TELKOM
NIKA
, Vol.14, No
.1, March 2
0
1
6
, pp. 119~1
2
8
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v14i1.2668
119
Re
cei
v
ed Se
ptem
ber 10, 2015; Revi
se
d Jan
uary 7, 2016; Accept
ed Ja
nua
ry 2
0
, 2016
Carrier Aggregation Technique to Improve Capacity in
LTE-Advanced Network
Iskandar*, R. Galih
Schoo
l of Elect
r
ical En
gin
eeri
ng an
d Info
rma
tics, Bandun
g Institute of T
e
chno
log
y
,
Jala
n Ganes
ha
No. 10, B
and
u
ng 40
13
2, INDONESIA
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: iskand
ar@st
e
i.itb.ac.id
A
b
st
r
a
ct
Carrier
aggr
eg
ation (CA) is
a major fe
atur
e in
LT
E-a
d
va
nced tec
hno
lo
gy that all
o
w
s
netw
o
rk
provi
der to
use
mor
e
tha
n
o
n
e
carrier s
i
multa
neo
usly to
incr
ease c
a
p
a
city. CA uses tw
o o
r
mor
e
i
ndiv
i
d
u
a
l
compo
nent car
r
iers (CCs) of
the same (int
ra-ba
nd)
a
nd
different (int
er-ban
d) frequ
en
cy bands. In thi
s
pap
er, w
e
ev
al
uate
CA tech
ni
ques
w
i
th the
test par
a
m
eter
usin
g BLE
R
a
nd thr
oug
hp
ut
insid
e
th
e test
e
d
area
of LT
E n
e
tw
ork. T
he tested are
a
is d
e
sig
ned to
r
e
p
r
esent
most co
mmo
n
e
n
viro
n
m
e
n
t in w
h
ic
h
LT
E
w
ill be
d
epl
oye
d
. We use
3
0
MH
z
fr
equ
ency
ba
ndw
idth
in t
he si
mul
a
tion
to ev
alu
a
te i
n
tra-ba
nd
an
d i
n
ter
-
ban
d CA tech
niq
ue perfor
m
ance. T
w
o CCs is allocate
d
for primary ser
v
ing cel
l
(PCel
l
) and sec
ond
ary
servin
g cell (S
Cell). Si
mu
lati
on result show
s for the
dow
nlink, BLER ten
d
to increase,
as the freque
n
c
y
config
uratio
n i
s
incre
a
se. Ho
w
e
ver, for the upli
n
k BLE
R
is
z
e
r
o
. W
e
foun
d that for CA c
onfig
uratio
n w
i
t
h
w
i
der PCell
ba
ndw
idth resu
lti
ng better thro
u
ghp
ut. In
CA intra-ban
d, the throu
ghp
ut of non-co
ntigu
ous
CA
is better than CA contig
uous.
T
hen in the int
e
r-ba
nd CA
, CC combin
atio
n w
i
th a low
e
r fre
que
ncy prod
uc
e
s
hig
her thro
ugh
put.
Ke
y
w
ords
:
Lo
ng T
e
rm Ev
olut
ion (LT
E
) Adva
nced, carri
er a
ggre
gatio
n, BL
ER, through
put
Copy
right
©
2016 Un
ive
r
sita
s Ah
mad
Dah
l
an
. All rig
h
t
s r
ese
rved
.
1. Introduc
tion
Long Te
rm Evolution (LTE
) Rele
ase 8 has b
een the
primary bro
adba
nd tech
nologi
e
s
whi
c
h attra
c
ti
ng mu
ch the
attention in cellular
com
m
unication [1]. With current
pea
k data rates
achi
eved i
n
LTE Relea
s
e
8, research
es
of furth
e
r enh
an
ceme
nts h
a
ve b
e
en
carrie
d o
u
t to
increa
se
data
rate
s in
LTE
Rele
ase 8. I
t
is ta
rgeted
i
n
IMT-A
req
u
i
reme
nts th
at the pe
ak dat
a
rate ca
n be 1 Gbp
s
for uplink an
d 50
0 Mbps for d
o
wnli
nk. The
s
e enh
an
ce
ments have
bee
n
con
s
id
ere
d
in
3GPP as a part of LTE-a
d
vanced (L
T
E
-A) and al
so kno
w
n a
s
LTE Rele
ase
10
specifications [2-4].
LTE-A re
quirements
have
been ad
dre
s
sed by the
3
rd
Gene
ratio
n
Partne
rship
Project
(3GPP)
and
the sta
nda
rdi
z
ation bo
dy be
hind L
T
E as
the pa
rt of its
Rele
ase 10. I
n
LTE-A, the
r
e
are
set of feat
ure
s
.
Carrie
r
aggregatio
n
(CA) i
s
one
of
the mo
st d
e
m
ande
d featu
r
es i
n
L
T
E-A [
5
]-
[6]. It has ma
ximum five compon
ent ca
rrie
r
s
(
CCs)
of 20 MHz b
and
width that
can be a
s
sig
ned
to the user e
quipme
n
t (UE) to achieve
a tota
l transmissi
on ba
nd
width of up to 100 MHz. CA is
then
calle
d a
s
a
way to
co
mbine
multipl
e
comp
onent
carrie
rs (CCs) to
have
scal
able
ban
dwi
d
th
expan
sion. T
h
is
carrie
r co
mbination i
s
not only
co
nfigure
d
in the
same
(contig
uou
s) freque
ncy
but also in
different (n
o
n
-contigu
o
u
s
) frequen
cy
b
and
s. Gene
rally, this scheme is a
b
l
e
to
provide m
a
ximum flexibility in utilizing
the scarcity of radio resources.
