
or.
T
he
MM
S
E
de
tec
ti
on
proc
es
s
wi
l
l
c
on
t
i
n
ue
t
o
do
c
ha
nn
e
l
p
i
l
ot
s
en
s
i
ng
,
tes
ti
n
g
s
i
gn
al
qu
a
l
i
ty
,
es
ti
m
ate
the
S
i
gn
al
t
o
No
i
s
e
R
ati
o
(
S
NR)
,
arr
an
ge
C
ha
nn
e
l
Ma
tr
i
x
(
CM)
form
ati
on
,
Chan
ne
l
s
el
ec
ti
o
n
&
es
ti
ma
ti
on
.
T
he
MM
S
E
proc
e
s
s
ed
da
ta
wi
l
l
be
fi
l
tere
d
for
ma
x
i
mu
m
Evaluation Warning : The document was created with Spire.PDF for Python.
◼
IS
S
N: 16
93
-
6
93
0
T
E
L
KO
MNIK
A
V
ol
.
17
,
No
.
6,
D
ec
em
b
er 20
19
:
27
8
2
-
2789
2784
c
orr
el
ati
on
d
ete
c
ti
on
,
wh
i
c
h
i
s
the
pa
r
t
of
S
IC
m
eth
o
d.
T
hi
s
me
tho
d
i
s
u
s
ed
to
de
t
ec
t
the
s
eq
u
en
t
i
a
l
l
y
proc
es
s
ed
da
ta
ac
c
ordi
ng
t
o
the
us
ers
’
s
y
mb
ol
s
an
d
r
eg
e
ne
r
a
te
the
d
ata
t
o
tr
an
s
fer
i
t
t
hrou
gh
the
n
ew
c
ha
n
ne
l
.
W
h
i
l
e
MM
S
E
wi
l
l
c
on
s
t
an
t
l
y
m
on
i
tor
t
he
s
i
gn
a
l
q
ua
l
i
ty
to
r
ea
l
ter
th
e
c
h
an
n
el
c
oe
ff
i
c
i
en
t,
the
S
IC
wi
l
l
he
l
p
M
M
S
E
to
i
mp
r
ov
e
i
ts
pe
r
f
orm
an
c
e
by
fas
t
l
y
proc
es
s
i
ng
s
eq
u
en
t
i
a
l
da
t
a
s
o
th
a
t
MM
S
E
c
an
r
es
el
ec
t
an
y
pa
r
a
me
t
ers
at
a
ny
mo
me
nt
t
o
r
ed
uc
e
i
nte
r
r
u
pti
on
an
d
da
ta
l
os
s
.
A
t
t
he
en
d
of
tr
an
s
mi
s
s
i
on
proc
es
s
,
the
RF
mo
du
l
at
i
o
n
w
i
l
l
mo
du
l
at
e
th
e
d
ata
t
he
n
fi
l
t
er
wi
t
h
S
pe
c
tr
u
m
S
ha
pi
n
g
F
i
l
ter
(
S
S
F
)
an
d
tr
a
ns
mi
t
t
hrou
gh
the
c
ha
nn
e
l
.
A
s
y
nc
hro
n
i
z
er
i
s
us
ed
i
n
tr
an
s
m
i
s
s
i
o
n
proc
es
s
to
s
y
nc
hron
i
z
e
an
y
d
i
s
r
up
te
d
op
erat
i
o
n.
O
n
t
he
r
ec
e
i
v
er
s
i
de
t
he
s
i
gn
al
w
i
l
l
be
d
e
mo
du
l
at
ed
an
d
r
es
ha
p
ed
wi
th
S
S
F
.
A
fte
r
de
mo
du
l
at
i
on
th
e
s
a
me
c
on
c
ep
t
of
propos
e
d
MM
S
E
wi
l
l
be
us
e
d
to
d
ec
od
e
the
da
ta.
T
he
s
y
nc
hron
i
z
er
o
n
th
e
r
ec
ei
v
er
s
i
d
e
a
nd
tr
an
s
m
i
tte
r
s
i
d
e
w
i
l
l
be
s
y
nc
hro
n
i
z
ed
t
og
e
the
r
throug
h
M
MS
E
.
F
i
na
l
l
y
,
t
he
de
c
o
de
d
da
t
a
w
i
l
l
be
r
eframe
d
us
i
n
g
s
am
e
S
I
C
me
th
od
.
T
h
i
s
c
om
bi
na
t
i
on
(
S
I
C
-
MM
S
E
)
c
an
r
ed
uc
e
t
he
c
ha
n
ne
l
s
h
ortage
an
d
pe
r
form
an
c
e
l
o
s
s
es
.
T
he
t
ota
l
proc
es
s
o
f
propos
ed
s
y
s
tem
for
tr
an
s
m
i
s
s
i
on
u
ni
t
an
d
proc
es
s
of
r
ec
ei
v
er
u
ni
t
are
s
ho
w
n
i
n Fi
gu
r
e
1.
F
i
gu
r
e
1.
P
r
op
os
e
d s
y
s
tem
ap
prox
i
ma
t
i
on
fo
r
tra
ns
mi
s
s
i
on
a
nd
r
ec
e
i
v
er
2.1.
S
ymb
o
liz
e
S
amp
ling
A
m
ul
t
i
p
l
e
us
er
w
i
th
mu
l
ti
pl
e
no
de
s
f
or
ba
s
e
s
tat
i
on
(
B
S
)
was
c
o
ns
i
d
ered
ba
s
ed
o
n
ti
me
di
v
i
s
i
on
d
up
l
ex
(
T
DD)
me
t
ho
d
w
he
r
e
u
pl
o
ad
a
nd
d
o
wnl
o
ad
c
ha
nn
e
l
d
ata
l
i
nk
s
c
on
s
i
de
r
w
i
th
i
n
c
oh
erenc
e
i
nte
r
v
a
l
i
n
a
po
i
nt
to
po
i
nt
MIM
O
s
y
s
tem
.
C
on
s
i
d
erin
g
the
s
y
s
tem
h
av
e
N
nu
mb
ers
of
no
de
s
o
n
a
ba
s
e
s
tat
i
o
n
pe
r
c
el
l
s
ha
v
i
n
g
M
n
u
mb
er
of
tot
a
l
an
t
en
n
as
pe
r
c
el
l
s
a
n
d
K
nu
mb
er
of
s
i
ng
l
e
an
t
en
n
ae
k
no
wn
as
us
er
term
i
n
al
(
UT
)
i
n
ea
c
h
c
el
l
.
F
or
K
an
ten
na
us
er
t
e
r
mi
na
l
to
ba
s
e
s
tat
i
o
n J
c
an
b
e e
x
pres
s
ed
as
:
H
jk
=
B
jk
G
jk
w
he
r
e,
H
jk
i
s
the
fa
di
n
g’
s
on
J
s
tat
i
on
f
or
K
nu
mb
er
of
s
i
ng
l
e
an
ten
na
e
.
B
jk
i
s
the
fad
i
n
g
c
oe
ffi
c
i
en
t
of
l
arge
s
c
al
e
a
nd
G
jk
i
s
the
fad
i
ng
c
o
eff
i
c
i
en
t
of
s
ma
l
l
s
c
al
e
[
7].
Her
e
,
B
jk
r
ep
r
es
en
t
pa
th
l
os
s
an
d
s
ha
do
w
fad
i
ng
of
t
he
c
ha
n
ne
l
.
