T
E
L
KO
M
N
I
KA
T
e
lec
om
m
u
n
icat
ion
,
Com
p
u
t
i
n
g,
E
lec
t
r
on
ics
an
d
Cont
r
ol
Vol.
18
,
No.
5
,
Oc
tober
2020
,
pp
.
2341~2351
I
S
S
N:
1693
-
6930,
a
c
c
r
e
dit
e
d
F
ir
s
t
G
r
a
de
by
Ke
me
nr
is
tekdikti
,
De
c
r
e
e
No:
21/E
/KP
T
/2018
DO
I
:
10.
12928/
T
E
L
KO
M
NI
KA
.
v18i5.
14
640
2341
Jou
r
n
al
h
omepage
:
ht
tp:
//
jour
nal.
uad
.
ac
.
id/
index
.
php/T
E
L
K
OM
N
I
K
A
Range
expansion
method
on
heterogeneous
network
to
increase
picocell
covera
ge
H
ad
i
S
u
p
r
ia
d
i,
Has
an
ah
P
u
t
r
i
D
i
p
l
o
m
a
III
T
el
ec
o
mmu
n
i
ca
t
i
o
n
T
e
ch
n
o
l
o
g
y
,
A
p
p
l
i
e
d
Sci
en
ce
Sch
o
o
l
,
T
el
k
o
m
U
n
i
v
er
s
i
t
y
,
In
d
o
n
es
i
a
Ar
t
icle
I
n
f
o
AB
S
T
RA
CT
A
r
ti
c
le
h
is
tor
y
:
R
e
c
e
ived
Nov
19,
2019
R
e
vis
e
d
M
a
r
31,
2020
Ac
c
e
pted
Apr
30,
2020
In
t
h
i
s
s
t
u
d
y
,
p
i
c
o
cel
l
p
l
a
n
n
i
n
g
w
as
carri
e
d
o
u
t
o
n
h
et
ero
g
e
n
eo
u
s
n
e
t
w
o
rk
s
b
y
ap
p
l
y
i
n
g
t
h
e
ra
n
g
e
ex
p
an
s
i
o
n
me
t
h
o
d
.
T
h
e
cas
e
s
t
u
d
y
w
a
s
co
n
d
u
ct
e
d
i
n
Co
b
l
o
n
g
Su
b
d
i
s
t
ri
c
t
-
Ban
d
u
n
g
o
n
t
h
e
1
8
0
0
M
H
z
freq
u
en
cy
.
H
e
t
ero
g
en
e
o
u
s
n
et
w
o
r
k
(H
e
t
N
e
t
)
i
s
a
s
y
s
t
em
t
h
at
co
m
b
i
n
es
m
i
cro
ce
l
l
n
et
w
o
r
k
s
a
n
d
s
mal
l
cel
l
n
et
w
o
rk
s
(p
i
co
c
el
l
an
d
femt
o
ce
l
l
).
T
h
e
ap
p
l
i
ca
t
i
o
n
o
f
t
h
e
ran
g
e
ex
p
a
n
s
i
o
n
met
h
o
d
i
n
p
i
co
ce
l
l
w
a
s
ai
med
t
o
b
r
o
ad
e
n
t
h
e
s
co
p
e
o
f
p
i
co
ce
l
l
.
Fo
r
t
h
e
s
i
m
u
l
a
t
i
o
n
,
A
t
o
l
l
3
.
3
.
0
s
o
f
t
w
are
w
i
t
h
o
b
s
er
v
at
i
o
n
al
p
aramet
er
s
w
a
s
i
mp
l
emen
t
ed
,
i
n
cl
u
d
i
n
g
RSRP,
SIN
R,
t
h
r
o
u
g
h
p
u
t
,
an
d
u
s
er
co
n
n
ect
e
d
.
T
h
e
p
l
a
n
n
i
n
g
res
u
l
t
s
s
h
o
w
e
d
t
h
a
t
t
h
e
ap
p
l
i
cat
i
o
n
o
f
ex
p
an
s
i
o
n
met
h
o
d
i
n
crea
s
ed
t
h
e
c
o
v
era
g
e
an
d
q
u
al
i
t
y
o
f
n
et
w
o
r
k
,
w
h
ere
t
h
e
RSRP
v
al
u
e
≥
-
9
0
d
Bm
w
as
9
7
.
7
2
%
,
SIN
R
≥
5
d
B
w
as
7
0
.
9
9
%
,
u
p
l
i
n
k
t
h
ro
u
g
h
p
u
t
w
a
s
1
7
.
8
0
Mb
p
s
,
d
o
w
n
l
i
n
k
t
h
ro
u
g
h
p
u
t
w
as
2
1
.
3
7
M
b
p
s
,
an
d
u
s
er
c
o
n
n
ec
t
e
d
w
as
9
9
.
2
%
.
K
e
y
w
o
r
d
s
:
Atoll
3.
3
.
0
He
tNe
t
C
ove
r
a
ge
P
icoc
e
ll
R
a
nge
e
xpa
ns
ion
Th
i
s
i
s
a
n
o
p
en
a
c
ces
s
a
r
t
i
c
l
e
u
n
d
e
r
t
h
e
CC
B
Y
-
SA
l
i
ce
n
s
e
.
C
or
r
e
s
pon
din
g
A
u
th
or
:
Ha
s
a
na
h
P
utr
i
,
Diplom
a
I
I
I
T
e
lec
omm
unica
ti
on
T
e
c
hnology,
Applied
S
c
ienc
e
S
c
hool
,
T
e
lkom
Unive
r
s
it
y
,
T
e
lekomunikas
i
S
tr
e
e
t
,
No.
1
T
e
r
us
a
n
bua
h
ba
tu
,
B
a
ndung,
I
ndone
s
ia.
E
mail:
ha
s
a
na
hputr
i@t
e
lkom
univer
s
it
y.
a
c
.
id
1.
I
NT
RODU
C
T
I
ON
B
a
s
e
d
on
the
t
r
a
f
f
ic
da
ta
in
C
oblong
s
ubdis
tr
ict
[1
,
2]
,
the
number
of
us
e
r
s
a
c
c
e
s
s
ing
L
T
E
ne
twor
ks
on
s
e
ve
r
a
l
s
it
e
s
r
e
a
c
he
d
mor
e
than
2
,
000
us
e
r
s
pe
r
da
y
be
c
a
us
e
in
th
e
a
r
e
a
ther
e
we
r
e
s
e
ve
r
a
l
tour
is
t
a
t
tr
a
c
ti
ons
,
c
oll
e
ge
c
a
mpus
e
s
a
nd
s
hopping
c
e
nter
s
r
e
s
ult
ing
in
the
ove
r
load
c
a
pa
c
it
y.
B
e
s
ides
,
the
dr
ive
tes
t
r
e
s
ult
s
s
howe
d
s
e
ve
r
a
l
point
s
e
xpe
r
ienc
e
d
poo
r
c
ove
r
a
ge
with
a
n
a
ve
r
a
ge
R
S
R
P
va
lue
o
f
<
-
90
dB
m
due
to
a
lar
ge
n
umber
of
tall
buil
dings
a
nd
lar
ge
tr
e
e
s
that
c
a
us
e
d
the
de
c
r
e
a
s
e
of
s
ignal
qua
li
ty
.
I
n
the
c
ondit
ion
of
u
r
ba
n
a
r
e
a
s
,
one
o
f
the
s
olut
ions
to
ove
r
c
ome
th
e
pr
oblem
is
by
planni
ng
he
ter
oge
ne
ous
ne
twor
ks
in
the
f
or
m
of
picoc
e
ll
.
T
he
r
e
s
e
a
r
c
h
[3
-
5]
e
xplain
ed
the
picoc
e
ll
plannin
g
in
C
e
ntr
a
l
C
im
a
hi
s
ubdis
tr
ict
wa
s
c
onduc
ted
by
s
e
lec
ti
ng
s
it
e
s
ha
ving
high
tr
a
f
f
ic
s
.
T
he
r
e
s
ult
s
o
f
s
tudy
ba
s
e
d
on
the
r
e
f
e
r
e
nc
e
s
ignal
r
e
c
e
ived
powe
r
(
R
S
R
P
)
pa
r
a
mete
r
s
,
s
ignal
to
no
is
e
r
a
ti
o
(
S
I
NR
)
,
thr
ough
put,
a
nd
us
e
r
c
onne
c
ted
s
howe
d
that
the
picoc
e
ll
c
ould
be
the
s
olut
ion
to
ove
r
c
ome
the
of
f
load
tr
a
f
f
ic
on
c
a
p
a
c
it
y
a
nd
a
ls
o
im
p
r
ove
d
the
L
T
E
ne
two
r
k
pe
r
f
or
m
a
nc
e
on
c
ove
r
a
ge
.
T
he
r
e
f
or
e
,
the
planning
of
he
ter
oge
ne
ous
ne
twor
ks
us
ing
s
mall
c
e
ll
s
is
ve
r
y
a
ppr
op
r
iate
to
be
a
ppli
e
d
in
de
ns
e
ur
ba
n
a
r
e
a
s
that
a
ll
ow
a
s
ur
ge
o
f
c
a
pa
c
it
y.