Witne
ssi
ng the patterns of broad
ba
nd pen
etrati
on incre
a
ses every year, LTE-A
techn
o
logy is promisi
ng to
fulfillment of high-
spe
ed data. However, before su
ch technolo
g
y
is
impleme
n
ted,
it is n
e
cessary to
study
the net
work
perfo
rman
ce.
The
proble
m
is th
at ho
w to
determi
ne th
e type of
CA
sele
cted
by the regul
ator
assign
ed to
e
a
ch
ope
rato
r
is not
an
swered.
We have le
ss kno
w
led
ge o
f
which o
ne p
e
rform
be
tter
than others, contiguo
us o
r
non-co
ntiguo
us
intra or i
n
ter-band. Investi
gation to this probl
em is
v
e
ry ra
re. In [6] we foun
d that the do
wnl
i
nk
perfo
rman
ce
of
CA
i
s
studi
ed
a
nd discu
s
sed. Ho
wev
e
r
,
th
e s
t
ud
y is
limited
to
in
te
r
-
b
a
n
d
C
A
on
ly
with different
traffic load. T
herefo
r
e, to make
it cle
a
r
and comp
reh
ensive di
scu
s
sion, we pro
p
o
se
intra a
nd inte
r-b
and
CA te
chni
que
with
different traffic lo
ad a
nd
d
e
sig
nated
co
verage
area i
n
whi
c
h LTE-A
is deployed.
We pro
p
o
s
e
an evaluat
io
n of CA tech
nique sele
ction in a desi
g
ned
LTE-A netwo
rk in te
rms
of BLER and th
roug
hput in o
r
de
r to kn
ow
whi
c
h on
e th
e best te
chni
que
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 14, No. 1, March 2
016 : 119 – 1
2
8
120
of the CA. In
addition to the freque
ncy combinatio
n, it will also be o
b
se
rved thro
u
ghput ba
sed
on
combi
nation
of primary an
d se
cond
ary comp
one
nt carri
er ba
nd
wi
dth. This stud
y is based o
n
the
acce
ss net
wo
rk
plan
ning
case
stu
d
y in t
y
pical
urb
an environ
ment by
usi
ng com
puter sim
u
lati
on.
LTE radio n
e
t
work plan
nin
g
with differe
nt propa
gatio
n model an
d the techni
que
to improve its
quality u
s
ing
joint dete
c
tio
n
techniq
ue f
o
r the
next
wirele
ss comm
unication te
chnolo
g
ies ha
ve
been inve
stig
ated in [7]-[8]
.
In this contri
bution, we
be
gin with the d
e
sig
n
ing L
T
E-A cove
rag
e
in
particula
r area and th
en
simulate th
e CA tec
hni
que
s in that
area to fin
d
the be
st CA
perfo
rman
ce
so that the re
sult wo
uld be
use
d
by the operato
r
in opti
m
izing thei
r spectrum.
Figure 1. LTE-A ca
rrie
r
agg
regatio
n tech
nique
The
re
st of t
he p
ape
r i
s
a
s
follo
w. Se
ct
ion 2
p
r
ovide
s
a
n
ove
r
vie
w
of
differe
nt types of
CA. Sectio
n
3
outline
s
an
L
T
E-advan
ce
d
network
plan
ning
simul
a
tio
n
with
the
ma
in focus on
th
e
con
s
id
eratio
n
s
in u
r
ba
n en
vironme
n
t as
the simul
a
ted
area.
Re
sult
s an
d an
alysi
s
are p
r
e
s
ent
ed
in Section 4. Finally, some
con
c
lu
sion
s
are d
r
a
w
n in
Section 5.
2. Carrier
Aggr
egation
T
e
c
hniques
Today in
mo
st of the count
ry, there
are
almo
st non
e of
a cellula
r service
p
r
ovid
er
o
w
n
s
contin
uou
s spectrum of 100 MHz.
This be
cau
s
e of
the spect
r
u
m
sca
rcity and the numb
e
r of
servi
c
e
provider rel
a
tively many
in
a
co
untry. The
r
ef
ore, it
eme
r
g
e
s th
e
ne
w
method
calle
d CA
techni
que
[9-10]. The
r
e
are thre
e diffe
rent mod
e
s of
CA, na
mely
intra-ban
d
co
ntiguou
s, intra-
band
no
n-co
ntiguou
s, a
n
d
inter-ba
nd
CA [14-1
8
]. 3G
PP’s radio
a
c
ce
ss net
work (RAN)
worki
ng
grou
p 4
(RA
N
4
)
i
s
re
sp
o
n
sibl
e fo
r
set
t
ing pe
rforma
nce
requi
rem
ents
of
CA
t
e
ch
niqu
e whi
c
h
initially limits
the agg
re
gation up t
o
two
com
pon
ent
carrie
rs only.
With thi
s
me
thod, one
UE
is
able to co
nsu
m
e up to 100
MHz
spe
c
tru
m
band
width
in total.
In CA te
ch
ni
que
ea
ch
ca
rrie
r
or chan
nel, which i
s
agg
re
gated,
is called
co
mpone
nt
carrie
rs (CCs). Each
CCs might
have
band
width
of
1.4, 3,
5, 1
0
o
r
2
0
M
H
z. The
maxim
u
m
numbe
r of
ch
annel
s that
can be
agg
reg
a
ted is five
, so the maximu
m amou
nt of band
width th
at
can
be
used
is 1
00 M
H
z.
The n
u
mb
er
of agg
reg
a
te
d carriers
ca
n differe
nt be
tween
do
wnli
nk
and u
p
lin
k, b
u
t the num
be
r of upli
n
k
CCs
never
mo
re tha
n
CCs
downlin
ks [6]
.