T
h
e
ma
tr
i
x
was
de
n
ote
d
by
up
pe
r
c
as
e
an
d
b
ol
d
up
pe
r
c
as
e
us
e
d
for
v
ec
tor
i
de
nti
f
i
c
at
i
on
s
.
T
he
G
jk
i
s
th
e
tot
a
l
no
d
al
fa
di
ng
eff
ec
t
i
nd
uc
ed
i
n
pe
r
c
el
l
’
s
c
ap
ac
i
ty
c
a
n b
e
r
ep
r
e
s
en
ted
by
[7
],
G
jk
=
CM (0, I
m)
w
he
r
e,
C
an
d
M
are
th
e
c
ap
ac
i
ty
s
um
r
at
e
&
nu
mb
e
r
of
B
S
an
t
en
n
as
r
es
pe
c
ti
v
el
y
a
nd
Im
i
s
the
i
nd
i
c
ati
on
f
un
c
ti
on
of
M
.
S
o,
H =
G
√
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
MNIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
A
wi
r
el
es
s
prec
o
di
ng
t
ec
hn
i
qu
e f
or m
i
l
l
i
me
tr
e
-
wav
e
...
(
Roun
ak
ul
Is
l
am
B
o
by
)
2785
h
ere,
H
i
s
c
ha
nn
el
fa
di
ng
,
B
r
ep
r
es
en
ts
the
l
arge
-
s
c
al
e
di
ag
o
na
l
ma
tr
i
x
an
d
G
r
ep
r
es
en
ts
the
s
ma
l
l
-
s
c
al
e
m
atri
x
e
ac
h
c
ol
u
mn
r
e
pres
en
ts
a
c
ha
nn
e
l
fr
om
UT
to
B
S
.
W
he
n
the
nu
m
be
r
of
B
S
an
t
en
n
as
i
nc
r
e
as
e
th
e
c
ha
nn
e
l
,
t
he
ap
pr
ox
i
ma
tes
ortho
go
n
al
ma
tr
i
x
wi
l
l
be
l
im
→
∞
=
B
.
E
ac
h
ter
mi
na
l
i
s
as
s
i
gn
ed
wi
th
a
pi
l
ot
s
e
ns
i
n
g
for
k
nu
mb
er
of
s
i
ng
l
e
a
nte
nn
as
,
t
h
e
s
en
s
i
n
g
pi
l
ot
s
k
wi
th
po
w
er
e
qu
a
l
t
o,
s
k,t
=
[s
k1
;
s
k2
….s
kt
]T
a
nd
at
e
ac
h
B
S
s
t
ati
on
,
‖
‖
2
=
0,
i
f
j
≠
k
an
d
the
tr
an
s
m
i
tte
d
po
wer
i
s
e
qu
a
l
f
or
a
l
l
p
i
l
ots
.
F
or
th
e
c
on
v
en
ti
on
a
l
de
t
ec
ti
o
n
t
he
r
ec
ei
v
er
v
ec
tor
ma
tr
i
x
y
c
an
be
de
n
ote
d
by
y
∈
C
M×
1
[8]
or y
=
Hx
+
n
(
1)
w
he
r
e,
C
i
s
c
h
an
n
el
ma
tr
i
x
an
d
c
om
p
l
ex
ad
di
t
i
v
e
w
hi
t
e
ga
us
s
i
an
no
i
s
e
(
A
W
G
N)
v
ec
tor,
H
∈
C
M×
K
,
x
i
s
the
s
y
mb
o
l
v
ec
tor
s
en
t
by
K
us
er
c
an
be
d
en
o
te
by
x
∈
C
k×
1
an
d
nu
mb
er
of
no
de
s
n.
If
the
s
y
mb
o
l
err
or v
ec
tor
e t
h
en
,
e =
x
-
ẋ
(
2)
h
ere,
ẋ
i
s
th
e
r
ec
e
i
v
i
n
g
s
i
g
na
l
.
A
s
s
um
i
n
g
c
orr
el
ati
on
p
aramet
er
σ
i
s
k
n
own
pe
r
fec
tl
y
at
the
ba
s
e
s
tat
i
o
ns
an
d
h[
n
] b
e
th
e
c
h
an
ne
l
v
ec
tor
be
tw
ee
n
a UT
an
d
a B
S
a
t ti
me
t
.
T
he
n [
9
]
,
h[t
] =
σ
h [
t
-
1]
+
ev
[t]
(
3)
h
ere,
t
i
s
ti
me
i
nd
ex
an
d
e[
t]
i
s
wh
i
te
no
i
s
e
w
i
th
z
ero
me
an
a
nd
tem
po
r
a
l
c
orr
e
l
a
ti
on
p
arame
ter
σ
2
o
bta
i
ne
d
t
hrou
gh
t
he
Y
ul
e
-
W
al
k
er
eq
u
ati
on
[7]
.
T
he
c
ha
nn
e
l
mo
de
l
a
bo
v
e
i
s
k
no
wn
as
the
s
tat
i
o
na
r
y
ergo
di
c
G
au
s
s
-
Ma
r
k
ov
bl
oc
k
fa
d
i
n
g c
ha
n
ne
l
mo
d
el
[8
].
2.2.
MM
S
E
Det
ec
t
ion
P
r
o
c
es
s
F
or
the
MI
MO
m
od
el
eq
u
ati
o
n
ac
c
ordi
n
g
to
r
ef
eren
c
e
no
[7]
,
wh
ere
r
ec
e
i
v
i
ng
s
i
gn
al
v
ec
tor
ŷ
fr
om
r
ec
ei
v
er
s
i
g
n
al
y
an
d
t
he
fi
br
i
no
us
n
orm
‖
‖
2
t
o
l
i
mi
t
s
ph
er
e
o
f
v
a
l
i
d
i
ty
of
ge
ne
r
al
no
r
m.
ŷ
=
y
−
H
ẋ
=
H (
e +
x
)
(
3)
w
he
r
e,
x
i
s
tr
an
s
mi
tt
ed
s
y
mb
o
l
ma
s
s
ag
es
an
d
ẋ
i
s
th
e
r
ec
e
i
v
ed
s
y
mb
o
l
ma
s
s
ag
es
.
E
r
r
or
v
ec
tor
e
s
ho
u
l
d
be
z
ero
f
or
i
d
e
al
c
om
mu
ni
c
at
i
on
s
y
s
tem
.
S
o,
t
ha
t
the
err
or
de
t
ec
t
i
on
s
h
ou
l
d
b
e
ov
erc
om
e
fr
om
r
ec
e
i
v
er
s
i
gn
a
l
v
ec
tor.
S
o
me
r
es
ea
r
c
he
r
ex
pres
s
es
the
c
om
pre
s
s
i
ng
s
en
s
i
n
g
me
th
od
s
,
wh
ere
the
y
pr
op
os
e
d
t
o
na
tural
l
y
c
on
s
i
de
r
the
s
y
mb
ol
err
or
v
ec
tor
e
[7]
.
In
c
om
pres
s
i
ng
s
en
s
i
ng
me
tho
ds
M
s
h
ou
l
d
b
e
l
es
s
the
n
K
,
bu
t
i
f
M
be
c
o
me
s
mo
r
e
the
n
eq
ua
l
to
K
,
th
i
s
s
y
s
tem
w
i
l
l
b
e
i
mp
r
ac
ti
c
al
.