He
ter
oge
ne
ous
ne
twor
k
is
a
s
c
he
me
int
r
oduc
e
d
by
the
thi
r
d
ge
ne
r
a
ti
on
pa
r
tner
s
hip
pr
ojec
t
(
3
GPP
)
,
whic
h
c
a
n
c
ombi
ne
mac
r
oc
e
ll
ne
twor
ks
with
s
m
a
ll
c
e
ll
s
(
picoc
e
ll
s
a
nd
f
e
mt
oc
e
ll
s
)
[6
-
8]
.
He
ter
o
ge
ne
ous
n
e
twor
k
(
He
tNe
t)
is
a
ne
twor
k
c
ons
is
ti
ng
of
mac
r
oc
e
ll
ne
twor
ks
a
nd
s
mall
c
e
ll
ne
twor
ks
,
mac
r
oc
e
ll
n
e
twor
ks
ha
ve
high
powe
r
leve
ls
;
mea
nwhile,
s
mall
c
e
ll
ne
twor
ks
ha
ve
low
t
r
a
ns
mi
s
s
ion
powe
r
or
c
omm
onl
y
r
e
f
e
r
r
e
d
to
a
s
low
powe
r
node
(
L
P
N)
ne
twor
ks
[9
,
10]
.
T
h
e
L
P
N
ne
twor
k
c
a
n
be
in
the
f
o
r
m
o
f
p
icoc
e
ll
s
,
f
e
mt
oc
e
ll
s
,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
5
,
Oc
tober
2020:
2341
-
2351
2342
r
e
mot
e
r
a
dio
he
a
ds
(
R
R
H)
,
a
nd
r
e
lay
node
s
.
On
he
ter
oge
ne
ous
ne
twor
ks
,
mac
r
oc
e
ll
s
a
r
e
r
e
s
pons
ibl
e
f
or
p
r
ovidi
ng
ove
r
a
ll
c
ove
r
a
ge
of
long
-
ter
m
e
volut
ion
(
L
T
E
)
ne
twor
k
.
M
a
c
r
oc
e
ll
s
in
he
ter
oge
ne
ous
n
e
twor
ks
a
r
e
e
Node
B
that
tr
a
ns
mi
t
high
powe
r
up
to
46
dB
m
[
11
,
12]
.
S
mall
c
e
ll
s
a
r
e
im
pleme
nted
und
e
r
the
s
c
ope
of
mac
r
oc
e
ll
s
to
maximi
z
e
the
c
ove
r
a
ge
a
nd
to
incr
e
a
s
e
the
c
a
pa
c
it
y
of
L
T
E
-
A
ne
twor
k
s
ys
tems
[
13]
.
S
mall
c
e
ll
s
pr
ovide
s
e
r
vice
s
a
t
s
e
ve
r
a
l
point
s
that
e
xpe
r
ienc
e
of
f
load
t
r
a
f
f
ic
,
e
s
pe
c
ially
on
the
e
dge
of
c
e
ll
[
14]
.
T
hus
,
the
s
mall
c
e
ll
s
c
a
n
s
e
r
ve
us
e
r
s
who
c
a
nnot
be
s
e
r
ve
d
by
mac
r
oc
e
ll
s
.
W
it
h
low
powe
r
a
r
ound
23
-
30
dB
m,
s
mall
c
e
ll
s
only
ha
ve
a
s
mall
c
ove
r
a
ge
(
up
to
300
m)
;
the
r
e
f
or
e
,
i
n
the
he
ter
oge
ne
ous
ne
twor
ks
,
s
m
a
ll
c
e
ll
s
will
pr
ovide
maximu
m
c
ove
r
a
ge
[
15
]
.
W
it
h
a
good
c
oll
a
bor
a
ti
on
be
twe
e
n
mac
r
oc
e
ll
s
a
nd
s
m
a
ll
c
e
ll
s
,
he
ter
oge
ne
ous
ne
twor
ks
will
pr
ovide
f
lexible,
ine
xpe
ns
ive
a
nd
e
f
f
icie
nt
s
olut
ions
to
incr
e
a
s
e
the
c
ove
r
a
ge
a
nd
c
a
pa
c
it
y
of
L
T
E
-
Adva
nc
e
d
s
ys
tems
[
16
,
17
]
.
T
he
main
pr
oblem
in
he
ter
oge
ne
ous
ne
twor
k
p
lanning
is
the
ne
c
e
s
s
it
y
to
e
ns
ur
e
that
the
p
icoc
e
ll
s
c
a
n
s
e
r
ve
the
us
e
r
s
on
the
c
e
ll
e
dge
a
r
e
a
pr
ope
r
ly
.
One
of
wa
ys
to
ove
r
c
ome
thi
s
pr
oblem
is
by
incr
e
a
s
ing
a
r
e
a
s
s
e
r
ve
d
by
the
picoc
e
ll
s
,
whic
h
is
c
a
ll
e
d
a
s
r
a
nge
e
xpa
ns
ion
[
18
,
19]
.
I
n
the
r
a
nge
e
xpa
ns
ion
s
ys
tem,
a
n
of
f
s
e
t
bias
is
a
dde
d
to
the
R
S
R
P
picoc
e
ll
s
to
incr
e
a
s
e
the
c
ove
r
a
ge
.
B
e
f
or
e
us
ing
the
r
a
nge
e
xpa
ns
ion
method,
the
us
e
r
s
a
t
the
e
dge
of
c
e
ll
would
c
hoos
e
to
be
s
e
r
ve
d
by
ma
c
r
oc
e
ll
s
be
c
a
us
e
of
the
high
R
S
R
P
va
lue
[
20
,
21
]
.
I
n
the
non
-
r
a
nge
e
xpa
ns
ion
method,
the
s
e
lec
ti
on
of
s
e
r
ving
c
e
ll
s
wa
s
ba
s
e
d
on
the
highes
t
R
S
R
P
va
lue
or
c
omm
only
c
a
ll
e
d
the
maximum
R
S
R
P
.
W
he
r
e
a
s
,
a
f
ter
us
ing
t
he
r
a
nge
e
xpa
ns
ion
method,
the
s
e
r
ving
c
e
ll
wa
s
c
hos
e
n
ba
s
e
d
on
the
R
S
R
P
va
lue
plus
the
of
f
s
e
t
bias
of
the
picoc
e
ll
.
W
he
n
the
of
f
s
e
t
b
ias
va
lue
=
0
dB
,
then
the
us
e
r
would
be
s
e
r
ve
d
by
a
mac
r
oc
e
ll
,
whe
r
e
a
s
if
the
o
f
f
s
e
t
bias
va
lue
=
1
-
6
dB
,
the
us
e
r
would
be
s
e
r
ve
d
by
a
pico
c
e
ll
[
22]
.
I
nter
-
c
e
ll
int
e
r
f
e
r
e
nc
e
c
oor
dination
(
I
C
I
C
)
wa
s
in
tr
oduc
e
d
in
L
T
E
r
e
lea
s
e
8
a
s
a
method
that
c
a
n
ove
r
c
ome
the
pr
oblem
o
f
int
e
r
f
e
r
e
nc
e
wit
h
us
e
r
s
a
t
the
e
dge
of
c
e
ll
,
whe
r
e
int
e
r
e
Node
B
c
a
n
c
om
muni
c
a
te
thr
ough
the
X2
int
e
r
f
a
c
e
to
r
e
duc
e
int
e
r
-
c
e
ll
int
e
r
f
e
r
e
nc
e
.
I
n
he
ter
oge
ne
ous
ne
twor
ks
,
to
make
the
u
s
e
r
s
c
a
n
ge
t
s
e
r
vice
s
f
r
om
picoc
e
ll
s
with
s
ignal
s
tr
e
ngth
f
r
o
m
s
tr
onge
r
mi
c
r
oc
e
ll
s
,
picoc
e
ll
s
ne
e
d
to
do
the
c
oo
r
dination
of
c
ontr
ol
c
ha
nne
l
a
nd
da
ta
c
ha
nne
l
int
e
r
f
e
r
e
nc
e
with
mor
e
domi
na
nt
mac
r
o
c
e
ll
in
ter
f
e
r
e
nc
e
.
I
n
L
T
E
r
e
lea
s
e
10,
it
wa
s
int
r
oduc
e
d
e
nha
nc
e
d
int
e
r
-
c
e
ll
int
e
r
f
e
r
e
n
c
e
c
oor
dination
(
e
I
C
I
C
)
,
whic
h
is
the
de
ve
lopm
e
nt
of
I
C
I
C
method
th
a
t
s
uppor
t
he
ter
oge
ne
ous
ne
twor
ks
[
23
,
24]
.
T
he
di
f
f
e
r
e
nc
e
is
in
the
a
ddit
ion
of
the
I
C
I
C
domain
ti
me,
whe
r
e
thi
s
method
us
e
s
a
lm
os
t
blank
s
ubf
r
a
me
(
AB
S
)
.
W
he
n
a
m
a
c
r
oc
e
ll
c
onf
igu
r
e
s
a
nd
s
e
nds
inf
or
mation
to
a
picoc
e
ll
a
bout
the
AB
S
pa
tt
e
r
n
,
the
us
e
r
c
a
n
c
onne
c
t
to
picoc
e
ll
a
nd
r
e
c
e
ive
d
ownlink
inf
or
mation,
both
c
ontr
ol
a
nd
us
e
r
da
ta.