Each
CCs
can
also
have dif
f
erent ba
nd
width. The CA
techni
que
i
s
a way of h
o
w
to ag
gre
g
a
t
e some
of the
CCs. As
sho
w
n in Fi
gure
1, three te
chn
i
que
s or
type
s of CA o
n
LT
E-A are
de
scribed a
nd can
be
explain a
s
follows:
a)
Intra-b
and
co
ntiguou
s CA:
the method u
s
i
ng two or m
o
re
chan
nel
s in one freq
ue
ncy
band seq
uent
ially.
b)
Intra-b
and
no
n-contigu
o
u
s
CA: the meth
od to a
g
g
r
eg
ating
chan
nel
s
within a
si
n
g
le
freque
ncy ba
nd, but there i
s
a dista
n
ce betwe
en CC’
s
.
c)
Inter-b
and n
on-contig
uou
s CA: the me
thod to aggregatin
g different ban
ds of
freque
ncy ch
annel
s, it is useful if there i
s
fragm
entati
on ban
d.
In the
CA
sy
stem, the
r
e
a
r
e t
w
o type
s
of se
r
v
in
g ce
ll, th
o
s
e ar
e
Pr
ima
r
y Se
r
v
in
g
C
e
ll
(PCell
) an
d Secon
d
a
r
y Serving
Cell (SCell).
While
PCell have
the task to
bring P
r
ima
r
y
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Carrie
r Aggre
gation Te
chni
que to Im
prove Cap
a
cit
y
in LTE-Ad
van
c
e
d
Network (Iskan
dar)
121
Comp
one
nt
Carrie
r
(PCC) d
o
wnlink a
n
d
upli
n
k
an
d
handl
e the
Radio
Re
so
urce Control (RRC)
con
n
e
c
tion, SCell have the
task to b
r
ing t
he Se
co
nda
ry Compon
ent
Carrier
(SCC). There is onl
y
one P
C
C in t
he
CA sy
ste
m
, but the
SCC could
be
more
than
on
e. Cove
rag
e
of PCell
and
SCell
may vary dep
endin
g
on the
frequen
cy it is bein
g
used.
In this pa
per,
we a
s
sume
that there a
r
e
several LTE-
A network
sc
enar
ios
w
i
th FDD
duplexing
scheme
s
. Two
CCs a
nd t
he am
ount
of agg
regate
d
ban
dwi
d
th
on ea
ch
CA
config
uratio
n
is 30 M
H
z. We u
s
e freq
uen
cy com
b
i
nation that fo
llows some
CA co
nfigu
r
a
t
ions
determi
ned
by 3GPP release 10, 11, and 12,
i.e. 3GPP E-UT
RA Ca
rri
er Aggregati
o
n
(TS.36.101
). They are
sum
m
ari
z
ed in Ta
ble 1, Table 2
,
and Table 3,
resp
ectively [7].
Table 1. CA Intra-Ba
nd Co
ntiguou
s Con
f
iguration S
c
e
nario
s
Name
CA Configuration
E-UTRA Ba
nd
Band Name
CC Band
w
i
dth
(
M
Hz)
PCell SCell
A1 3C
3
1800
+
15
15
A2 3C
3
1800
+
10
20
A3 3C
3
1800
+
20
10
Table 2. CA Intra-Ba
nd No
n-Contigu
o
u
s
Configu
r
atio
n Scena
rio
s
.
Name
CA Configuration
E-UTRA Ba
nd
Band Name
CC Band
w
i
dth
(
M
Hz)
Space
(MHz
)
PCell SCell
B1
3A - 3A
3
1800 +
15
15
15
B2
3A - 3A
3
1800 +
10
20
20
B3
3A - 3A
3
1800 +
20
10
20
Table 3. CA Inter-Ba
nd No
n-Contigu
o
u
s
Configu
r
atio
n Scena
rio
s
.
Name
CA Configuration
E-UTRA Ba
nd
Band Name
Cell T
y
pe
CC Band
w
i
dth
(
M
Hz)
C1
3A - 28A
3 1800
+
SCell
15
28 700
APT
PCell
15
C2
3A - 28A
3 1800
+
Scell
20
28 700
APT
Pcell
10
C3
3A - 28A
3 1800
+
Scell
10
28 700
APT
Pcell
20
C4
1A - 8A
1 2100
Scell
20
8 900
Pcell
10
C5
7A - 28A
7 2600
Scell
15
28 700
APT
Pcell
15
C6
3A - 7A
3 1800
+
Pcell
15
7 2600
Scell
15
3. LTE-A Netw
o
r
k Planni
ng Simulation
Radi
o net
work pla
nnin
g
is an initial p
r
o
c
e
ss i
n
ou
r
work. Th
e p
r
oce
s
s be
gin
s
with the
coverage
di
mensi
onin
g
i
n
the
pa
rticul
ar
envir
o
n
me
nt,
as
a test
ed urb
an are
a
to evaluate CA
perfo
rman
ce.
The digital
map of that
area i
s
u
p
loa
ded to the n
e
twork pl
anni
ng sim
u
lato
r an
d
then all pa
ra
meters are consi
dered a
n
d
take
n into
con
s
id
eratio
n
in the simul
a
tion. Radi
o l
i
nk
cal
c
ulatio
n is a main task
in the cove
ra
ge dime
nsi
o
n
i
ng to estimat
e
the value o
f
the maximum
allowable pat
h
loss (MAPL
)
an
d
ra
dio si
gnal received
power
(RS
R
P) thre
sh
old.