F
or
MI
MO
mu
l
t
i
-
a
nte
n
na
mo
d
e,
t
he
M
i
s
ge
ne
r
al
l
y
gre
ate
r
the
n
eq
u
al
K
, t
h
e rec
e
i
v
er s
i
gn
al
v
ec
tor
l
at
er fi
l
ter
by
m
atri
x
W
MM
S
E
i
s
g
i
v
en
by
:
W
MMSE
=
=
ℎ
ℎ
+
(
4)
where,
W
i
s
a
prede
fi
n
e
f
i
l
t
er
ma
tr
i
x
,
W
MMSE
i
s
the
fi
l
te
r
ma
tr
i
x
for
MM
S
E
m
atri
x
f
or
an
d
A
W
G
N
(
G
au
s
s
i
an
no
i
s
e)
v
ec
tor
n
∈
CM
for
CM
(
0,
I
m
)
[
8].
B
y
Ma
x
i
mu
m
A
P
os
teri
or
(
M
A
P
)
d
ete
c
ti
on
k
no
wn
as
de
t
ec
ti
o
n
s
y
s
tem
de
tec
ti
on
me
th
od
t
he
op
ti
ma
l
d
ete
c
ti
on
é
c
an
be
fo
un
d
fr
om
the
r
ef
erenc
e
pa
p
er no
[10
]
.
é
≅
a
r
g
ma
x
∈
ˆA
(
1
√
2
2
⁄
)
ex
p [
−
0
.
707
2
(
(
|
|
ŷ
−
He
|
|
)
2
2
] P
r
(
e)
(
5)
A
c
c
ordi
n
g
to
th
e
p
ap
er
t
he
ap
prox
i
ma
ti
o
n
i
s
be
c
a
us
e
of
e
an
d
n
de
p
en
de
nc
y
a
n
d
ma
y
om
i
t
wh
i
l
e
S
NR
i
nc
r
ea
s
es
an
d
c
a
n
b
e
prec
i
s
e
at
h
i
g
h
S
N
Rs
[10
].
P
r(e)
i
s
prob
a
bi
l
i
ty
o
f
pri
orit
y
err
or
s
y
mb
ol
.
W
he
n
B
P
K
S
v
al
ue
s
are
+
1
&
-
1,
ˆA
i
s
th
e
f
i
ni
te
al
ph
a
be
t
ha
v
i
ng
th
e
v
al
u
es
-
2,
0
&
+
2
an
d
for
th
e
no
nz
ero
v
a
l
ue
of
A´
de
tec
ti
o
n
err
or
b
ec
o
me
s
-
2
&
2.
I
f
tr
an
s
m
i
tt
ed
s
y
mb
ol
s
are
fr
om
-
1
to
1,
the
n
the
p
os
s
i
bi
l
i
ty
of
the
e
w
i
l
l
be
n
o
z
eroes
fr
om
+
2
to
-
2
an
d
po
s
s
i
bl
e
prob
ab
i
l
i
ty
Evaluation Warning : The document was created with Spire.PDF for Python.
◼
IS
S
N: 16
93
-
6
93
0
T
E
L
KO
MNIK
A
V
ol
.
17
,
No
.
6,
D
ec
em
b
er 20
19
:
27
8
2
-
2789
2786
c
an
b
e
0
.5
P
.
W
he
n
λ
i
s
t
he
de
gree
of
s
p
ars
i
ty
,
|
|
e
|
|
=
0.2
5
|
|
e
|
|
2
.
If,
e
i
s
the
e
l
em
en
t
of
ˆA
an
d
e
i
s
th
e s
y
mb
ol
err
or v
ec
tor f
or i
n
i
ti
al
i
tera
ti
o
n,
by
s
o
l
v
i
n
g (5)
:
e
=
H
H
ˆy
H
H
H
+
0
.
5
λ
e
=
ˆy
; [i
f,
ˆA
i
s
fi
n
i
te
an
d
i
ni
t
i
a
l
l
y
é
=
e
]
(
6)
h
ere,
M
i
s
M
MS
E
d
ete
c
ti
on
me
th
od
w
i
th
tun
ea
bl
e
de
gre
e
of
s
p
ares
λ
,
where,
λ
is
the
r
ep
l
ac
em
en
t
of
no
i
s
e.
If,
Q
θ(*)
i
s
v
ec
tor
di
v
i
di
n
g
fu
nc
ti
o
n
an
d
θ
op
ti
ma
l
t
hres
ho
l
d
t
he
n
,
op
ti
ma
l
d
ete
c
ti
on
,
é=
Q
θ(e
)
fo
r
di
s
c
r
ete
fu
nc
t
i
on
[1
0
]. S
o,
w
e c
a
n rewr
i
te:
é =
Q
θ(e
)
=
2s
i
n(e
)
I ;
[|
|
e
|
|
>
θ]
(
7)
where,
“
I”
i
s
t
he
i
nd
i
c
ati
on
fun
c
ti
on
.
If
,
th
e
op
t
i
m
al
t
hres
ho
l
d,
θ
=
{
θ
1
,
θ
2
,
θ
3
,
…,
θ
n
}
a
nd
for
the
no
n
-
z
ero
c
om
po
ne
nts
,
e
=
{
0,
±
2};
[i
.
e.
|
|
e
|
|
<
θ
].
S
i
mi
l
arly
,
Q
P
S
K
d
ete
c
ti
on
the
eq
ui
v
a
l
en
t
tr
an
s
form w
i
th
r
ea
l
(
R)
an
d
i
ma
gi
na
r
y
(
I)
, whe
r
e
I(e)
an
d R(e)
pa
r
ts
o
f
x
,
é =
2s
i
n
[ {R(
e) +
I(e)
}
T]; [
W
he
r
e
, e
ˆA
]
(
8)
h
ere,
y
i
ni
t
i
a
l
r
ec
ei
v
er
s
i
gn
al
s
,
n
i
s
t
he
G
au
s
s
i
an
n
oi
s
e,
e(
l
)
i
s
th
e
l
th
s
y
mb
ol
err
or
v
ec
tor.
(
8)
i
s
the
prio
r
prob
ab
i
l
i
ty
d
ete
c
t
i
on
of
e.
I
f,
e(
l
)
i
s
n
on
-
z
ero,
for
t
he
n
th
en
tr
y
,
−
1
i
s
of
−
1
,
the
for
n
th
en
tr
y
of
l
th
s
y
mb
o
l
,
=
W
(
l
−
1)
+
∑
1
−
≠
+
−
1
(
9)
s
o,
G
au
s
s
i
an
ap
prox
i
m
ate
s
wi
th
fo
l
l
o
wi
n
g v
ari
an
c
e
,
(
−
1
)
2
=
∑
4
(
1
−
)
2
−
1
+
{
∑
(
−
1
)
≠
}
(
10
)
2.3
.
S
IC
A
lgo
r
it
h
m
Cons
i
d
erin
g
t
he
mm
-
wav
e
MI
MO
s
y
s
tem
w
i
th
D
i
s
tr
i
bu
t
ed
A
nt
en
n
a
S
y
s
tem
(
DA
S
)
c
on
fi
g
urat
i
on
,
wh
ere,
nu
m
be
r
o
f
b
as
e
an
t
en
n
a
M
B
h
av
i
ng
k
n
um
be
r
of
s
i
ng
l
e
an
te
nn
a
a
nd
N
nu
mb
er
r
em
ote
r
a
di
o
h
ea
ds
.
If
th
e
Q
us
er
al
s
o
e
qu
i
pp
e
d
w
i
th
M
U
a
nte
n
na
,
the
r
ec
e
i
v
i
n
g
an
te
nn
as
,
M
R
=
M
B
+
Nk
≥
Q
M
U
[1
9
-
26
]
.