T
he
pr
e
s
e
nt
s
tudy
opti
mi
z
e
d
the
c
ove
r
a
ge
o
f
L
T
E
ne
twor
ks
that
e
xpe
r
ienc
e
d
of
f
load
c
a
pa
c
it
y,
by
planning
he
ter
oge
ne
ous
ne
twor
ks
in
the
f
o
r
m
of
picoc
e
ll
s
a
t
1
,
800
M
Hz
f
r
e
que
nc
y
a
nd
a
pplyi
n
g
r
a
nge
e
xpa
ns
ion
to
e
xpa
nd
the
c
ove
r
a
ge
of
pi
c
oc
e
ll
s
.
R
a
nge
e
xpa
ns
ion
c
a
n
maximi
z
e
the
pe
r
f
or
manc
e
of
He
tNe
t
be
c
a
us
e
it
c
a
n
im
p
r
ove
the
c
ove
r
a
ge
of
downlink
s
on
picoc
e
ll
s
[
25]
.
I
n
picoc
e
ll
p
lanning,
it
wa
s
p
e
r
f
or
med
the
c
a
lcula
ti
on
of
li
nk
budge
t
a
nd
c
a
pa
c
it
y
dim
e
n
s
ioni
ng
[
26
,
27
]
.
T
he
n,
it
wa
s
c
ompar
e
d
to
the
r
e
s
ult
s
of
planning
he
ter
oge
ne
ous
ne
twor
ks
a
nd
ne
twor
ks
wi
thout
he
ter
oge
ne
ous
(
homogene
ous
)
.
T
his
a
r
ti
c
le
is
one
of
the
r
e
c
omm
e
nda
ti
ons
f
or
c
e
ll
ular
ne
twor
k
p
r
ovider
s
in
opti
mi
z
ing
the
c
ove
r
a
ge
o
f
he
ter
oge
ne
ous
ne
tw
or
ks
a
nd
is
a
r
e
f
e
r
e
nc
e
f
o
r
r
e
s
e
a
r
c
he
r
s
in
he
ter
oge
ne
ous
ne
twor
k
r
e
s
e
a
r
c
h.
2.
RE
S
E
AR
CH
M
E
T
HO
D
I
n
thi
s
s
e
c
ti
on,
it
is
e
xplaine
d
the
s
tage
s
of
picoc
e
ll
planning
on
the
He
tNe
t
ne
twor
k
by
a
pplyi
ng
the
r
a
nge
e
xpa
ns
ion
method.
T
he
s
tage
s
in
thi
s
p
lan
a
r
e
in
a
c
c
or
da
nc
e
with
the
c
e
ll
ula
r
pr
ovider
s
tanda
r
d
ope
r
a
ti
ng
pr
oc
e
dur
e
(
S
O
P
)
.
T
he
planning
wa
s
s
tar
te
d
by
e
va
luating
ge
ogr
a
phica
l
c
ondit
ions
,
e
v
a
luating
e
xis
ti
ng
s
it
e
s
,
e
va
luating
ini
ti
a
l
dr
ive
tes
t,
c
ove
r
a
ge
a
nd
c
a
pa
c
it
y
dim
e
ns
ioni
ng,
modeling
the
He
tNe
t
ne
twor
k
with
r
a
nge
e
xpa
ns
ion,
a
nd
c
onf
igur
ing
mi
c
r
oc
e
ll
s
a
nd
picoc
e
ll
s
.
2.
1.
T
h
e
e
valu
a
t
ion
o
f
ge
og
r
ap
h
ical
c
on
d
it
ion
s
Ge
ogr
a
phica
l
c
ondit
ions
in
the
C
oblong
s
ubdi
s
tr
ict
a
r
e
a
indi
c
a
te
that
s
e
ve
r
a
l
point
s
be
c
ome
the
c
e
nter
of
the
c
r
owd,
whic
h
c
a
n
c
a
us
e
ove
r
load
c
a
pa
c
it
y.
Als
o,
the
ge
ogr
a
phica
l
s
it
ua
ti
on
whe
r
e
ther
e
a
r
e
many
lar
ge
tr
e
e
s
a
nd
une
ve
n
e
a
r
th
c
ontou
r
s
c
a
n
c
a
us
e
poor
c
ove
r
a
ge
,
whe
r
e
us
e
r
s
do
not
ge
t
maximum
s
e
r
vic
e
a
t
that
point
.
T
he
obs
e
r
va
ti
on
wa
s
c
a
r
r
ied
out
in
4
(
f
our
)
l
oc
a
ti
ons
including
J
a
lan
I
r
.
H.
J
ua
nda
(
Da
go)
,
Ga
s
i
bu
F
ield,
B
a
ndung
Z
oo,
a
nd
C
ihampe
las
.
F
igur
e
1
s
hows
the
s
ur
ve
y
r
e
s
ult
s
of
the
ge
ogr
a
phica
l
c
ondit
ions
of
the
obs
e
r
va
ti
on
a
r
e
a
.
I
r
.
H.
J
ua
nda
S
t
r
e
e
t
ha
s
une
ve
n
c
ontour
s
a
nd
tend
s
to
uphil
l,
a
ls
o
ther
e
a
r
e
many
tr
e
e
s
f
ound
ther
e
.
As
one
of
point
s
f
o
r
the
S
unda
y
c
a
r
-
f
r
e
e
da
y
e
ve
nt
i
n
the
c
it
y
o
f
B
a
ndung,
th
is
s
tr
e
e
t
is
a
ls
o
the
c
e
nter
o
f
c
r
owd.
M
e
a
nwhile,
in
Ga
s
ibu,
ther
e
is
a
ma
r
ke
t
on
S
und
a
ys
.
Ga
s
ibu
is
f
r
e
que
ntl
y
us
e
d
a
s
a
ve
nue
f
or
mus
ic
e
ve
nts
vis
it
e
d
by
many
pe
ople
.
Als
o,
the
Ga
s
ibu
f
ield
ha
s
a
lwa
ys
be
e
n
one
of
the
c
e
nter
of
c
r
owd
du
r
in
g
t
he
Ne
w
Ye
a
r
's
E
ve
a
nd
T
a
kbe
e
r
night
e
ve
nt.
T
he
r
e
f
or
e
,
the
s
e
r
vice
pr
ovider
mus
t
maximi
z
e
it
s
ne
twor
k
to
mee
t
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
R
ange
e
x
pans
ion
me
thod
on
he
ter
oge
ne
ous
ne
tw
or
k
to
incr
e
as
e
picoc
e
ll
c
ov
e
r
age
(
Hadi
Supr
iadi
)
2343
the
ne
e
ds
of
us
e
r
s
ther
e
.
Ac
c
or
ding
to
C
NN
I
ndone
s
ia
(
5/6)
,
the
number
o
f
vis
it
or
s
to
the
B
a
ndung
Z
o
o
dur
ing
the
E
id
holi
da
y
wa
s
28
thous
a
nd
pe
ople,
whic
h
c
a
us
e
d
the
ove
r
load
c
a
pa
c
it
y.
B
e
s
ides
,
in
the
B
a
nd
ung
Z
oo
,
ther
e
a
r
e
s
ha
dy
tr
e
e
s
that
c
a
n
c
a
u
s
e
poor
c
ove
r
a
ge
.
T
hus
,
the
a
ddit
ion
o
f
picoc
e
ll
s
is
ne
e
de
d
to
opti
mi
z
e
the
ne
twor
k
in
ter
ms
o
f
c
ove
r
a
ge
a
nd
c
a
pa
c
it
y.
On
C
ihampe
las
s
tr
e
e
t,
many
s
hop
ping
c
e
nter
s
a
r
e
f
r
e
que
ntl
y
vis
it
e
d
by
vis
it
o
r
s
f
r
om
va
r
ious
r
e
gions
.
T
he
r
e
is
a
ls
o
a
s
kywa
lk
o
r
a
pe
de
s
tr
ian
br
idge
buil
t
on
the
4
50
-
mete
r
br
idge,
whic
h
c
a
n
r
e
duc
e
the
qua
li
ty
of
s
ignal
f
o
r
u
s
e
r
s
be
low
it
.
F
igur
e
1
.
T
he
s
ur
ve
y
r
e
s
ult
o
f
obs
e
r
va
ti
on
a
r
e
a
s
2.
2.
T
h
e
e
valu
a
t
ion
o
f
i
n
it
ial
d
r
ive
t
e
s
t
B
a
s
e
d
on
the
mea
s
ur
e
ment
of
s
ignal
by
the
dr
ive
tes
t
method
us
ing
T
E
M
S
P
oc
ke
t
s
of
twa
r
e
in
C
oblong
S
ubdis
tr
ict,
the
r
e
s
ult
s
s
hown
in
F
igur
e
s
2
a
nd
3
.
F
r
om
F
igur
e
2
,
it
c
a
n
be
s
e
e
n
that
the
r
e
s
ult
of
R
S
R
P
p
a
r
a
mete
r
r
e
por
ti
ng
of
the
L
T
E
ne
two
r
k
c
ondit
ion
in
C
oblong
S
ubdis
tr
ict
–
B
a
ndung.
S
e
ve
r
a
l
point
s
ha
d
poor
ne
twor
k
c
ondit
ions
with
ye
ll
ow
a
nd
r
e
d
indi
c
a
tor
s
,
whic
h
wa
s
a
n
R
S
R
P
va
lue
<
90
d
B
m
with
a
pe
r
c
e
ntage
o
f
53.