In this pa
pe
r
we
use multipl
e
freque
ncy co
nfiguratio
ns,
therefo
r
e several differe
nt paramet
ers are u
s
ed in the
cal
c
ulatio
n. All the pa
ramet
e
rs u
s
ed i
n
th
e ra
dio lin
k b
udget
simulat
i
on a
r
e
summ
arized in
Tabl
e
4 an
d T
able
5. In the
first
step,
we
cal
c
ulate lin
k
bud
get u
s
ing
si
m
p
le fo
rmula
to
find the
nu
m
ber
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 14, No. 1, March 2
016 : 119 – 1
2
8
122
of site
re
quire
d to
cove
r the
tested
simul
a
ted a
r
e
a
. Receiver
sensitivity is
calcul
a
t
ed by u
s
in
g t
he
followin
g
equ
ation.
SINR
NF
N
x
x
S
rb
R
12
000
15
10
174
.
log
(1)
Table 4. Tx-Rx Parameters Use
d
in Cov
e
rag
e
Dim
e
n
s
ioni
ng
Par
a
meter
s
Dow
n
link
Uplink
T
r
a
n
sm
it
te
r En
d (
e
N
B
)
Max TX Po
wer
(
d
Bm)
43
23
Number of
T
x
An
tenna
2
1
T
x
Diver
s
ity
G
a
in (
d
B)
3
0
Feeder Loss
per length
(dB/m)
0.06
-
Feeder Length (
m
)
35
-
Feeder
Loss/Body Loss (dB)
2.1
1
Connector Loss (
d
B)
0.5
-
Recei
v
e
r En
d (
U
E)
Noise Figureur
e
(dB)
8
4
Required SIN
R
a
t
Cell Edge (dB)
-4.13
-5.11
Fast Fade Ma
rgi
n
(dB)
4.5
4.5
Interfer
ence Mar
g
in (dB)
5.5
4.5
Bod
y
Loss/Fee
d
e
r Loss (dB)
1
2.1
Additional Gain (
d
B)
0
3
Table 5. Prop
agation Pa
ra
meters Used
in Coverage
Dimen
s
io
ning
Parameters
T
y
pe of
Area
Urban
Suburban
eNB antenna
height (m)
30
30
UE antenna h
e
ight (m)
1.5
1.5
Cell Edge Proba
bility
75%
75%
Outdoo
r Standa
r
d
Deviation Shadow
ing
7
7
UE antenna cor
r
e
ction factor (dB)
0 0.01
Cl
utter c
o
rrec
t
i
o
n
(dB)
0
-9.3
Ar
ea size (
k
m
2
)
104.84
62.83
Table 6. Nu
mber of e
N
B based on
cov
e
rag
e
dime
nsioning..
Parameter
Formulation
Area T
y
pe
Urban
Suburban
Number
sector
per
site
3 3
Cell/eNB coverage area [km
2
]
L
si
t
e
= 1.95 *
(d
2
)
d: cell diameter
2.96 9.99
Simulated coverage area [km
2
]
L
r
104.84
62.83
Simulated number of eNB p
e
r a
r
e
a
t
y
pe
N = L
r
/ sector nu
mber
36 7
Total simulated number of e
N
B
43
Whe
r
e
N
rb
denote
s
numb
e
r of Re
so
urce Blo
ck,
NF
is a noise F
i
gure
u
re of the re
ceive
r
, and
SINR
de
note
s
si
gnal to i
n
teferen
c
e
pl
us n
o
ise
rati
o. From all t
hose pa
rame
ters, MAPL
and
RSRP thre
sh
old value can
be cal
c
ulate
d
by using th
e followin
g
eq
uation
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Carrie
r Aggre
gation Te
chni
que to Im
prove Cap
a
cit
y
in LTE-Ad
van
c
e
d
Network (Iskan
dar)
123
R
S
R
T
T
S
L
G
G
P
MAPL
(2)
shadowing
MAPL
EIRP
RSPP
subscriber
per
min
(3)
Whe
r
e P
T
d
e
notes e
N
B po
wer tran
smit, G
T
is tran
smit antenna gai
n
,
G
R
is receiv
e antenn
a gai
n,
L
s
denote
s
a
n
interfe
r
en
ce margin, an
d S
R
denote
s
receiver
sen
s
itivity. MAPL
minimum of
all
scena
rio
s
will
be conve
r
ted
into cell radi
us u
s
ing O
k
u
m
ura
-
Hatta propa
gation m
odel provided
by
(4). Pro
pag
ation paramete
r
s for calculation
of the cell
radiu
s
a
r
e sh
own in Ta
ble
5.
d
h
h
a
h
f
dB
P
te
re
te
c
L
log
log
.
.
log
.
log
.
.
)
(
55
6
9
44
82
13
16
26
55
69
(4)
Whe
r
e
f
c
den
otes carrier freque
ncy use
d
in LTE,
h
te
denote
s
eNB
antenna hei
ght,
h
re
denotes
use
r
ante
nna
height, and
d
is the
cell
radiu
s
. For
an
alytical pu
rpo
s
e
we a
s
sum
e
a hexag
on
cell
sha
pe which is co
mpri
se
d of 3 secto
r
s, theref
o
r
e e
N
B
cell are
a
ca
n
be cal
c
ulate
d
by (5).
2
2
95
1
d
x
km
L
site
.