F
or
Q
us
er
M
U
nu
mb
er
f
l
at
fad
i
n
g
c
h
an
n
el
s
,
t
he
M
MS
E
pi
l
ot
s
k
was
c
o
ns
i
d
ered
b
efo
r
e
no
w
c
an
be
r
ewr
i
tt
en
as
v
ec
toral
for
m,
s
k
∈
C
M
u
×1
.
F
r
o
m
the
mo
d
el
as
G
au
s
s
-
Ma
r
k
o
v
bl
oc
k
fa
di
ng
c
h
an
n
el
s
ho
wn
a
bo
v
e
i
n
(
3),
th
e
d
ata
v
ec
tor
s
k
ha
v
e
z
ero me
a
n.
T
he
S
IC
al
go
r
i
thm
r
el
i
es
on
s
eq
u
en
t
i
al
d
ete
c
t
i
on
r
ec
ei
v
er
s
i
g
na
l
s
,
wh
ere
i
t
i
s
r
e
qu
i
r
ed
to
eq
ua
l
i
z
e
the
c
h
an
n
el
ma
tr
i
c
es
W
MMSE
gi
v
e
n
i
n
(
4),
t
he
n
c
arr
i
e
r
c
h
an
n
el
s
c
an
ge
t
th
e
h
i
gh
er
S
i
g
na
l
to
Int
erfer
en
c
e
No
i
s
e
Rat
i
o
(
S
INR)
.
F
r
o
m
t
he
r
efe
r
en
c
e
no
[2
6
]
th
e
S
INR
pe
r
s
y
mb
o
l
for
I
th
i
t
erati
on
f
or the
j
th
nu
mb
e
r
of
s
y
mb
o
l
s
i
s
th
us
c
an
b
e
ex
pres
s
ed
as
,
=
(
)
−
2
(
)
2
(
|
s
ki
|)
2
(
13
)
where,
i
s
the
a
mp
l
i
tu
de
,
G
au
s
s
i
an
ap
prox
i
m
ate
s
v
ar
i
a
nc
e
,
p
i
l
ot
s
k
for
I
th
i
t
erat
i
on
.
T
he
f
ad
i
ng
ma
tr
i
x
H
k
f
or
k
us
er,
ha
v
i
n
g
N+
1
s
ub
ma
tr
i
x
i
n
e
ac
h
r
em
ot
e
r
a
di
o
h
ea
d,
t
he
n
,
H
k
=
[H
k1
,
H
k2
,
…,
H
k(N+
1)
]
T
.
W
he
n
the
s
y
mb
ol
i
s
de
c
i
d
es
ac
c
ordi
n
g
a
d
ec
i
s
i
on
wi
l
l
be
m
ad
e
de
p
e
nd
s
on
MM
S
E
op
erat
or
g
i
v
en
i
n
(
1
1).
In
s
tea
d
o
f
ex
ec
ut
i
n
g
d
on
’
t
c
are
s
i
gn
de
c
i
s
i
on
,
i
t
i
s
p
os
s
i
bl
e
to
us
e
op
erat
or
Q
as
s
oft
s
wi
tc
h
throug
h
the
hy
p
erbo
l
i
c
tan
g
en
t
no
n
-
l
i
ne
ar
de
tec
tor
w
ho
s
e
a
r
gu
me
nt
i
s
wei
g
hte
d
by
a
n
es
ti
m
ati
on
of
th
e
S
INR
[
2
6
].
S
o
,
the
ex
pres
s
i
o
n
f
or
s
k
for
I
th
i
t
erati
o
n
c
a
n
be
gi
v
en
i
n Q
P
S
K
c
o
ns
tel
l
at
i
o
n a
s
,
=
0
.
707
[
ta
n
h
{
(
)
/
(
2
)
}
+
ta
n
h
{
(
)
/
(
2
)
}
]
(
14
)
f
i
na
l
l
y
,
f
or
th
e
de
c
o
de
d
c
a
s
e
i
n
r
ec
e
i
v
er
en
d,
wh
i
l
e
al
l
s
y
m
bo
l
s
are
r
etri
ev
ed
,
the
d
on
’
t
c
are
de
c
i
s
i
o
n w
i
l
l
p
erfor
m f
or the
r
es
ul
ti
ng
ou
t
pu
t y
=
(
y
1
, y
2
,
…, y
n
)
T
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
MNIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
A
wi
r
el
es
s
prec
o
di
ng
t
ec
hn
i
qu
e f
or m
i
l
l
i
me
tr
e
-
wav
e
...
(
Roun
ak
ul
Is
l
am
B
o
by
)
2787
3.
Me
as
u
r
ement
and
S
im
u
latio
n
s
F
or
M
A
T
LA
B
s
i
m
ul
a
ti
o
n
we
us
ed
G
au
s
s
i
an
no
i
s
e
as
r
e
ferenc
e
wi
t
h
di
f
ferent
S
NR
l
ev
el
s
to
a
na
l
y
s
e
th
e
p
erfor
ma
nc
e
of
th
e
pr
op
os
e
d
S
IC
-
M
M
S
E
s
y
s
tem
.
I
n
t
hi
s
s
i
mu
l
at
i
on
pr
oc
es
s
we
ha
v
e
c
o
mp
ar
ed
r
es
ul
ts
wi
th
c
on
v
e
nt
i
on
al
m
m
-
wav
e
MI
MO
s
y
s
tem
an
d
MM
S
E
s
y
s
tem
.
F
or
the
s
i
mu
l
ati
on
proc
es
s
we
fi
r
s
t
c
on
s
i
d
ered
t
he
n
um
be
r
of
a
nte
nn
a
e
pe
r
c
e
l
l
s
M=
10
0
0.
F
or
the
proc
es
s
,
i
n
i
ti
al
l
y
we
de
tec
te
d
s
y
mb
ol
v
ec
tor
j
us
i
ng
c
on
v
en
t
i
o
na
l
MIM
O
s
y
s
tem
an
d
propos
e
d
M
MS
E
.
F
or
th
e
ou
t
pu
t
S
NRs
prio
r
i
ty
pro
ba
b
i
l
i
ty
for
c
o
nv
en
t
i
on
al
a
nd
M
MS
E
we
fol
l
owe
d e
qu
at
i
o
ns
from
the
r
efe
r
en
c
e
pa
p
ers
[7
-
10
]
s
h
own
i
n (1
5):
l
im
→
∞
l
o
g
(
)
l
o
g
=
−
;
l
im
→
∞
l
o
g
−
(
)
l
o
g
=
−
(
15
)
T
he
n,
the
de
gree
of
s
p
ars
i
ty
λ
,
c
an
b
e
o
bta
i
ne
d
fr
om
λ
=
ln
[
2
(
1
−
)
]
;
Cons
i
de
r
i
ng
the
MM
S
E
l
i
ne
ar d
ete
c
ti
on
,
fo
r
t
he
I
th
i
terat
i
on
th
e e
r
r
or
prob
ab
i
l
i
ty
é
for
S
IC
-
M
MS
E
b
as
ed
MIM
O
was
ob
t
ai
ne
d
fr
om
(
8),
w
he
r
e
op
t
i
ma
l
t
hres
ho
l
d
was
ob
t
ai
ne
d
by
s
o
l
v
i
n
g
the
(
7).
T
hi
s
propos
e
d
r
es
e
arc
h
w
as
c
on
du
c
t
o
n
T
i
me
D
i
v
i
s
i
o
n
Dupl
ex
(
T
DD)
me
t
ho
d
.