7%
,
whic
h
mea
ns
that
the
por
ti
on
of
C
oblong
s
ubdis
tr
ict
ha
d
a
non
-
opti
mal
ne
twor
k
in
ter
ms
of
c
ove
r
a
ge
.
F
igur
e
3
s
hows
the
r
e
por
ti
ng
r
e
s
ult
s
f
or
the
S
I
NR
<
5
dB
pa
r
a
mete
r
with
a
r
e
d
indi
c
a
tor
ha
ving
a
pe
r
c
e
ntage
o
f
45.
91%
.
T
he
s
ize
o
f
the
S
I
NR
va
lue
will
a
f
f
e
c
t
the
va
lue
of
thr
oughput
ob
taine
d
by
the
us
e
r
.
F
r
om
the
r
e
por
ti
n
g
a
bove
,
it
c
a
n
be
c
onc
luded
that
the
c
ondit
ion
of
L
T
E
ne
t
wor
k
in
the
C
oblong
s
ubdis
tr
ict
wa
s
not
opt
im
a
l
b
e
c
a
us
e
it
did
no
t
a
c
hieve
the
s
tanda
r
d
ope
r
a
tor
of
R
S
R
P
pa
r
a
mete
r
s
,
whic
h
mus
t
ha
ve
a
pe
r
c
e
ntage
of
90%
90
-
90
dB
m
a
nd
S
I
NR
90%
≥
5
dB
.
T
his
pr
ob
lem
c
a
n
be
r
e
s
ult
ed
f
r
om
s
e
ve
r
a
l
f
a
c
tor
s
,
s
uc
h
a
s
the
in
f
luenc
e
of
ge
o
gr
a
phica
l
c
ondit
ions
,
the
number
of
tall
buil
dings
,
a
nd
t
he
number
of
s
ha
de
tr
e
e
s
or
of
f
load
t
r
a
f
f
ic
.
M
or
e
ove
r
,
the
C
oblong
Dis
tr
ict
ha
s
s
e
ve
r
a
l
plac
e
s
that
f
unc
ti
on
a
s
the
c
e
nter
of
c
r
owd,
s
uc
h
a
s
s
hopping
c
e
nter
s
,
tour
is
t
a
tt
r
a
c
ti
ons
,
a
nd
c
oll
e
ge
c
a
mpus
e
s
.
T
he
p
r
oblem
in
t
he
C
oblong
s
ubdis
tr
ict
oc
c
ur
ed
be
c
a
u
s
e
only
the
m
a
c
r
oc
e
ll
s
pr
ovided
the
c
ove
r
a
ge
.
T
he
r
e
f
or
e
,
i
t
is
ne
c
e
s
s
a
r
y
to
c
onduc
t
the
opti
mi
z
a
ti
on
to
maxim
ize
the
e
xis
ti
ng
ne
twor
k
s
e
r
vice
s
,
one
of
them
is
by
c
onduc
ti
ng
the
picoc
e
l
l
planning
a
t
point
s
that
mi
ght
oc
c
ur
of
f
loading
tr
a
f
f
ic
a
nd
poor
c
ove
r
a
ge
,
whe
r
e
the
picoc
e
ll
s
will
he
lp
the
m
i
c
r
oc
e
ll
s
to
s
e
r
ve
the
us
e
r
s
who
a
r
e
a
t
the
c
e
nter
of
c
r
owd
or
poor
c
ove
r
a
ge
.
F
igur
e
2.
R
e
por
ti
ng
R
S
R
P
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
5
,
Oc
tober
2020:
2341
-
2351
2344
F
igur
e
3.
R
e
por
ti
ng
S
I
NR
2.
3.
T
h
e
e
valu
a
t
ion
o
f
e
xis
t
in
g
s
it
e
s
I
n
the
C
oblong
s
ubdis
tr
ict
,
a
n
a
na
lys
is
of
the
c
a
pa
c
it
y
of
e
xis
ti
ng
s
it
e
s
wa
s
c
a
r
r
ied
ou
t
by
obs
e
r
ving
s
e
ve
r
a
l
s
it
e
s
that
ha
d
h
igh
t
r
a
f
f
ic
s
.
T
he
r
e
f
or
e
,
t
he
he
ter
oge
ne
ous
ne
twor
k
planning
c
ould
be
c
onduc
ted,
by
a
dding
picoc
e
ll
s
in
s
e
ve
r
a
l
s
it
e
s
ha
ving
ove
r
load
c
a
pa
c
it
y.
B
a
s
e
d
on
the
tr
a
f
f
ic
da
ta
obtain
e
d
f
r
om
the
ope
r
a
tor
,
ther
e
we
r
e
5
s
it
e
s
in
the
C
oblong
s
u
bdis
tr
ict
ha
ving
high
tr
a
f
f
ic
s
.
T
a
ble
1
s
hows
the
t
otal
us
e
r
s
f
r
om
the
f
ive
s
it
e
s
in
C
oblong
Dis
tr
ict
take
n
3
da
ys
in
a
r
ow.
T
he
f
ive
s
it
e
s
ha
d
a
tot
a
l
of
us
e
r
s
in
a
da
y
r
e
a
c
hing
mor
e
than
2
,
000
us
e
r
s
.
I
t
s
howe
d
that
the
numb
e
r
of
us
e
r
s
a
c
c
e
s
s
ing
the
L
T
E
ne
twor
k
wa
s
hig
h.
T
hus
,
he
ter
oge
ne
ous
ne
twor
k
planning
wa
s
ne
e
d
e
d
to
make
the
tr
a
f
f
ic
s
ur
ge
in
mac
r
oc
e
ll
s
c
a
n
be
ha
ndled
by
s
mall
c
e
ll
s
,
whic
h
then
les
s
e
n
the
bur
de
n
on
mac
r
oc
e
ll
s
t
ha
t
ha
ve
high
t
r
a
f
f
ic
s
.
T
a
ble
1.
T
r
a
f
f
ic
da
ta
of
s
ite
s
S
it
e
_N
a
me
D
a
te
T
ot
a
l
U
s
e
r
N
umbe
r
24 Hour
L
T
E
_I
S
T
A
N
A
_D
A
G
O
26/
01/
2019
2196
27/
01/
2019
2500
28/
01/
2019
3395
L
T
E
_I
R
_H
A
J
I
_J
U
A
N
D
A
26/
01/
2019
2764
27/
01/
2019
2161
28/
01/
2019
2022
L
T
E
_S
U
M
U
R
B
A
N
D
U
N
G
26/
01/
2019
2460
27/
01/
2019
2365
28/
01/
2019
2211
L
T
E
_T
A
M
A
N
S
A
R
I
I
T
B
A
X
26/
01/
2019
2891
27/
01/
2019
3572
28/
01/
2019
2640
L
T
E
_P
A
S
T
E
U
R
_C
I
P
E
D
E
S
26/
01/
2019
2489
27/
01/
2019
2462
28/
01/
2019
2786
2.
4.
T
h
e
m
o
d
e
ll
in
g
of
h
e
t
e
r
oge
n
e
ou
s
n
e
t
wor
k
T
he
s
olut
ion
whic
h
is
e
f
f
icie
nt
to
ha
ndle
the
inc
r
e
a
s
ing
tr
a
f
f
ic
de
mand
by
us
ing
s
mall
c
e
ll
s
ins
tea
d
a
dd
mor
e
B
T
S
mac
r
os
.
T
he
he
ter
oge
ne
ous
ne
twor
k
planning
wa
s
c
a
r
r
ied
out
by
a
pplyi
ng
s
mall
c
e
ll
s
withi
n
the
c
ove
r
a
ge
a
r
e
a
of
mac
r
oc
e
ll
s
.
F
igur
e
4
s
hows
the
he
ter
oge
ne
ous
ne
twor
k
modeling
with
the
a
ppl
i
c
a
ti
on
of
r
a
nge
e
xpa
ns
ion
method
.
I
n
thi
s
he
ter
oge
ne
ous
ne
t
wor
k
planning
,
the
s
m
a
ll
c
e
ll
us
e
d
wa
s
a
picoc
e
ll
th
a
t
wor
ks
a
t
a
f
r
e
que
nc
y
of
1
,
800
M
Hz
or
e
qua
l
to
the
wor
k
ing
f
r
e
que
nc
y
a
t
a
mac
r
oc
e
ll
a
nd
a
ba
ndwidth
of
10
M
Hz
.
P
icoc
e
ll
s
we
r
e
plac
e
d
a
t
the
e
dge
of
the
m
a
c
r
oc
e
ll
s
c
ove
r
a
ge
to
maximi
z
e
c
ove
r
a
ge
a
nd
to
incr
e
a
s
e
the
ne
twor
k
ca
pa
c
it
y.
T
he
r
a
nge
of
picoc
e
ll
s
is
ve
r
y
li
mi
ted
inf
l
ue
nc
e
d
by
powe
r
a
nd
int
e
r
f
e
r
e
nc
e
f
r
om
s
tr
onge
r
ma
c
r
oc
e
ll
s
,
e
s
pe
c
ially
a
t
the
e
dge
of
c
e
ll
.
T
he
r
a
nge
e
xpa
ns
ion
method
will
a
dd
of
f
s
e
t
bias
to
the
r
e
c
e
ived
s
ignal
s
tr
e
ngth
(
R
S
S
)
of
s
mall
c
e
ll
s
to
incr
e
a
s
e
down
li
nk
(
DL
)
c
ove
r
a
ge
.