(5)
And the num
ber of calcul
a
t
ed eNB is:
site
area
design
L
L
N
(6)
By usin
g all
those
pa
ram
e
ters a
nd Eq
uation
(1
) to
(6
),
we
are
able to
calcu
l
ate the
requi
re
d n
u
m
ber of total
sit
e
(eNB
) in
th
e teste
d
a
r
e
a
as summa
ri
zed in
Ta
ble
6. Up
to thi
s
st
ep
,
we have do
n
e
the covera
ge planni
ng i
n
the tested
area
so that the numbe
r o
f
eNB within the
spe
c
ified
pa
rameter compl
i
es to
LTE te
chni
cal
spe
c
ification
ha
s
b
een
derive
d
and
de
scribe
d in
Figure 2. T
h
e calculated
numbe
r
of e
N
B ba
se
d u
p
on the
above
assum
p
tion
is 4
5
. The
n
we
deploy tho
s
e
eNB throu
g
hout the test
ed are
a
in the simul
a
tion
environm
ent
. Each eNB
is
installe
d 3 sector a
n
tenn
a to improve
capa
city
an
d at the sa
me time red
u
ce
co-ch
a
n
nel
interferen
ce. The next ste
p
of network planni
ng is
to
evaluate the capa
city dimensi
oning in t
he
tested a
r
ea.
3.1. Capa
cit
y
Dimensioning
In the
capa
cit
y
plannin
g
p
r
oce
s
s, we ha
ve
assum
ed t
hat LTE
su
bscrib
ers are
predicte
d
for the next 5 years. Th
eref
ore, the network
th
rou
ghp
ut and cell through
put are
determi
ned.
Table 7. Nu
mber of Sub
s
crib
er b
a
sed
upon the Pre
d
iction
Parameter
Value
Population of Bandung Cit
y
2.455.517
Population gro
w
t
h
ever
y
yea
r
1.1%
Population in 5
years
2.593.574
Cellular penetrati
on (2013
)
120%
Number of
mobile users
3.112.289
LTE penet
ration
3%
Number of
LTE
users
93.369
Table 8. Nu
mber of e
N
B Based o
n
Ca
pacity Planni
ng
Item
Urban
Suburban
DL UL
DL
UL
Net
w
ork
th
rough
put
(Kbit)
261,640.67
1,000,935.77
65,348.26
232,005.3
Cell average thro
ughput
54,399.95
65,279.95
54,399.95
65,279.95
Cell Number
5
16
2
4
Number of
eNB
6
2
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 14, No. 1, March 2
016 : 119 – 1
2
8
124
Total numbe
r of
eNB
8
Figure 2. eNB position on
a di
gital map
of a tested area
Usi
ng thi
s
predicte
d
traffic cal
c
ul
ation.
The d
a
ta
whi
c
h
are
ne
ede
d in
clude
po
pulatio
n
data obtain
e
d
from stati
s
tical muni
cip
a
l city o
ffice [9], the population growt
h
in the city is
assume
d to be 1.1% per year [10],
mobile
pen
etration is 12
0
%
[11], LTE penetratio
n
is
assume
d to be 3%, and the perce
ntag
e of subscrib
er
s b
e
lon
g
to particul
a
r p
r
ovider is 4
2
.1%.
The predi
cte
d
numb
e
r of
LTE sub
s
cri
b
ers
of the pa
rticula
r
p
r
ovid
er is
sho
w
n i
n
Table 7. T
h
e
values
of traffic model pa
rameters are
deter
mi
ned b
a
se
d upo
n the developm
e
n
t servi
c
es from
the data [12]. For the amo
unt of aggre
gated ban
dwi
d
th of 30 MHz a
nd the n
u
mbe
r
of use
r
s
listed in Ta
bl
e 7, the value of netwo
rk
throug
hput
a
nd cell th
rou
g
hput as
well
as the
cal
c
ul
ation
of the numbe
r of eNB ca
n be se
en in T
able 8, wh
ile
the numbe
r o
f
required e
N
B in the network
is cal
c
ul
ated
usin
g Equatio
n (7) a
nd (8
) resp
ectively.
Figure 3. Percenta
ge of co
vered a
r
ea
by
RSRP and
CINR of wh
ole
scena
rio
s
.
Figure 4. Sector traffic map
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Carrie
r Aggre
gation Te
chni
que to Im
prove Cap
a
cit
y
in LTE-Ad
van
c
e
d
Network (Iskan
dar)
125
hput
cellThroug
oughput
networkThr
DL
DL
cell
DL
of
Number
(7)
hput
cellThroug
oughput
networkThr
UL
UL
cell
UL
of
Number
(8)
3.2. Nominal
Plan
In the n
e
two
r
k
de
sign
p
r
oce
s
s, the to
tal site
numb
e
r i
s
d
e
termi
ned from th
e
high
est
numbe
r of co
verage
or
ca
pacity dime
n
s
ioni
ng in o
r
der to me
et both a
s
pe
ct. In the pro
c
e
s
s of
laying site, the result take
s 46 eNB to cover the
entire city of the test
ed
a
r
ea which can
be seen
in Figu
re
2.
Each
site
ha
s th
ree
secto
r
s,
and
ea
c
h
s
e
c
t
or cons
ists
of two trans
m
itters
with
the
same a
z
im
uth, one tran
smitter for PCell and t
he
other for S
c
ell. To examine the network
throug
hput a
nd the netwo
rk BLE
R
in the tested a
r
ea
, we have to kno
w
RS
RP and CI
NR in
sid
e
the tested a
r
e
a
.
Table 9. Weighting di
strib
u
tion of simul
a
ted are
a
ba
sed on mo
rph
o
logy
No Clutter
Class Weight
%
Indoo
r
1 unclassif
i
ed
0
0
2 w
a
t
e
r_bo
dies
0
0
3 seaport
1
0
4 airport
1
0
5 open_are
a
1
0
6 forest
1
0
7 villages
2
0
8 industrial_commercial
3
0
9 residential
4
0
10 urban
3
0
11 building
3
0
12 Rural
1
0
Figure 5. BLER value of wh
ole scen
ario
s
Figure 6. Do
wnlin
k thro
ug
hput of whol
e
scena
rio
s
4.