S
o
,
to
d
ete
r
m
i
n
e
the
S
pe
c
tr
a
l
E
ffi
c
i
en
c
y
(
S
E
)
fo
r
S
IC
-
MM
S
E
i
s
ex
pres
s
e
d [
27
]:
η
hM
M
S
E
=
(
−
−
)
(
16
)
where
T
p
=
prea
mb
l
e
p
eri
od
,
T
t
=
tr
ai
l
er
ti
me
p
erio
d,
T
f
=
fr
am
e
du
r
at
i
o
n
an
d
N
s
=
nu
mb
er
of
s
y
mb
ol
s
i
n
a
t
t
i
m
e
s
l
ot
,
Ni
=
nu
mb
er
of
i
nfo
r
m
ati
on
bi
t
s
.
B
y
r
es
ol
i
v
i
ng
t
h
e
eq
u
ati
o
ns
i
n
M
A
T
L
A
B
fi
na
l
l
y
w
e g
ot
S
E
f
or th
e S
I
C
-
MM
S
E
.
S
i
mu
l
ati
ng
the
prop
os
ed
s
y
s
tem
i
n
MA
T
LA
B
t
he
pe
r
forma
nc
e
o
f
S
IC
-
MM
S
E
was
ac
hi
ev
e
d.
F
or
th
e
c
o
mp
aris
on
an
d
b
en
c
hm
ark
i
ng
we
al
s
o
s
i
m
ul
ate
d
t
he
c
on
v
en
t
i
o
na
l
mm
-
W
av
e,
wh
ere
the
s
i
mu
l
ati
o
n
was
do
ne
by
S
pe
c
tr
al
E
ffi
c
i
en
c
y
(
S
E
)
[b
i
t/s
/Hz
/c
e
l
l
]
v
s
Num
be
r
of
B
S
A
nte
n
na
s
(
M).
F
i
g
ure
2
s
ho
ws
the
c
om
pa
r
i
s
on
of
th
e
S
pe
c
tr
a
l
E
ff
i
c
i
en
c
y
(
S
E
)
wi
th
the
i
nc
r
ea
s
e
nu
mb
er
of
B
S
an
ten
na
at
b
as
e
s
tat
i
on
fo
r
c
on
v
en
ti
o
na
l
or
s
i
ng
l
e
m
i
l
l
i
m
etre
-
W
av
e
s
y
s
tem
,
where
i
t
i
s
de
pi
c
ted
t
ha
t
s
pe
c
tr
a
l
eff
i
c
i
en
c
y
i
nc
r
ea
s
es
fr
om
0
to
ma
x
i
m
um
13
6
bi
ts
/s
/Hz
/c
e
l
l
wi
t
h t
h
e i
nc
r
ea
s
e o
f
ba
s
e
an
t
en
n
as
f
r
om
0
to
10
0
0.
W
he
r
e,
th
e
pa
r
am
ete
r
s
are
op
ti
mi
z
e
d
for
the
be
tte
r
p
e
r
forman
c
e,
the
m
ax
i
mu
m
S
E
was
r
ec
ord
ed
t
o
1
57
b
i
ts
/s
/Hz
/c
el
l
for
the
1
00
0
n
u
mb
er
of
a
nte
nn
as
.
F
i
gu
r
e
2.
C
on
v
en
t
i
o
na
l
m
m
-
wav
es
MI
MO
a
nd
op
t
i
m
i
z
e
d m
m
-
w
av
es
MI
MO
s
p
ec
tr
a
l
ef
f
i
c
i
e
nc
y
pe
r
forma
nc
e w
i
th
the
i
nc
r
e
as
e n
u
mb
er
of
A
nte
nn
as
.
T
he
S
IC
-
M
MS
E
s
i
mu
l
at
i
o
n
i
n
F
i
gu
r
e
3
s
ho
ws
be
tt
er
pe
r
for
ma
nc
e
th
an
c
on
v
en
ti
on
al
mm
-
W
av
e
MI
MO
s
y
s
tem
a
fte
r
op
t
i
m
i
z
ati
on
.
B
ef
ore
o
p
ti
m
i
z
ati
on
t
he
m
ax
i
m
um
S
E
was
fo
un
d
to
13
3~
13
2
b
i
ts
/s
/Hz
/c
el
l
w
hi
l
e
nu
m
be
r
of
a
nte
nn
as
w
as
ma
x
i
mu
m.
W
he
r
e,
a
fte
r
op
ti
mi
z
a
ti
o
n
Evaluation Warning : The document was created with Spire.PDF for Python.
◼
IS
S
N: 16
93
-
6
93
0
T
E
L
KO
MNIK
A
V
ol
.
17
,
No
.
6,
D
ec
em
b
er 20
19
:
27
8
2
-
2789
2788
t
he
v
a
l
u
e
c
r
os
s
ed
19
5
bi
ts
/s
/Hz
/c
el
l
.
E
v
ery
s
y
s
tem
r
e
qu
i
r
es
op
t
i
m
i
z
at
i
on
,
w
he
r
e
thi
s
propos
e
d
s
y
s
tem
pe
r
forme
d
al
mo
s
t
t
he
eq
u
al
t
o
op
t
i
m
i
z
ed
c
on
v
en
ti
on
a
l
m
m
-
wav
e
MI
MO
s
y
s
tem
,
bu
t
aft
er
op
ti
mi
z
at
i
o
n
i
t
r
ap
i
d
l
y
i
nc
r
ea
s
ed
.
F
i
g
ure
4
s
ho
ws
the
pe
r
forma
nc
e
c
om
p
ar
i
s
on
s
i
mu
l
ati
on
b
l
oc
k
for
the
bo
th
me
tho
ds
ha
v
i
n
g
s
am
e
pa
r
a
me
ters
,
s
pe
c
tr
al
ef
fi
c
i
en
c
y
ac
c
ordi
n
g
to
i
n
c
r
ea
s
e
nu
m
be
r
of
an
ten
na
s
.
F
i
gu
r
e
3.
P
r
op
os
e
d
S
IC
-
M
MS
E
an
d o
pti
mi
z
e
d
S
IC
-
M
MS
E
s
pe
c
tr
a
l
ef
f
i
c
i
en
c
y
pe
r
forman
c
e
ac
c
ordi
ng
to
i
nc
r
ea
s
e
nu
m
be
r
of
an
te
nn
a
s
F
i
gu
r
e
4.
T
h
e p
erf
orma
nc
e
of
bo
th
m
eth
od
s
’
s
pe
c
tr
al
ef
fi
c
i
en
c
y
ac
c
ordi
n
g t
o
i
nc
r
ea
s
e
nu
mb
er of
an
t
en
n
as
4.
Co
n
clus
ion
and
Fu
t
u
r
e
W
o
r
k
T
hi
s
p
ap
er
ha
s
pres
en
t
ed
a
c
om
m
un
i
c
ati
on
me
t
ho
d
wh
i
c
h i
s
th
e
c
om
bi
n
ed
me
tho
d
ol
o
gy
of
MM
S
E
an
d
S
IC
t
ec
hn
i
qu
e
f
or
mm
-
W
av
e
MIM
O
ba
s
ed
wi
r
e
l
es
s
c
om
m
un
i
c
ati
on
s
y
s
tem
.