T
he
o
f
f
s
e
t
bias
va
lue
that
c
a
n
be
us
e
d
f
or
picoc
e
ll
s
is
be
twe
e
n
0
-
6
d
B
.
W
he
n
the
of
f
s
e
t
va
lue
is
0
dB
,
then
the
picoc
e
ll
doe
s
not
us
e
the
r
a
nge
e
xp
a
ns
ion
method.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
R
ange
e
x
pans
ion
me
thod
on
he
ter
oge
ne
ous
ne
tw
or
k
to
incr
e
as
e
picoc
e
ll
c
ov
e
r
age
(
Hadi
Supr
iadi
)
2345
T
hus
,
the
s
e
r
ving
c
e
ll
will
be
pos
s
ibl
y
c
onduc
ted
by
mac
r
oc
e
ll
s
,
e
s
pe
c
ially
f
o
r
the
us
e
r
s
who
a
r
e
on
the
e
dge
of
c
e
ll
.
How
e
ve
r
,
whe
n
picoc
e
ll
s
ha
ve
a
n
of
f
s
e
t
bias
,
the
us
e
r
s
who
a
r
e
withi
n
the
picoc
e
ll
s
s
c
ope
s
a
r
e
not
s
e
r
ve
d
by
mac
r
oc
e
ll
s
.
T
a
ble
2
s
hows
the
plannin
g
s
c
e
na
r
ios
including
(
1)
the
s
im
ulation
o
f
e
xis
ti
ng
s
it
e
s
a
nd
(
2)
the
s
im
ulation
with
picoc
e
ll
planning
by
plac
e
ment
ba
s
e
d
on
the
point
of
ove
r
load
c
a
pa
c
it
y
a
nd
poor
c
ove
r
a
ge
.
F
igur
e
4
.
He
tNe
t
modelli
ng
T
a
ble
2.
T
he
p
lanning
of
s
c
e
na
r
io
S
ys
te
m P
a
r
a
me
te
r
S
c
e
na
r
io
1
S
c
e
na
r
io
2
M
a
c
r
o
P
ic
o
M
a
c
r
o
P
ic
o
F
r
e
q
ue
n
cy
(
M
H
z)
1
,
800
-
1
,
800
1
,
800
S
it
e
N
umbe
r
16
0
16
15
2
.
5.
Cove
r
age
p
lan
n
in
g
To
c
onduc
t
the
c
ove
r
a
ge
planning
c
a
lcula
ti
on
r
e
quir
e
s
the
s
pe
c
if
ica
ti
ons
of
de
vice
us
e
d.
T
a
ble
3
s
hows
the
s
pe
c
if
ica
ti
on
of
Hua
we
i
de
vice
s
us
e
d
in
thi
s
c
ove
r
a
ge
planning
.
Ne
xt
is
c
a
lcula
t
ing
the
po
we
r
li
nk
budge
t
to
ge
t
the
va
lue
of
M
a
xim
um
Al
lowe
d
P
a
t
h
L
os
s
(
M
APL
)
.
T
a
ble
4
s
hows
the
r
e
s
ult
of
c
a
lcula
ti
on
of
P
icoc
e
ll
L
ink
B
udge
t.
Af
ter
obtaining
the
va
lue
of
M
APL
picoc
e
ll
,
the
lowe
s
t
va
lue
wa
s
take
n
a
s
a
d
ownlink
M
APL
f
o
r
picoc
e
ll
s
wi
th
a
va
lue
o
f
125
.
427
dB
m.
T
he
M
APL
va
lue
wa
s
us
e
d
to
c
a
lcula
te
the
c
e
ll
r
a
d
ius
.
T
he
c
a
lcula
ti
on
of
c
e
ll
r
a
dius
wa
s
pe
r
f
or
med
us
ing
in
(
1
)
a
nd
(
2
)
.
a
(
hm)
=
(
1
.
1
log(
f
c
)
–
0
.
7
)
hm
–
(1
.
56
log(
f
c
)
–
0
.
8)
(
1)
a
(
hm)
=
(
1
.
1
log(
1
,
800)
–
0
.
7)
1.
5
–
(
1.
56
log(
1
,
80
0)
–
0
.
8)
a
(
hm
)
=
4.
320
–
4
.
278
a
(
hm
)
=
0.
042
M
APL
=
46.
3
+
33
.
9
(
L
og
f
c
)
–
13.
82
L
og
hb
–
a
(
hm
)
+
(
44.
9
–
6.
55
L
og
hb
)
log
d
(
2)
125,
427
=
46
.
3
+
33.
9
(
log
1
,
800)
–
13
.
82
log
(
7
)
–
0
.
042
+
(
44
.
9
–
6.
55
log
(
7)
)
log
d
125
,
427
=
46
.
3
+
33
.
9
(
3
.
255
)
–
13
.
82
(
0
.
845)
–
0
.
042
+
(
44
.
9
–
6.
55
(
0
.
845)
)
l
og
d
125,
427
=
46
.
3
+
110,
344
–
11
,
677
–
0.
042
+
(
44
.
9
–
5
,
534
)
log
d
125,
427
=
46
.
3
+
110,
344
–
11
,
677
–
0.
042
+
39,
3
66
log
d
125,
427
=
144
,
925
+
39.
366
log
d
-
19.
498
=
39.
366
log
d
−
19
.
498
39
.
366
=
log
d
-
0.
495
=
log
d
10
-
0.
495
=
d
d
=
0.
319
km
Af
ter
obtaining
the
va
lue
of
c
e
ll
r
a
dius
,
the
ne
x
t
s
tep
wa
s
pe
r
f
or
mi
ng
the
c
e
ll
a
r
e
a
c
a
lcula
ti
on
t
o
c
ove
r
the
e
nti
r
e
a
r
e
a
.
T
he
c
e
ll
a
r
e
a
c
a
n
be
c
a
lcula
ted
us
ing
i
n
(
3
)
.
L
c
e
ll
=
2.
6
x
d
2
(
3)
L
c
e
ll
=
2.
6
x
(
0.
319
)
2
L
c
e
ll
=
0.
264
km
2
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
5
,
Oc
tober
2020:
2341
-
2351
2346
T
a
ble
3.
T
he
s
pe
c
if
ica
ti
ons
of
de
vice
U
pl
in
k
D
ow
nl
in
k
P
a
r
a
me
te
r
of
T
r
a
ns
mi
tt
e
r
(
U
E
)
P
a
r
a
me
te
r
of
T
r
a
ns
mi
tt
e
r
(
e
N
ode
B
)
P
a
r
a
me
te
r
V
a
lu
e
P
a
r
a
me
te
r
V
a
lu
e
M
a
x T
x P
ow
e
r
23 dB
m
M
a
x T
x P
ow
e
r
30 dB
m
A
nt
e
nna
H
e
ig
ht
1.5 m
G
a
in
T
x
17 dB
i
G
a
in
T
x
0 dB
i
A
nt
e
nna
H
e
ig
ht
7 m
P
a
r
a
me
te
r
of
R
e
c
e
iv
e
r
(
e
N
ode
B
)
C
a
be
l
L
os
s
0.5 dB
G
a
in
R
x
17 dB
i
T
he
r
ma
l
N
oi
s
e
-
132.2 dB
m
N
oi
s
e
F
ig
ur
e
2.3 dB
N
oi
s
e
F
ig
ur
e
2.3 dB
T
he
r
ma
l
N
oi
s
e
-
132.2 dB
m
P
a
r
a
me
te
r
of
R
e
c
e
iv
e
r
(
U
E
)
C
a
be
l
L
os
s
0.5 dB
A
nt
e
nna
H
e
ig
ht
1.5 m
A
nt
e
nna
H
e
ig
ht
7
B
ody L
os
s
3 dB
T
a
ble
4.
T
he
ca
lcula
ti
on
of
M
APL
Tx
F
or
mul
a
UL
DL
M
a
x. T
ot
a
l
T
x P
ow
e
r
(
dB
m)
A
23
30
R
B
t
o D
is
tr
ib
ut
e
C
3
50
S
ubc
a
r
r
ie
r
s
t
o D
is
tr
ib
ut
e
P
ow
e
r
D
=
12*C
36
600
S
ubc
a
r
r
ie
r
P
ow
e
r
(
dB
m)
E
=
A
-
10*L
og(
D
)
7,
436
2,
218
T
x A
nt
e
nna
G
a
in
(
dB
i)
G
18
18
T
x C
a
bl
e
L
os
s
(
dB
)
H
0.5
0
E
I
R
P
pe
r
S
ubc
a
r
r
ie
r
(
dB
m)
J
=
E
+
G
-
H
23,
936
20,
218
Rx
F
or
mul
a
UL
DL
S
I
N
R
(
dB
)
K
-
6
.
19
-
6.
19
R
x N
oi
s
e
F
ig
ur
e
(
dB
)
L
2.
3
7
R
e
c
e
iv
e
r
S
e
ns
it
iv
it
y (
dB
m)
M
=
K
+
L
–
174 +
10*L
og10(
15000)
-
136,
129
-
131,
429
R
x B
ody L
os
s
(
dB
)
P
0
0
I
nt
e
r
f
e
r
e
nc
e
M
a
r
gi
n (
dB
)
Q
2.