Simulation Results a
nd Analy
s
is
4.1. RSRP a
nd CIN
R
Cov
e
rage
The
ove
r
all
result pre
d
icti
on
of RSRP
and CI
NR
ca
n be
seen
in
Figure 3
ab
o
v
e. Based
on
simulatio
n
usin
g the
ab
ove pa
ramet
e
rs, th
e p
e
rcentage
of av
erag
e
RSRP
in the
simul
a
ted
area
is ab
ove
95% for all
scen
ario
s
of A1 to A3
, B1 t
o
B3, an
d
C1
to C6. We
fo
und th
e value
of
averag
e
RSRP is
above
-8
1.29 dBm
wit
h
the
mini
mu
m value
of
-1
00 dBm
in
all
scen
ari
o
s.
T
he
CINR mini
mu
m that
spe
c
if
ied in
the
ne
twork of
LTE
-
A is -6.5 dB
. Based
up
o
n
the
data
a
n
d
assumptio
n
a
bove, our
sim
u
lation ha
s shown t
he cov
e
rag
e
for do
wnlin
k CI
NR
predi
ction i
s
95%
whe
r
ea
s for
the uplin
k is 96%. From the result
s of
these predi
ctions, the de
sign
ed net
wo
rk
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 14, No. 1, March 2
016 : 119 – 1
2
8
126
meets the ta
rget and the
r
e
f
ore we are a
b
le to us
ed th
is area a
s
the
simulate
d area to dete
r
mi
ne
the thro
ughp
ut and BLE
R
value
of the imple
m
en
tation of CA
techni
que
s i
n
LTE-Adva
n
c
e
d
netwo
rks.
Figure 7. Upli
nk thro
ughp
u
t
of whole
scena
rio
s
Figure 8. Do
wnlin
k thro
ug
hput ba
sed o
n
band
width P
C
ell da
n Scell
Figure 9. Do
wnlin
k thro
ug
hput ba
sed o
n
CA type
4.2. BLER an
d Throughp
ut
No
w we a
r
e
evaluating th
e network th
rough
put
an
d
BLER u
s
e M
onte-Ca
rlo
si
mulation
method, which its re
sult
s
clo
s
er to
real
ity
becau
se t
h
is meth
od g
enerate u
s
e
r
s ba
se
d on t
he
data traffic. T
he u
s
e
r
follo
ws the
dist
rib
u
tion of t
r
affic map
s
as sho
w
n i
n
Fig
u
re
4, and
the
we
ight
to be con
s
ide
r
ed a
c
cording
to the Poisson distri
but
io
n is cal
c
ulate
d
by weightin
g distrib
u
tion
in
Table 9 [13]. Simulation
s perform
ed 5 times for e
a
ch
sce
nari
o
wit
h
the numbe
r of iterations
in
simulatio
n
s is up to
10
0 ti
mes. T
he
re
sults of th
e th
rough
put a
nd
BLER of
all t
he
simulatio
n
will
be avera
ged.
The
re
sults
of Monte
Ca
rlo
simulatio
n
s fo
r do
wnli
nk B
L
ER value
s
sh
own
in Fi
gure
8. For
the uplin
k, BLER value i
s
ze
ro for
all
scena
rio
s
,
so
that the upli
n
k effe
ctive throu
ghp
ut is
not
different from
the p
e
a
k
th
rough
put. In t
he d
o
wnlin
k
side
there i
s
a tra
n
smi
s
sio
n
e
rro
r
whi
c
h
ha
s
value ra
nge
s from 0.6% to 0.7%. In the intr
a
-
ba
nd
CA only uses the 1
800
MHz f
r
equ
en
cy
downlin
k and
459.61 kbp
s
for uplin
k. While the
avera
ge increa
se
s
of combi
natio
n 15 MHz +
15
MHz to 20
M
H
z
+
20 M
H
z is 1.8
7
Mb
p
s
on
the d
o
wnlink
and
539
.3 kbp
s
on th
e uplin
k. So,
an
increa
se in
throu
ghp
ut of
PCell ba
nd
wi
dth 10 M
H
z to 20 M
H
z in
the do
wnlin
k
can
re
ach 3.
88
Mbps, an
d 99
8.9 kbp
s
on t
he uplin
k.
Furthe
rmo
r
e,
if base
d
on t
he type of
CA, The
graph
in Figu
re 9
shows that in t
he intra-
band
CA, through
put for intra-ban
d no
n-contigu
o
u
s
is hi
ghe
r th
an throug
hpu
t for intra
-
b
a
n
d
contig
uou
s. T
h
is i
s
du
e to the intra
-
b
and
cont
ig
uou
s
CA ch
ann
el u
s
ed to
be very close toget
her,
giving rise to the presen
ce of adjace
n
t chan
nel
inte
rfere
n
ce, whi
c
h ma
ke
s ge
tting lower SINR
value. Th
e g
r
eater the SI
NR valu
e, the
highe
r th
e m
odulatio
n typ
e
u
s
ed,
so th
e be
arer
ch
a
nnel
codi
ng
rate a
nd its
efficien
cy high
er, a
n
d
re
sultin
g th
e high
er th
ro
ughp
ut be
cau
s
e the
num
be
r of
transmitted bi
ts per
symbol
is more.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Carrie
r Aggre
gation Te
chni
que to Im
prove Cap
a
cit
y
in LTE-Ad
van
c
e
d
Network (Iskan
dar)
127
Acco
rdi
ng to
Oku
m
ura-Hat
t
a pro
pag
atio
n mod
e
l in
Equation
(3
), t
he lo
we
r fre
q
uen
cie
s
will
have
a smaller pathloss, so
th
at at
the same
dist
ance from th
e transmitter,
it woul
d produce
a better ra
dio
cha
nnel
con
d
ition for th
e
use
r
. Th
e bet
ter conditio
n
of the chan
ne
l and th
e g
r
e
a
ter
SINR a
nd ef
ficien
cy bea
rer
will re
sult
highe
r thro
ughp
ut. Sce
nario
C1
an
d C5
both
use
freque
nci
e
s
o
f
700 MHz, b
u
t the frequ
e
n
cy of 18
00
MHz
C5
re
place
d
with
a h
i
gher
(2
600
MHz).