T
he
c
om
b
i
ne
d
me
t
ho
d
was
propos
ed
to
r
ed
uc
e
the
r
e
l
ati
v
e
orth
og
o
na
l
c
oo
r
di
na
t
i
on
an
d
mu
l
ti
p
l
e
no
de
d
ete
c
ti
on
pro
bl
e
m
wh
i
l
e
tr
a
ns
mi
tt
er
or
r
ec
ei
v
er
mo
v
es
.
T
h
e
de
v
e
l
op
me
n
t
o
f
th
e
e
qu
ati
on
s
was
do
ne
by
c
om
pa
r
i
n
g,
r
ea
di
n
g
an
d
r
e
op
t
i
m
i
z
i
n
g
the
ex
i
s
te
d
s
ev
eral
c
o
nc
ep
ts
.
F
r
om
the
s
i
mu
l
ati
on
i
t
c
an
f
ou
nd
th
at,
th
e
pro
po
s
ed
c
om
b
i
ne
d
tec
h
ni
q
ue
f
or
w
i
r
el
es
s
p
ower
c
om
mu
ni
c
at
i
on
i
s
be
tte
r
th
an
c
o
nv
en
t
i
on
al
m
m
-
wav
e
MIM
O
.
T
ho
u
gh
,
th
e
P
r
op
os
ed
S
IC
-
M
MS
E
r
eq
ui
r
e
o
pti
mi
z
at
i
o
n
f
or
be
t
ter
pe
r
f
orma
nc
e
m
ore
c
o
m
bi
n
ed
tec
hn
i
qu
e
w
i
th
b
ett
e
r
op
t
i
m
i
z
ati
on
c
an
l
e
ad
a
be
tt
er
pe
r
for
ma
nc
e
the
n
s
i
n
gl
e
on
e
.
In
f
utu
r
e
we
wo
ul
d
l
i
k
e
to
i
mp
r
ov
e
thi
s
r
es
e
arc
h
by
ad
d
i
ng
mo
r
e
s
y
s
tem
to
g
eth
er f
or opt
i
m
al
pe
r
form
an
c
e a
nd
c
o
mp
are wi
th
r
ec
e
n
t res
ea
r
c
h.
5.
Ac
kno
w
ledg
ement
s
T
hi
s
p
ap
er
was
pa
r
t
o
f
w
ork
s
c
on
du
c
te
d
un
d
er
t
he
IIU
M
R
es
ea
r
c
h
In
i
ti
ati
v
e
G
r
an
t
S
c
he
m
e
(
RIG
S
16
-
33
4
-
04
9
8
&
RIG
S
17
-
0
31
-
0
60
6).
T
h
e
au
t
ho
r
s
wou
l
d
a
l
s
o
l
i
k
e
t
o
ac
k
no
wl
ed
g
e
al
l
s
up
p
orts
gi
v
e
n
by
t
he
II
UM
Res
e
arc
h
M
an
a
ge
m
en
t
Cent
r
e
thro
ug
h
th
e
gra
nt
an
d
R
A
Y
R
&
D
f
or the
i
r
r
es
ea
r
c
h s
up
po
r
t.
Ref
er
en
ce
s
[1
]
Aa
l
to
Un
i
v
e
rs
i
ty
,
A
T&T
,
BUPT
,
CC
,
Eri
c
s
s
o
n
,
Hu
a
wei
,
In
t
e
l
,
KT
Corp
o
ra
ti
o
n
,
No
k
i
a
,
NTT
DO
CO
M
O
,
NYU
,
Q
u
a
l
c
o
m
m
,
Sa
m
s
u
n
g
,
U.
Bri
s
to
l
,
a
n
d
USC
.
Wh
i
te
p
a
p
e
r
o
n
5
G
c
h
a
n
n
e
l
m
o
d
e
l
f
o
r
b
a
n
d
s
u
p
t
o
1
0
0
GH
z
.
2
0
1
6
O
c
t.
v
2
.3
.
[2
]
Su
n
S,
Rap
p
a
p
o
rt
TS,
Th
o
m
a
s
TA,
G
h
o
s
h
A,
Ngu
y
e
n
HC
,
Ko
v
á
c
s
IZ
,
Rod
r
i
g
u
e
z
I,
Ko
y
m
e
n
O
,
Pa
rty
k
a
A.
In
v
e
s
t
i
g
a
t
i
o
n
o
f
p
re
d
i
c
ti
o
n
a
c
c
u
ra
c
y
,
s
e
n
s
i
ti
v
i
t
y
,
a
n
d
p
a
ra
m
e
t
e
r
s
ta
b
i
l
i
ty
o
f
l
a
rg
e
-
s
c
a
l
e
p
ro
p
a
g
a
ti
o
n
p
a
th
l
o
s
s
m
o
d
e
l
s
f
o
r
5
G
wir
e
l
e
s
s
c
o
m
m
u
n
i
c
a
ti
o
n
s
.
IEEE
Tra
n
s
a
c
ti
o
n
s
o
n
Ve
h
i
c
u
l
a
r
Te
c
h
n
o
l
o
g
y
.
2
0
1
6
;
6
5
(5
):
2
8
4
3
-
60.
[3
]
Rag
h
a
v
a
n
V,
Pa
rt
y
k
a
A,
A
k
h
o
o
n
d
z
a
d
e
h
-
As
l
L
,
T
a
s
s
o
u
d
j
i
M
A
,
Ko
y
m
e
n
O
H,
Sa
n
e
l
l
i
J
.
M
i
l
l
i
m
e
te
r
wa
v
e
c
h
a
n
n
e
l
m
e
a
s
u
r
e
m
e
n
ts
a
n
d
i
m
p
l
i
c
a
t
i
o
n
s
f
o
r
PHY
l
a
y
e
r
d
e
s
i
g
n
.
IEEE
Tra
n
s
a
c
ti
o
n
s
o
n
An
te
n
n
a
s
a
n
d
Pro
p
a
g
a
ti
o
n
.
2
0
1
7
;
6
5
(1
2
)
:
6
5
2
1
-
65
3
3
.
[4
]
Rus
e
k
F
,
Pe
rs
s
o
n
D,
L
a
u
BK,
L
a
rs
s
o
n
EG
,
M
a
rz
e
t
ta
TL
,
E
d
fo
rs
O
,
Tu
f
v
e
s
s
o
n
F.
Sc
a
l
i
n
g
u
p
M
I
M
O
:
O
p
p
o
rtu
n
i
ti
e
s
a
n
d
c
h
a
l
l
e
n
g
e
s
wit
h
v
e
ry
l
a
rg
e
a
rra
y
s
.
a
rXi
v
p
r
e
p
ri
n
t
a
rXi
v
:
1
2
0
1
.
3
2
1
0
.
2
0
1
2
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
MNIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
A
wi
r
el
es
s
prec
o
di
ng
t
ec
hn
i
qu
e f
or m
i
l
l
i
me
tr
e
-
wav
e
...
(
Roun
ak
ul
Is
l
am
B
o
by
)
2789
[5
]
Roh
W,
Se
o
l
J
Y,
Pa
r
k
J
,
L
e
e
B,
L
e
e
J
,
Ki
m
Y
,
Cho
J
,
Che
u
n
K,
Ary
a
n
fa
r
F.
M
i
l
l
i
m
e
te
r
-
wa
v
e
b
e
a
m
fo
rm
i
n
g
a
s
a
n
e
n
a
b
l
i
n
g
te
c
h
n
o
l
o
g
y
f
o
r
5
G
c
e
l
l
u
l
a
r
c
o
m
m
u
n
i
c
a
ti
o
n
s
:
Th
e
o
re
ti
c
a
l
fe
a
s
i
b
i
l
i
t
y
a
n
d
p
ro
to
t
y
p
e
r
e
s
u
l
t
s
.