72
2.72
M
in
. S
ig
na
l
R
e
c
e
pt
io
n S
tr
e
ngt
h
(
dB
m)
R
=
M
+
P
+
Q
-
133.409
-
128.209
P
a
th
L
os
s
F
or
mul
a
UL
DL
P
e
ne
tr
a
ti
on L
os
s
(
dB
)
S
15
15
S
ha
dow
F
a
di
ng M
a
r
gi
n (
dB
)
T
8
8
M
A
P
L
T
=
J
-
R
-
S
-
T
134,
345
125,
427
2.
6.
Capacit
y
p
lan
n
in
g
T
o
de
ter
m
ine
the
picoc
e
ll
s
c
a
pa
c
it
y
de
pe
nds
on
t
he
a
ll
oc
a
ted
ba
ndwidth.
I
n
the
planning
pr
oc
e
s
s
include
s
the
c
a
lcula
ti
on
of
f
or
e
c
a
s
ti
ng
n
u
mber
of
u
s
e
r
s
,
s
e
r
vice
m
ode
l
p
a
r
a
mete
r
s
,
t
r
a
f
f
ic
m
ode
l,
s
it
e
c
a
pa
c
it
y,
a
nd
n
umber
o
f
s
it
e
s
.
F
or
e
c
a
s
ti
ng
n
umber
of
u
s
e
r
c
a
l
c
ulation
include
s
the
number
of
r
e
s
idents
in
the
obs
e
r
va
ti
on
a
r
e
a
,
population
gr
owth
f
a
c
tor
s
,
pr
oduc
ti
ve
a
ge
po
pulation
(
a
ge
d
15
-
65
ye
a
r
s
)
,
c
e
ll
phone
us
e
r
s
,
ma
r
ke
t
s
ha
r
e
ope
r
a
tor
s
,
a
nd
L
T
E
s
e
r
vice
pe
ne
tr
a
ti
on
.
T
o
c
a
lcula
te
the
f
o
r
e
c
a
s
ti
ng
n
umber
of
u
s
e
r
s
us
e
s
in
(
4
)
:
P
t
=P
o
(
1
+
r
)
n
(
4)
whe
r
e
P
t
is
r
e
s
idents
in
ye
a
r
t,
P
o
is
r
e
s
idents
in
the
ba
s
e
ye
a
r
,
r
is
population
gr
owth
,
a
nd
n
is
ba
s
e
ye
a
r
.
B
a
s
e
d
on
the
c
a
lcula
ti
on
in
T
a
ble
5
,
i
t
is
known
that
L
T
E
us
e
r
s
f
or
the
ne
xt
5
ye
a
r
s
a
r
e
19,
064
pe
oples
.
T
a
ble
5.
F
or
e
c
a
s
ti
ng
n
umber
R
e
s
id
e
nt
s
N
umbe
r
P
r
odu
c
ti
ve
A
ge
(
15
-
65)
P
opul
a
ti
on G
r
ow
th
F
a
c
to
r
M
a
r
ke
t
S
ha
r
e
O
pe
r
a
to
r
L
T
E
U
s
e
r
s
f
or
th
e
N
e
xt
5
ye
a
r
s
132
,
002
75%
1
.
7%
17
.
7%
19
,
064
T
he
s
e
r
vice
model
wa
s
us
e
d
to
f
ind
out
the
mi
nim
u
m
a
mount
of
th
r
oughput
s
o
that
the
c
us
tom
e
r
s
c
a
n
a
c
c
e
s
s
a
va
il
a
ble
s
e
r
vice
s
on
the
L
T
E
ne
twor
k.
T
a
ble
6
s
hows
the
pa
r
a
mete
r
s
of
s
e
r
vice
model
a
nd
obtaine
d
thr
oughput
va
lue
pe
r
s
e
s
s
ion
f
or
e
a
c
h
type
s
of
L
T
E
s
e
r
vice
s
f
r
om
the
upli
nk
o
r
downlink
di
r
e
c
ti
on.
On
e
a
c
h
s
ide
r
e
quir
e
s
da
ta:
be
a
r
e
r
r
a
te
(
kbps
)
,
s
e
s
s
ion
ti
me,
a
nd
s
e
s
s
ion
duty
r
a
ti
o.
T
he
ne
xt
s
tep
is
c
a
lcula
ti
ng
the
c
e
ll
a
ve
r
a
ge
thr
oughput
us
ing
i
n
(
5
)
a
nd
s
hown
in
T
a
bl
e
7
:
C
e
ll
A
ve
r
a
ge
T
hr
oughput
=
∑
[
S
I
NR
P
r
oba
bil
it
y
(
P
n
)
x
C
e
ll
C
a
pa
c
it
y]
(
5)
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
R
ange
e
x
pans
ion
me
thod
on
he
ter
oge
ne
ous
ne
tw
or
k
to
incr
e
as
e
picoc
e
ll
c
ov
e
r
age
(
Hadi
Supr
iadi
)
2347
T
a
ble
6.
T
he
pa
r
a
mete
r
of
s
e
r
vice
m
ode
l
T
r
a
f
f
ic
B
e
ha
vi
or
U
P
L
I
N
K
D
O
W
N
L
I
N
K
UL
DL
B
e
a
r
e
r
S
e
s
s
io
n
S
e
s
s
io
n
B
le
r
B
e
a
r
e
r
S
e
s
s
io
n
S
e
s
s
io
n
B
le
r
T
hr
oughput/
S
e
s
s
io
n
(
k
bps
)
T
hr
oughput/
S
e
s
s
io
n
(
k
bps
)
R
a
te
(
kbps
)
T
ime
(
s
)
D
ut
y
R
a
ti
o
R
a
te
(
kbps
)
T
im
e
(
s
)
D
ut
y
R
a
ti
o
V
O
I
P
26.
9
80
0.
4
0.
01
26
.
9
80
0.
4
0.
01
869.
495
869.
495
V
id
e
o P
hone
62.53
70
1
0.01
62.53
70
1
0.01
4,421.313
4,421.313
V
id
e
o
C
onf
e
r
e
nc
e
62.53
1800
1
0.01
62.53
1800
1
0.01
113,690.9
113,690.9
R
e
a
l
T
im
e
G
a
mi
ng
31.26
1800
0.2
0.01
125.06
1800
0.4
0.01
11,367.27
9,095.,73
S
tr
e
a
mi
ng
M
e
di
a
31.26
3600
0.05
0.01
250.11
3600
0.95
0.01
5,683.636
864,016.4
I
M
S
S
ig
na
ll
in
g
15.63
7
0.2
0.01
15.63
7
0.2
0.01
22.103
22.103
W
e
b
B
r
ow
s
in
g
62.53
1800
0.05
0.01
250.11
1800
0.05
0.01
5,684.545
22,737.27
F
il
e
T
r
a
ns
f
e
r
140.69
600
1
0.01
750.34
600
1
0.01
85,266.67
454,751.5
E
ma
il
140.69
50
1
0.01
750.34
15
1
0.01
7,105.556
11,368.79
P
2P
F
il
e
S
ha
r
in
g
250.11
1200
1
0.01
750.34
1200
1
0.01
30,3163.6
909,503.03
T
a
b
l
e
7.
C
e
ll
a
ve
r
a
ge
t
hr
oughpu
t
D
ir
e
c
ti
on
M
a
c
r
oc
e
ll
(
M
bps
)
P
ic
oc
e
ll
(
M
bps
)
UL
14
.
70813024
29
.
41629024
DL
12
.
25677024
24
.
51357024
Af
ter
obtaining
the
va
lue
of
s
ingl
e
-
us
e
r
thr
oughpu
t
a
nd
c
e
ll
thr
oughput,
the
ne
xt
s
tep
is
de
ter
mi
n
ing
the
number
o
f
r
e
qui
r
e
d
picoc
e
ll
s
.
T
he
c
a
lcula
ti
on
o
f
c
e
ll
s
number
wa
s
ba
s
e
d
on
the
e
s
ti
mate
d
maximum
number
of
us
e
r
s
that
c
ould
s
e
r
ve
d
by
one
picoc
e
ll
s
it
e
.
B
e
c
a
us
e
the
planning
a
r
e
a
is
a
n
ur
ba
n
a
r
e
a
,
the
pe
a
k
o
f
the
a
ve
r
a
ge
r
a
ti
o
us
e
d
wa
s
20%
with
a
DL
c
e
ll
loa
d
of
100
%
.
F
igur
e
s
5
a
nd
6
a
r
e
the
c
a
lcula
ti
on
o
f
maximum
us
e
r
f
or
e
xis
ti
ng
s
it
e
s
a
nd
p
ico
s
it
e
s
.
F
r
om
the
c
a
lcula
ti
on
r
e
s
ult
obtaine
d
the
maximum
of
us
e
r
numb
e
r
s
e
r
ve
d
by
the
e
xis
ti
ng
s
it
e
s
wa
s
840
us
e
r
s
pe
r
s
it
e
.
T
he
r
e
a
r
e
16
e
xis
ti
ng
s
it
e
s
in
C
oblong
S
ubdis
tr
ict
a
r
e
a
.
T
he
e
xis
ti
ng
s
it
e
s
e
na
ble
to
s
e
r
ve
a
s
many
a
s
13,
440
us
e
r
s
(
840
us
e
r
s
x
16
s
it
e
s
)
.
F
igur
e
5.