As a
re
sult,
the thro
ugh
p
u
t de
cre
a
sed
0.59 M
b
p
s
,
from 7
9
.24
Mbps to 7
8
.75 Mb
ps on
the
downlin
k and
0.76 Mbp
s
o
n
the uplin
k. Furthe
rm
o
r
e,
in scena
rio
C6, the freque
ncy of 700 M
H
z
C1
scen
ario
cha
nge
d to 2
600 M
H
z. Th
e de
crea
se i
n
thro
ugh
put
from
C1 to
C6 is greate
r
t
han
C1 to
C5,
whi
c
h i
s
1.68
Mb
ps
on
the
do
wnlin
k
and
1.
39 M
b
p
s
o
n
t
he u
p
lin
k. Th
rough
put of
int
e
r-
band
CA
scenari
o
C6
e
v
en lower 0.
72 Mb
ps fro
m
intra
-
ba
nd
with the
sa
me ba
nd
width
combi
nation
s
, those are scenari
o
A1 an
d B1 wh
ich o
n
ly use
s
the 1800 M
H
z fre
quen
cy ban
d.
In PCell and
SCell ba
ndwi
d
th com
b
inati
on 10 M
H
z
+ 20 MHz, the
r
e is a
scen
a
r
io C2
(700 MHz
an
d
180
0
MHz) and C4
(900 MHz and 210
0
MHz),
wich have thro
ugh
put re
spe
c
tivel
y
77.56 Mbp
s
and 75.6
4
Mbps fo
r do
wn
link, and 5
3
.03 and 5
1
.0
9 for the upli
n
k. Ju
st like
the
combi
nation
of band
width
15 MHz
+ 15
MHz, thi
s
ti
me the scen
ari
o
that use
s
lo
wer f
r
equ
en
ci
es,
ie C2, re
sulti
ng better through
put than
the sce
nari
o
C4 that u
s
e
s
high
er freque
nci
e
s. T
he
differen
c
e in
throug
hput of
these two scen
ario
s o
n
the do
wnlin
k
dire
ction rea
c
hes
1.92 Mb
ps,
and upli
n
k u
p
to 1.08 Mbps.
5. Conclu
sion
Based
on the
LTE network acce
ss
de
sig
n
in this
stud
y, we ca
n co
nclu
de
several thing
s
.
First, the BLER on uplin
k is zero for all scen
ario
s,
whil
e the downli
n
k BLER is 0.6
3
% to 0.79
%,
the do
wnlin
k
BLER i
s
gre
a
ter fo
r the
h
i
gher fre
quen
cy configu
r
ati
on. Fo
r throu
ghput from t
h
e
netwo
rk si
mu
lation, CA
co
nfigur
atio
n
wi
th wid
e
r P
C
el
l ban
dwidth
result
better th
roug
hput. T
h
en,
based on type of CA in the intra-ba
nd, the
thro
ughp
ut non-contiguo
us CA is better than
contig
uou
s
CA. And the last, in the in
ter-b
and
CA, CC
co
mbina
t
ion with a l
o
wer f
r
eq
uen
cy
prod
uces hig
her th
rou
ghp
ut. CA co
nfiguratio
n
that
pro
d
u
c
e
s
th
e high
est th
rough
put in t
he
desi
gn a
r
ea t
he city is inte
r-ban
d no
n-co
ntiguou
s
CA
with the com
b
ination of 7
0
0 MHz and
1
800
MHz, a
nd th
e band
width
combin
ation
of the
primary and
seconda
ry se
rving cell 2
0
M
H
z
+
10MHz, the throu
ghp
ut is 80.95 Mbp
s
o
n
t
he downlin
k and 5
3
.83
Mbps o
n
the uplin
k.
Referen
ces
[1]
LT
E
for UMT
S
- Evolutio
n to L
T
E-Ad
vanced. John W
i
l
e
y
& S
ons, Ltd. 201
1.
[2]
A Ghosh, R
Ratasuk, B M
ond
al, N Ma
n
galv
edh
e, T
T
homas. LT
E-adv
anc
ed: n
e
xt-gen
eratio
n
w
i
re
le
ss b
r
oa
db
a
n
d
te
ch
no
l
ogy
.
IEEE Wireless Comm
unic
ations.
20
10; 1
7
(3): 153
6-1
2
8
4
.
[3]
D Erik, S Parkvall, J Sköld. 4G
LT
E/L
T
E-Adv
ance
d
for Mobi
le Broa
db
and.
Oxford: Elsev
i
e
r
. 2011.
[4]
L Ha
e
y
ou
ng,
S Vahi
d, K Moessn
er. A Surve
y
of R
adi
o Res
ource
Mana
geme
n
t for Spect
r
u
m
Aggre
gati
on in
LT
E-Advanced
.
Commu
n
ic
ati
ons Surveys &
T
u
torials.
20
14
; 16(2): 745-7
6
0
.
[5]
Al-Shi
bl
y, MAM Haba
ebi, M
H
Che
b
il J,
C
a
rrier a
ggre
gat
ion i
n
Lo
ng T
e
rm Ev
oluti
on-
Advanc
ed.