IEEE
c
o
m
m
u
n
i
c
a
ti
o
n
s
m
a
g
a
z
i
n
e
.
2
0
1
4
;
5
2
(2
):
1
0
6
-
1
1
3
.
[6
]
Rag
h
a
v
a
n
V,
S
u
b
ra
m
a
n
i
a
n
S,
Cez
a
n
n
e
J
,
Sa
m
p
a
t
h
A,
Ko
y
m
e
n
O
H,
L
i
J
.
S
i
n
g
l
e
-
u
s
e
r
v
e
r
s
u
s
m
u
l
t
i
-
u
s
e
r
p
re
c
o
d
i
n
g
fo
r
m
i
l
l
i
m
e
te
r
wa
v
e
M
I
M
O
s
y
s
te
m
s
.
IEEE
J
o
u
rn
a
l
o
n
Se
l
e
c
te
d
Are
a
s
i
n
Com
m
u
n
i
c
a
ti
o
n
s
.
2
0
1
7
;
3
5
(6
):
1
3
8
7
-
1
4
0
1
.
[7
]
An
d
e
rs
o
n
CR
,
Rap
p
a
o
rt
TS.
I
n
-
Bu
i
l
d
i
n
g
Wi
d
e
b
a
n
d
Pa
r
ti
t
i
o
n
L
o
s
s
M
e
a
s
u
r
e
m
e
n
ts
a
t
2
.5
G
Hz
a
n
d
6
0
G
Hz
.
a
rXi
v
p
re
p
r
i
n
t
a
rXi
v
:
1
7
0
1
.0
3
4
1
5
.
2
0
1
6
.
[8
]
Rag
h
a
v
a
n
V,
Ce
z
a
n
n
e
J
,
S
u
b
ra
m
a
n
i
a
n
S,
S
a
m
p
a
th
A,
Ko
y
m
e
n
O
.
Be
a
m
fo
r
m
i
n
g
tr
a
d
e
o
ff
s
fo
r
i
n
i
ti
a
l
UE
d
i
s
c
o
v
e
ry
i
n
m
i
l
l
i
m
e
t
e
r
-
wav
e
M
IM
O
s
y
s
te
m
s
.
IEEE
J
o
u
r
n
a
l
o
f
S
e
l
e
c
te
d
To
p
i
c
s
i
n
Si
g
n
a
l
Pro
c
e
s
s
i
n
g
.
2
0
1
6
;
1
0
(
3
):
5
4
3
-
5
59.
[9
]
Su
n
S,
Rap
p
a
p
o
rt
TS
,
He
a
th
RW,
Ni
x
A,
Ran
g
a
n
S.
M
IM
O
fo
r
m
i
l
l
i
m
e
te
r
-
w
a
v
e
wir
e
l
e
s
s
c
o
m
m
u
n
i
c
a
ti
o
n
s
:
Be
a
m
fo
rm
i
n
g
,
s
p
a
t
i
a
l
m
u
l
t
i
p
l
e
x
i
n
g
,
o
r
b
o
th
?
.
IEEE
Com
m
u
n
i
c
a
t
i
o
n
s
M
a
g
a
z
i
n
e
.
2
0
1
4
;
5
2
(1
2
):
1
1
0
-
1
2
1
.
[1
0
]
Rag
h
a
v
a
n
V,
Su
b
ra
m
a
n
i
a
n
S,
Cez
a
n
n
e
J
,
Sa
m
p
a
th
A,
K
o
y
m
e
n
O
,
L
i
J
.
Di
re
c
t
i
o
n
a
l
h
y
b
ri
d
p
re
c
o
d
i
n
g
i
n
m
i
l
l
i
m
e
te
r
-
wa
v
e
M
IM
O
s
y
s
t
e
m
s
.
2
0
1
6
IEEE
G
l
o
b
a
l
Com
m
u
n
i
c
a
t
i
o
n
s
Co
n
fe
re
n
c
e
(G
L
O
BECO
M
).
2
0
1
6
:
1
-
7.
[1
1
]
Ran
R,
Wa
n
g
J
,
O
h
SK,
Hon
g
SN.
Sp
a
rs
e
-
a
ware
m
i
n
i
m
u
m
m
e
a
n
s
q
u
a
re
e
r
ro
r
d
e
t
e
c
t
o
r
fo
r
M
IM
O
s
y
s
te
m
s
.
IEEE
Co
m
m
u
n
i
c
a
ti
o
n
s
L
e
tt
e
r
s
.
2
0
1
7
;
2
1
(1
0
):
2
2
1
4
-
221
7.
[1
2
]
Kh
a
i
z
u
ra
n
A
,
Rou
n
a
k
u
l
IB.
SIC
-
M
M
SE
M
e
t
h
o
d
b
a
s
e
d
Wi
re
l
e
s
s
Pre
c
o
d
i
n
g
Te
c
h
n
i
q
u
e
fo
r
M
i
l
l
i
m
e
tre
-
W
a
v
e
M
I
M
O
Sy
s
te
m
.
In
d
i
a
n
J
o
u
rn
a
l
o
f
S
c
i
e
n
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
.
2
0
1
9
:
1
2
(
9
):
1
-
11.
[1
3
]
Wa
n
g
Z,
L
i
M
,
L
i
u
Q
,
Swin
d
l
e
h
u
rs
t
AL
.
Hy
b
ri
d
p
re
c
o
d
e
r
a
n
d
c
o
m
b
i
n
e
r
d
e
s
i
g
n
wit
h
l
o
w
-
re
s
o
l
u
ti
o
n
p
h
a
s
e
s
h
i
f
te
rs
i
n
m
m
W
a
v
e
M
IM
O
s
y
s
te
m
s
.
IEEE
J
o
u
r
n
a
l
o
f
Se
l
e
c
t
e
d
To
p
i
c
s
i
n
Si
g
n
a
l
P
ro
c
e
s
s
i
n
g
.
2
0
1
8
;
1
2
(2
)
:
2
5
6
-
2
6
9
.
[1
4
]
El
Ay
a
c
h
O
,
Raj
a
g
o
p
a
l
S,
Ab
u
-
Su
rra
S,
Pi
Z
,
Hea
th
RW.
Sp
a
ti
a
l
l
y
s
p
a
rs
e
p
re
c
o
d
i
n
g
i
n
m
i
l
l
i
m
e
t
e
r
wav
e
M
IM
O
s
y
s
te
m
s
.
IEEE
tra
n
s
a
c
ti
o
n
s
o
n
wi
re
l
e
s
s
c
o
m
m
u
n
i
c
a
t
i
o
n
s
.
2
0
1
4
M
a
r;
1
3
(3
)
:
1
4
9
9
-
1
5
1
3
.
[1
5
]
Ve
n
u
g
o
p
a
l
K,
Al
k
h
a
te
e
b
A,
Pr
e
l
c
i
c
NG
,
He
a
th
RW.
Ch
a
n
n
e
l
e
s
t
i
m
a
ti
o
n
fo
r
h
y
b
ri
d
a
r
c
h
i
t
e
c
t
u
re
-
b
a
s
e
d
wid
e
b
a
n
d
m
i
l
l
i
m
e
te
r
wav
e
s
y
s
te
m
s
.
IEEE
J
o
u
rn
a
l
o
n
Se
l
e
c
te
d
Are
a
s
i
n
Com
m
u
n
i
c
a
ti
o
n
s
.