T
he
maximum
o
f
e
xis
ti
ng
s
it
e
us
e
r
T
he
maximum
number
of
us
e
r
s
that
c
ould
be
s
e
r
ve
d
by
one
picoc
e
ll
s
it
e
wa
s
393
us
e
r
s
with
70
%
c
e
ll
loading
to
a
nti
c
ipate
e
xc
e
s
s
ive
us
e
r
s
ur
ge
s
.
I
n
th
is
planning,
P
icoc
e
ll
wa
s
us
e
d
to
ove
r
c
ome
the
ove
r
load
c
a
pa
c
it
y
e
xpe
r
ienc
e
d
by
e
xis
ti
ng
s
it
e
s
.
T
he
c
a
lcula
ti
on
r
e
s
ult
o
f
f
or
e
c
a
s
ti
ng
number
of
us
e
r
s
in
C
oblong
s
ubdis
tr
ict
f
or
the
ne
xt
5
ye
a
r
s
a
s
many
a
s
19
,
064
us
e
r
s
.
T
he
e
xis
ti
ng
s
it
e
s
we
r
e
not
a
ble
to
s
e
r
ve
a
ll
us
e
r
s
in
the
r
e
gion
be
c
a
us
e
the
maximum
number
of
us
e
r
s
that
c
ould
be
s
e
r
ve
d
by
e
xis
ti
ng
s
it
e
s
w
a
s
13,
4
40
us
e
r
s
,
whic
h
mea
ns
a
tot
a
l
of
5,
624
us
e
r
s
did
not
ge
t
s
e
r
v
ice
s
f
r
om
e
xis
ti
ng
s
it
e
s
.
B
a
s
e
d
on
the
maximum
nu
mber
of
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
5
,
Oc
tober
2020:
2341
-
2351
2348
us
e
r
s
in
P
icoc
e
ll
,
it
took
15
P
icoc
e
ll
s
it
e
s
to
s
e
r
ve
a
ll
us
e
r
s
who
e
xpe
r
ienc
e
d
o
f
f
loa
d
c
a
pa
c
it
y.
I
n
he
ter
oge
ne
ous
ne
twor
k
plan
ning,
it
is
ne
c
e
s
s
a
r
y
to
c
onf
igur
e
picoc
e
ll
s
;
f
or
e
xa
mpl
e
,
by
a
dding
ha
ndove
r
mar
g
ins
a
nd
indi
vidual
c
e
ll
of
f
s
e
ts
that
f
unc
ti
on
a
s
r
a
nge
e
xpa
ns
ion
,
a
nd
a
dding
c
on
f
igur
a
ti
on
of
Almos
t
B
lank
S
ubf
r
a
me
(
AB
S
)
pa
tt
e
r
ns
on
mac
r
o
s
it
e
s
that
c
a
n
r
e
duc
e
the
im
pa
c
t
o
f
int
e
r
f
e
r
e
nc
e
be
twe
e
n
mac
r
oc
e
ll
s
a
nd
p
icoc
e
ll
s
.
F
oll
owing
is
the
c
on
f
igur
a
ti
on
of
the
AB
S
pa
tt
e
r
n
us
e
d
by
the
mac
r
o
s
it
e
:
AB
S
pa
tt
e
r
n
1:
1000000010000000100000001000
000010000000
AB
S
pa
tt
e
r
n
2:
0100000001000000010000000100
000001000000
AB
S
pa
tt
e
r
n
3:
00
10000000100000001000000010
000000100000
AB
S
pa
tt
e
r
n
4:
0001000000010000000100000001
000000010000
AB
S
pa
tt
e
r
n
5:
0000100000001000000010000000
100000001000
AB
S
pa
tt
e
r
n
6:
0000010000000100000001000000
010000000100
AB
S
pa
tt
e
r
n
7:
0000001000000010000000100000
00
1000000010
AB
S
pa
tt
e
r
n
8:
0000000100000001000000010000
000100000001
1
=
Almos
t
B
lank
S
ub
f
r
a
me
0
=
Ac
ti
ve
S
ub
f
r
a
me
F
igur
e
6.
T
he
maximum
o
f
picoc
e
ll
s
us
e
r
s
3.
S
I
M
UL
AT
I
ON
AN
D
RE
S
UL
T
S
S
c
e
na
r
io
1
s
im
ulate
d
the
e
xis
ti
ng
s
it
e
s
to
f
ind
ou
t
the
pe
r
f
or
manc
e
be
f
o
r
e
he
ter
oge
ne
ous
ne
twor
k
planning
wa
s
c
onduc
ted,
whic
h
then
the
r
e
s
ult
s
we
r
e
c
ompar
e
d
to
be
twe
e
n
be
f
o
r
e
a
nd
a
f
ter
picoc
e
ll
p
lanning.
I
n
the
C
oblong
s
ubdis
tr
ict
a
r
e
a
,
ther
e
we
r
e
16
e
xis
ti
ng
s
it
e
s
.
M
e
a
nwhile,
s
c
e
na
r
io
2
s
im
ul
a
ted
the
e
xis
t
ing
s
it
e
s
by
a
dding
picoc
e
ll
s
s
it
e
f
r
om
the
planning
r
e
s
ult
s
.
T
he
picoc
e
ll
plac
e
ment
us
e
d
in
thi
s
s
im
ulation
w
a
s
ba
s
e
d
on
the
poor
c
ondit
ion
of
R
S
R
P
a
nd
S
I
NR
a
nd
s
e
ve
r
a
l
point
s
of
ove
r
load
c
a
pa
c
it
y.
I
n
thi
s
s
im
ulation,
th
e
r
e
we
r
e
16
mac
r
oc
e
ll
s
a
nd
15
picoc
e
ll
s
.
F
igur
e
s
7
a
nd
8
a
r
e
the
dis
tr
ibut
ion
of
picoc
e
ll
s
it
e
s
in
the
C
oblong
a
r
e
a
.
F
igur
e
9
s
hows
the
c
ompar
is
on
r
e
s
ult
of
two
s
c
e
na
r
ios
,
na
mely
the
R
S
R
P
a
nd
S
I
NR
pa
r
a
mete
r
s
.
R
S
R
P
va
lues
obtaine
d
f
r
om
the
two
s
im
ulations
indi
c
a
t
e
d
that
the
R
S
R
P
a
r
e
a
c
ove
r
a
ge
met
the
s
tanda
r
d
of
KPI
ope
r
a
tor
that
wa
s
with
R
S
R
P
c
ove
r
a
ge
≥
-
90
dB
m
,
a
bove
90
%
.
Af
ter
picoc
e
ll
s
planning,
the
R
S
R
P
va
lue
incr
e
a
s
e
d
by
5%
,
whic
h
it
s
howe
d
that
picoc
e
ll
s
planning
incr
e
a
s
e
d
the
R
S
R
P
va
lue
in
ter
ms
of
c
ove
r
a
ge
.
W
hil
e
ba
s
e
d
on
the
S
I
NR
pa
r
a
mete
r
s
,
both
s
c
e
na
r
i
os
did
not
ye
t
mee
t
the
s
tanda
r
d
of
KPI
ope
r
a
tor
be
c
a
us
e
the
S
I
NR
c
ove
r
a
ge
va
lue
≥
5
dB
wa
s
s
ti
ll
be
low
90
%
.
T
he
va
lue
o
f
S
I
NR
≥
5
dB
in
s
im
ulation
1
wa
s
59.
55%
a
nd
in
s
im
ulation
2
wa
s
70.
99
%
.
I
t
c
a
n
be
r
e
s
ult
e
d
f
r
o
m
the
number
of
obs
tac
les
,
s
uc
h
a
s
lar
ge
tr
e
e
s
,
tall
buil
dings
,
a
nd
une
ve
n
e
a
r
th
c
ontour
s
s
o
that
the
qua
li
ty
of
the
s
ignal
powe
r
obtaine
d
wa
s
dis
tur
be
d
.
B
a
s
e
d
on
the
s
tanda
r
d
of
KPI
ope
r
a
tor
,
the
downlin
k
of
thr
oughput
va
lue
wa
s
12
M
bps
a
nd
the
upli
nk
wa
s
6
M
bps
.
F
igur
e
10
s
hows
that
both
s
c
e
na
r
ios
met
the
s
tanda
r
d
of
KPI
ope
r
a
tor
.
T
he
mos
t
s
i
gni
f
ica
nt
incr
e
a
s
e
in
downlink
thr
oughput
oc
c
ur
r
e
d
in
s
im
ul
a
ti
on
2.
I
t
r
e
s
ult
e
d
f
r
om
the
good
S
I
NR
va
lue
that
a
f
f
e
c
ted
the
type
of
modul
a
ti
on
us
e
d,
whe
r
e
the
S
I
NR
va
lue
will
be
dir
e
c
tl
y
pr
opor
ti
ona
l
to
the
th
r
ough
put
va
lue
be
c
a
us
e
the
S
I
NR
va
lue
wi
ll
r
e
pr
e
s
e
n
t
the
qua
li
t
y
of
the
s
ignal
r
e
c
e
ived
by
the
us
e
r
.
F
igur
e
11
s
hows
the
c
ompar
is
on
of
us
e
r
s
c
onne
c
ted
a
nd
r
e
jec
ted.
I
n
s
im
ulation
1,
it
s
howe
d
that
the
KPI
s
tanda
r
d
wa
s
not
a
c
hieve
d
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
R
ange
e
x
pans
ion
me
thod
on
he
ter
oge
ne
ous
ne
tw
or
k
to
incr
e
as
e
picoc
e
ll
c
ov
e
r
age
(
Hadi
Supr
iadi
)
2349
be
c
a
us
e
the
va
lue
of
r
e
jec
ted
us
e
r
wa
s
s
ti
ll
a
bove
1%
.