In
Procee
din
g
of IEEE Control a
nd S
y
stem Graduate
R
e
se
arc
h
Coll
oq
uium (I
CSGRC). 201
2
:
154-15
9.
[6]
Hua W
a
ng, C
Rosa, K P
eder
sen.
Perfor
ma
nce An
alysis
o
f
Dow
n
link I
n
te
r-Band
Carri
er
Aggre
gati
o
n
in LT
E-Adva
nc
ed.
In Procee
di
ngs of IEEE Vehicul
a
r T
e
c
hnol
og
y
Co
nfere
n
c
e
(VT
C
Fall). 2011: 1-5.
[7]
Reka
w
t
S
a
b
i
r
Hassa
n, T
A
Rahm
an, AY
Abd
u
lra
h
man
.
LT
E Covera
ge N
e
t
w
ork
P
l
an
nin
g
a
n
d
Comp
ariso
n
w
i
th Different Pro
pag
atio
n Mode
ls.
Journa
l T
e
lk
omnik
a
.
201
4; 12(1).
[8] Hendra
Setiaw
an
, F
i
rda
u
s.
T
he Most Possible Sc
heme
of Joint Serv
ice Detecti
on
for the Ne
xt
W
i
reless Com
m
unic
a
tion T
e
chno
log
i
es.
Jour
nal T
e
lk
o
m
nik
a
. 2013; 11(
1).
[9]
Centra
l age
nc
y statistics of Bandu
ng
Cit
y: http://
w
w
w
.
b
a
n
d
u
ngkota.b
p
s.go.i
d
/
[10]
Den
n
y
Kusum
a
He
ndra
n
i
ngr
at, et al.
Refa
rmi
ng
of F
r
eq
uency
70
0 M
H
z
A
n
a
l
ysis fo
r Lon
g T
e
r
m
Evoltio
n
(LT
E
)
in Ind
o
n
e
sia
Usin
g Li
nk
Budg
et Calc
ul
ation.
In th
e
proce
edi
ng
of Internatio
na
l
Confer
ence
on
Electrical En
gi
neer
ing a
nd Inf
o
rmatics (IJEEI 2011). 20
11: 1
-
6.
[11]
Indon
esi
a
n
asso
ciatio
n
of
Internet service
provi
ders:
http://
w
w
w
.
a
p
ji
i
.
or.id/v2/ind
e
x
.php/re
ad/p
age/
hal
aman-
data/
9/
statistik.htm.
[12]
ET
SI
T
S
136 101, v1
0.3.0. LT
E, E-U
T
RA, and UE
Rad
i
o T
r
ansmissio
n
an
d Rec
epti
on T
e
chnica
l
Specific
ation. 2
011.
[13]
W
e
iho
ng F
u
, Qingl
ia
ng Ko
ng, Yue Z
han
g, Xi
n Yan.
A resou
r
ce sched
uli
ng
alg
o
rith
m bas
e
d
on carri
er
weight in LTE-
Advanc
ed syst
em
with Carr
ier Aggregation.
In Proce
edi
ngs
of 22
nd
IEEE Wireless and
Optical Comm
unic
a
tion C
onf
erenc
e (W
OCC). 2013: 1-5.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 14, No. 1, March 2
016 : 119 – 1
2
8
128
[14]
Ran Z
h
ang, Z
hon
gmin
g Z
h
e
ng, Mia
o
W
a
n
g
, Xuemi
n
Sh
en, Li
ang-
Li
an
g Xie.
Equivalent capacity
ana
lysis of LT
E-Advanc
ed sy
stems w
i
th carrier ag
greg
atio
n
.
In Proceedi
ng
s of IEEE ICC
Confer
ence
.
201
3: 611
8-61
22.
[15]
DP Sh
arma, S
Gautam.
D
i
stribute
d
a
n
d
Pri
o
ritise
d Sch
e
d
u
lin
g to
Impl
e
m
e
n
t C
a
rrier
A
ggre
gatio
n
in
LT
E Advance
d
Systems.
In Procee
din
g
s
of 4
th
International Confer
ence on IEEE Advanced
Comp
uting & C
o
mmunic
a
tio
n
T
e
chnolog
ies (
A
CCT
). 2014: 390-
393.
[16]
Na Le
i, Cai
li
Guo,
Chu
n
y
an
F
eng, Yu Ch
en.
A multi-us
er MIMO resource sch
edu
lin
g sche
m
e f
o
r
carrier a
ggre
gatio
n scen
a
r
io.
In Proc
eed
ings
of Internati
o
n
a
l
Confer
ence
on W
i
rel
e
ss
Commun
i
cati
o
n
s and Si
gn
al
Processi
ng (W
CSP). 2011: 1-
5.
[17]
Yonis AZ
, Abd
u
lla
h MF
L.
Simulati
on of nov
e
l
non-
adj
ace
n
t compo
nent car
r
iers in LT
E-Ad
vance
d
.
In
Procee
din
g
s of
Internati
o
n
a
l C
onfere
n
ce on
E
l
ectr
on
ics D
e
si
gn, S
y
stems
a
nd A
ppl
icati
ons
(ICEDSA).
201
2: 254-
259.
[18]
Shaj
aia
h
H, A
bde
l-Ha
di A, C
l
anc
y C.
Uti
lity
Proportio
n
a
l
F
a
irness R
e
so
urce All
o
cati
on
w
i
th Carrier
Aggre
gati
on in
4G-LT
E
.
In P
r
oceedings
of
IEEE Militar
y
Communications Confer
ence (MILCOM).
201
3: 412-
417.
Evaluation Warning : The document was created with Spire.PDF for Python.