2
0
1
7
;
3
5
(9
):
1
9
9
6
-
2
0
0
9
.
[1
6
]
Al
k
h
a
te
e
b
A,
L
e
u
s
G
,
Hea
t
h
RW.
L
i
m
i
t
e
d
fe
e
d
b
a
c
k
h
y
b
ri
d
p
re
c
o
d
i
n
g
f
o
r
m
u
l
t
i
-
u
s
e
r
m
i
l
l
i
m
e
te
r
wav
e
s
y
s
te
m
s
.
IEEE
t
ra
n
s
a
c
ti
o
n
s
o
n
wi
re
l
e
s
s
c
o
m
m
u
n
i
c
a
ti
o
n
s
.
2
0
1
5
;
1
4
(1
1
):
6
4
8
1
-
64
9
4
.
[1
7
]
So
h
ra
b
i
F,
L
i
u
YF
,
Yu
W.
O
n
e
-
b
i
t
p
re
c
o
d
i
n
g
a
n
d
c
o
n
s
te
l
l
a
t
i
o
n
ra
n
g
e
d
e
s
i
g
n
f
o
r
m
a
s
s
i
v
e
M
IM
O
w
i
th
Q
AM
s
i
g
n
a
l
i
n
g
.
IEEE
J
o
u
rn
a
l
o
f
Se
l
e
c
te
d
T
o
p
i
c
s
i
n
Si
g
n
a
l
Pr
o
c
e
s
s
i
n
g
.
2
0
1
8
;
1
2
(
3
):
5
5
7
-
5
7
0
.
[1
8
]
Sa
d
e
g
h
i
M
,
B
j
ö
rn
s
o
n
E,
L
a
r
s
s
o
n
EG
,
Yu
e
n
C,
M
a
r
z
e
t
ta
T
L
.
M
a
x
–
m
i
n
f
a
i
r
tra
n
s
m
i
t
p
re
c
o
d
i
n
g
fo
r
m
u
l
t
i
-
g
ro
u
p
m
u
l
ti
c
a
s
ti
n
g
i
n
m
a
s
s
i
v
e
M
I
M
O
.
IEEE
Tra
n
s
a
c
ti
o
n
s
o
n
Wi
re
l
e
s
s
Com
m
u
n
i
c
a
ti
o
n
s
.
2
0
1
8
;
1
7
(2
):
1
3
5
8
-
13
7
3
.
[1
9
]
Noh
S,
Z
o
l
to
w
s
k
i
M
D,
L
o
v
e
D
J
.
Tra
i
n
i
n
g
s
e
q
u
e
n
c
e
d
e
s
i
g
n
f
o
r
fe
e
d
b
a
c
k
a
s
s
i
s
te
d
h
y
b
ri
d
b
e
a
m
f
o
rm
i
n
g
i
n
m
a
s
s
i
v
e
M
IM
O
s
y
s
te
m
s
.
IE
EE T
ra
n
s
a
c
t
i
o
n
s
o
n
Co
m
m
u
n
i
c
a
ti
o
n
s
.
2
0
1
6
;
6
4
(1
):
1
8
7
-
200.
[2
0
]
Rus
s
e
l
l
DS,
Fi
s
c
h
e
r
L
G
,
Wa
l
a
PM
.
Cel
l
u
l
a
r
c
o
m
m
u
n
i
c
a
ti
o
n
s
s
y
s
te
m
wit
h
c
e
n
tr
a
l
i
z
e
d
b
a
s
e
s
ta
ti
o
n
s
a
n
d
d
i
s
tr
i
b
u
t
e
d
a
n
te
n
n
a
u
n
i
t
s
.
US
5
,6
5
7
,
3
7
4
(P
a
te
n
t).
1
9
9
7
.
[2
1
]
O
rte
g
a
AJ
,
Sa
m
p
a
i
o
-
Net
o
R.
Ran
d
o
m
-
M
u
l
ti
-
Bra
n
c
h
Su
c
c
e
s
s
i
v
e
I
n
te
rf
e
re
n
c
e
Ca
n
c
e
l
l
a
ti
o
n
d
e
te
c
ti
o
n
i
n
s
i
n
g
l
e
-
u
s
e
r
a
n
d
m
u
l
ti
-
u
s
e
r
M
I
M
O
e
n
v
i
ro
n
m
e
n
ts
.
[2
2
]
Roh
W,
P
a
u
l
r
a
j
A.
M
IM
O
c
h
a
n
n
e
l
c
a
p
a
c
i
t
y
fo
r
th
e
d
i
s
tri
b
u
te
d
a
n
te
n
n
a
.
Pro
c
e
e
d
i
n
g
s
I
EEE
5
6
t
h
Ve
h
i
c
u
l
a
r
Te
c
h
n
o
l
o
g
y
Co
n
f
e
re
n
c
e
.
2
0
0
2
;
2
:
7
0
6
-
709.
[2
3
]
Zh
u
a
n
g
H,
Dai
L
,
Xi
a
o
L
,
Y
a
o
Y.
Sp
e
c
tra
l
e
ff
i
c
i
e
n
c
y
o
f
d
i
s
tr
i
b
u
t
e
d
a
n
te
n
n
a
s
y
s
te
m
w
i
th
ra
n
d
o
m
a
n
te
n
n
a
l
a
y
o
u
t.
E
l
e
c
tr
o
n
i
c
s
L
e
tt
e
rs
.
2
0
0
3
2
0
;
3
9
(
6
):
4
9
5
-
49
6.
[2
4
]
Cas
ta
n
h
e
i
r
a
D,
G
a
m
e
i
r
o
A.
Di
s
tri
b
u
te
d
a
n
t
e
n
n
a
s
y
s
te
m
c
a
p
a
c
i
t
y
s
c
a
l
i
n
g
[
c
o
o
r
d
i
n
a
te
d
a
n
d
d
i
s
tri
b
u
te
d
m
i
m
o
].
IEEE
Wi
re
l
e
s
s
Co
m
m
u
n
i
c
a
ti
o
n
s
.
2
0
1
0
;
1
7
(
3
):
6
8
-
75.
[2
5
]
Sh
i
d
a
M
.
Di
s
tri
b
u
te
d
a
n
te
n
n
a
s
y
s
te
m
.
Uni
t
e
d
St
a
te
s
p
a
t
e
n
t
US 8
,9
2
3
,
9
0
8
.
2
0
1
4
.
[2
6
]
Deb
b
a
h
M
,
M
u
q
u
e
t
B,
D
e
C
o
u
rv
i
l
l
e
M
,
M
u
c
k
M
,
Si
m
o
e
n
s
S,
L
o
u
b
a
to
n
P.
A
M
M
SE
s
u
c
c
e
s
s
i
v
e
i
n
te
rf
e
re
n
c
e
c
a
n
c
e
l
l
a
ti
o
n
s
c
h
e
m
e
f
o
r a
n
e
w a
d
j
u
s
ta
b
l
e
h
y
b
ri
d
s
p
re
a
d
OFDM
s
y
s
te
m
.
VTC2
0
0
0
-
Sp
ri
n
g
2
0
0
0
IEEE
5
1
st
V
e
h
i
c
u
l
a
r
Te
c
h
n
o
l
o
g
y
Con
f
e
re
n
c
e
Pr
o
c
e
e
d
i
n
g
s
(Cat
.
N
o
.
0
0
CH
3
7
0
2
6
).
2
0
0
0
;
2
:
745
-
7
4
9
.
Evaluation Warning : The document was created with Spire.PDF for Python.