I
t
r
e
s
ult
e
d
f
r
om
the
L
T
E
ne
twor
ks
that
we
r
e
no
t
a
ble
to
c
ope
with
of
f
load
c
a
pa
c
it
y
due
to
the
us
e
r
s
ur
ge
s
.
M
e
a
nwhile,
in
s
im
ulation
2
a
f
ter
he
ter
oge
ne
ous
ne
twor
k
planning
with
picoc
e
ll
s
,
it
s
howe
d
that
the
s
c
e
na
r
i
o
wa
s
a
ble
to
s
e
r
ve
a
nd
ove
r
c
ome
the
us
e
r
s
ur
ge
s
.
B
e
c
a
us
e
picoc
e
ll
s
we
r
e
a
ble
to
e
a
s
e
the
tr
a
f
f
ic
bur
de
n
on
mac
r
o
s
it
e
s
,
whe
r
e
whe
n
mac
r
o
s
it
e
s
will
e
xpe
r
ienc
e
a
us
e
r
s
ur
ge
,
then
the
tr
a
f
f
ic
load
wil
l
im
media
tely
be
ha
n
dled
by
picoc
e
ll
s
.
F
igur
e
7.
T
he
dis
tr
ibut
ion
of
e
xis
ti
ng
s
it
e
s
(
S
c
e
na
r
i
o
1)
F
igur
e
8.
T
he
dis
tr
ibut
ion
of
picoc
e
ll
s
(
s
c
e
na
r
io
2)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
5
,
Oc
tober
2020:
2341
-
2351
2350
F
igur
e
9.
T
he
c
ompar
is
on
of
R
S
R
P
a
nd
S
I
NR
F
igur
e
10.
T
he
c
ompar
is
on
of
t
hr
oughput
va
lue
F
igur
e
11.
T
he
C
ompar
is
on
o
f
us
e
r
c
onne
c
ted
a
nd
us
e
r
r
e
jec
ted
4.
CONC
L
USI
ON
B
a
s
e
d
on
R
S
R
P
pa
r
a
mete
r
s
,
in
s
im
ulation
1,
the
a
ve
r
a
ge
R
S
R
P
va
lue
wa
s
-
69.
23
dB
m,
while
the
R
S
R
P
c
ove
r
a
ge
≥
-
90
dB
m
wa
s
93.
35%
.
W
h
e
r
e
a
s
in
s
im
ulation
2
,
the
R
S
R
P
a
ve
r
a
ge
va
lue
w
a
s
-
64.
96
dB
m,
while
the
R
S
R
P
c
ove
r
a
ge
wa
s
≥
-
90
dB
m
w
a
s
97.
72%
.
B
a
s
e
d
on
the
S
I
NR
pa
r
a
mete
r
s
,
in
s
im
ulation
1,
a
n
a
ve
r
a
ge
S
I
NR
va
lue
of
7
.
57
dB
wa
s
obtaine
d
,
while
the
S
I
NR
≥
5
dB
c
ove
r
a
ge
wa
s
59
.
55%
.
W
he
r
e
a
s
in
s
im
ulation
2,
a
n
a
ve
r
a
ge
S
I
NR
o
f
10
.
77
dB
wa
s
obtaine
d,
while
the
S
I
NR
≥
5
dB
c
ove
r
a
ge
wa
s
70.
99%
.
An
incr
e
a
s
e
in
the
a
ve
r
a
ge
va
lue
of
S
I
NR
in
s
im
u
lation
2
,
thi
s
will
a
f
f
e
c
t
the
va
lue
of
th
r
oughput.
B
a
s
e
d
on
the
pa
r
a
mete
r
s
of
th
r
oughput,
in
s
im
ulation
1,
the
a
ve
r
a
ge
va
lue
of
downlink
thr
oughput
is
17.
04
M
bps
a
nd
the
a
ve
r
a
ge
va
lue
of
upli
nk
thr
oughput
is
14
.
07
M
bps
.
I
n
s
im
ulation
2,
the
a
ve
r
a
ge
va
lue
of
d
ownlink
thr
oughput
is
21.
37
M
bps
a
nd
the
a
ve
r
a
ge
va
lue
of
upli
nk
thr
oughput
is
17
.
80
M
bps
.
B
a
s
e
d
on
us
e
r
c
onne
c
ted
pa
r
a
mete
r
s
,
s
im
ulation
r
e
s
ult
s
1
,
the
numbe
r
o
f
c
onne
c
ted
us
e
r
s
is
19
,
238
(
77.
2
%
)
us
e
r
s
a
nd
us
e
r
r
e
jec
ted
is
4,
391
(
22
.
8%
)
.
W
hil
e
the
r
e
s
ult
s
of
s
im
ulation
2
,
the
number
of
c
onne
c
ted
us
e
r
s
incr
e
a
s
e
d
s
igni
f
ica
ntl
y
to
a
s
many
a
s
19,
078
us
e
r
s
(
99
.
2%
)
a
nd
r
e
jec
ted
us
e
r
s
a
s
many
a
s
163
us
e
r
s
(
0.
8
%
)
.
B
oth
planning
s
c
e
na
r
ios
s
howe
d
that
he
ter
oge
ne
ous
ne
twor
k
planning
wa
s
highl
y
e
f
f
e
c
ti
ve
in
ove
r
c
omi
ng
c
ove
r
a
ge
a
nd
c
a
pa
c
it
y
is
s
ue
s
.
I
n
ter
ms
of
c
ove
r
a
ge
,
it
c
a
n
p
r
ovide
a
ve
r
y
good
R
S
R
P
a
nd
S
I
NR
va
lue.
M
e
a
nwhile,
in
ter
ms
o
f
he
ter
oge
ne
ous
ne
t
wor
k
c
a
pa
c
it
y,
it
c
a
n
incr
e
a
s
e
the
thr
oughput
a
nd
ove
r
c
ome
the
us
e
r
s
ur
ge
s
.
T
he
r
e
f
or
e
,
he
ter
oge
ne
ous
ne
twor
ks
s
uit
s
to
be
a
ppli
e
d
in
de
ns
e
ur
ba
n
a
r
e
a
s
t
ha
t
ha
ve
many
us
e
r
s
a
nd
in
a
r
e
a
s
that
ha
ve
many
obs
tac
les
.
T
h
is
a
r
ti
c
le
is
one
o
f
the
r
e
c
omm
e
nda
ti
ons
f
o
r
c
e
ll
ular
ne
twor
k
pr
ovider
s
in
opti
mi
z
ing
the
c
ove
r
a
ge
o
f
he
ter
oge
ne
ous
ne
twor
k
s
a
nd
a
r
e
f
e
r
e
nc
e
f
o
r
r
e
s
e
a
r
c
he
r
s
in
he
ter
oge
ne
ous
ne
twor
k
r
e
s
e
a
r
c
h.
RE
F
E
RE
NC
E
S
[1
]
T
.
Q
.
S.
Q
.
Cai
re,
G
i
u
s
ep
p
e,
J
effrey
G
.
A
n
d
re
w
s
,
“Sp
ec
i
al
i
s
s
u
e
o
n
h
et
er
o
g
e
n
eo
u
s
n
et
w
o
r
k
s
,
”
J.
Co
m
m
u
n
.
Net
w
o
r
ks
,
v
o
l
.
1
3
,
n
o
.
4
,
2
0
1
1
.
[2
]
Y.
-
C.
W
.
an
d
S.
L
.
W
an
g
,
“Smal
l
-
cel
l
Pl
a
n
n
i
n
g
Imp
ro
v
e
i
n
L
T
E
H
et
N
et
t
o
E
ff
i
ci
e
n
cy
E
n
erg
y
,
”
In
t
.
J.
Co
m
m
u
n
.
S
ys
t
.
,
v
o
l
.
3
1
,
n
o
.
5
,
p
.
e
3
4
9
2
,
2
0
1
7
.
[3
]
U
.
K
.
U
.
A
j
i
Mau
l
an
a,
A
rfi
an
t
o
Fah
m
i
,
“D
es
i
g
n
o
f
L
T
E
-
A
d
v
a
n
ced
H
e
t
ero
g
en
e
o
u
s
N
et
w
o
r
k
s
w
i
t
h
Pi
co
Cel
l
U
s
i
n
g
Ran
g
e
E
x
p
a
n
s
i
o
n
i
n
Ci
mah
i
Ci
t
y
,
”
P
r
o
cee
d
i
n
g
o
f
E
n
g
i
n
eer
i
n
g
,
2
0
1
8
.
[4
]
U
.
K
.
U
.
H
.
Fi
n
an
d
ri
y
an
t
o
,
A
.
Fah
mi
,
“A
n
al
y
s
i
s
o
f
Smal
l
Cel
l
Perfo
rma
n
ce
i
n
L
T
E
-
A
d
v
an
ce
d
H
et
er
o
g
e
n
eo
u
s
N
et
w
o
r
k
s
,
”
e
-
P
r
o
cee
d
i
n
g
o
f
E
n
g
i
n
ee
r
i
n
g
,
2
0
1
6
.
Evaluation Warning : The document was created with Spire.PDF for Python.