TELKOM
NIKA
, Vol.14, No
.2, June 20
16
, pp. 515~5
2
2
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v14i1.2646
515
Re
cei
v
ed Se
ptem
ber 3, 2015; Re
vi
sed
Ma
rch 21, 20
16; Accepted
April 10, 201
6
SC-FDMA LTE Performance through High Altitude
Platforms Communications (HAPS) Channel
Iskandar
*
, D. Hida
y
a
t
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
p
o
n
d
i
ng auth
o
r, e-mail: iska
ndar
@
s
tei.itb.ac.id
A
b
st
r
a
ct
It is known t
hat high altit
u
de
platfor
m
s
yst
em
(
H
APS)
is one
of the promising
wireless
technologies that expl
oit
m
any
advan
tages from
c
e
ll
ul
ar sat
e
llite system
as w
e
ll as from
cellular terr
estrial
system. HAPS is envis
age
d to
be a nov
el tec
hno
logy fo
r co
mmu
n
icati
on, b
r
oadc
asting, i
n
ternet back
b
o
n
e
,
earth
obs
erva
tion
an
d surv
eill
anc
e, a
n
d
als
o
for
na
vigati
on. On
e
of u
p
co
ming
techn
o
l
ogi
es
for
communic
a
tio
n
is a
lon
g
ter
m
term
evo
l
utio
n
(LT
E
) of
a c
e
l
l
ular
forth g
e
n
e
ratio
n
(4G).
Many tech
ni
qu
es
have
be
en
dev
elo
ped
to
mak
e
LT
E co
me in
to real
in th
e e
n
viro
nment
of cellu
lar terr
estrial. H
o
w
e
ver, L
T
E
that depl
oye
d
throu
gh HAPS i
s
a chall
e
n
g
in
g
due to di
ffere
n
t
its geometry and ch
an
nel. This pa
per a
i
ms at
eval
uatin
g the
perfor
m
a
n
ce
of a pi
lot-b
a
se
d cha
n
n
e
l esti
mati
on for
upl
i
n
k LT
E usi
n
g
SC-F
DMA ov
er
Ricea
n HAPS
communic
a
tio
n
cha
nne
l. Pil
o
t-base
d
cha
n
nel esti
matio
n
is used to e
s
timate
an u
p
l
i
nk
chan
nel
of
LT
E users
w
ho t
r
ans
mit th
e d
a
t
a to HAPS
as
a
base
transc
e
iver st
ation
(
B
T
S
). Analysis
is
perfor
m
e
d
to
deter
mi
ne th
e
effect of user
’s elev
atio
n a
n
g
le w
i
th r
e
spe
c
t to user p
o
si
tion i
n
sid
e
HA
PS
covera
ge, LT
E
chann
el b
a
n
d
w
idth, mod
u
l
a
ti
on type, an
d th
e Dop
p
ler fre
q
uency sh
ift effect. W
e
found t
h
a
t
user
’
s
e
l
ev
atio
n an
gle c
ontri
bute
ma
jor eff
e
ct to the pi
lo
t-based c
han
n
e
l esti
mati
on
of LT
E SC-F
DMA
perfor
m
a
n
ce.
System c
a
p
abi
lity to ov
erco
me fad
i
ng
e
ffect that us
ers w
i
th low
e
l
ev
atio
n a
ngl
e w
oul
d
b
e
nee
de
d to incr
ease the
perfo
rma
n
ce. In par
ticular to
kee
p
an accept
abl
e
performanc
e, in this pa
per
w
e
compe
n
sate
the c
h
a
nne
l
b
andw
idth,
cha
ngi
ng
mod
u
lati
on typ
e
, a
n
d
li
mit th
e D
o
p
p
l
er Effect thr
o
ugh
vehicl
e spe
ed l
i
mitatio
n
.
Ke
y
w
ords
:
Lon
g T
e
rm Ev
oluti
on, SC-F
D
M
A, High Altitude
Pl
atforms,
pilot-b
a
sed c
han
nel esti
mat
i
on,
Ricea
n fadi
ng
chan
nel.
Copy
right
©
2016 Un
ive
r
sita
s Ah
mad
Dah
l
an
. All rig
h
t
s r
ese
rved
.
1. Introducti
on
Long
Term
Evolution (L
TE) is th
e late
st gene
ration
o
f
mobile
cellu
lar
commu
nicatio
n
s
techn
o
logy
which i
s
devel
oped
by 3rd
Gen
e
ratio
n
Partnershi
p
Proje
c
t (3
GP
P) [1-3]. L
T
E is
desi
gne
d to
be an
efficient cell
ular
techn
o
logy o
n
the u
s
e o
f
frequen
cy
spe
c
tru
m
, hi
gh
transmissio
n data rate (m
ore than 5
0
Mbps o
n
the
uplink an
d 100 Mbp
s
o
n
the downli
n
k),
simple architecturally, support
high mobility comm
unications, l
o
w delay, and high throughput.
LTE uses S
i
ngle
Carrie
r Freq
uen
cy
Divisio
n
Mult
iple Access
(SC-
FDMA
) for
uplin
k a
n
d
Orthog
onal
F
r
equ
en
cy Divi
sion
Multiple
Acce
ss
(OF
D
MA)
for
the d
o
wnli
nk. LTE techn
o
logy h
a
s
many optio
ns to use spe
c
trum b
and
widt
h, starti
ng
fr
om 1
.
4
MH
z
,
3
MHz
,
5
MH
z
,
10
MH
z
,
15
MHz and
20
MHz.
Thi
s
bandwi
d
th flex
ibility brings LT
E to be the
best
technology ever
of cell
ular
comm
uni
cati
on and p
o
tent
ial to offer high-spee
d data
rate [4-5].
Tran
smi
ssi
on
infra
s
tru
c
ture tech
nolo
g
y plays
an im
portant
role
i
n
the reliabili
ty of the
system. Cell
ular LTE terrestrial
syste
m
co
mmo
nly used e
N
B as the radi
o
acce
ss net
work.
Therefore, in
cellul
a
r terre
s
trial system, i
t
needs the g
r
oun
d allo
cati
on and a la
rg
e numbe
rs of
towers to cov
e
r larg
e are
a
s
. This could
be a pro
b
le
m
in the cost of investment to
maintain a high
quality of
services and
coverage.
An idea of
LTE
deploym
ent via
HAPS must
be proven as an
infrast
r
ucture
solution that utiliz
es a transceiver
stations pl
aced
in t
he stratosphere. The
network
architecture f
o
r cellular LT
E deployed v
i
a H
APS can be shown in Figure 1. HAPS advantages
are hig
h
ele
v
ation angle
which bro
a
den Lin
e
of Sight (LOS
) and
coverage area
s, lowe
r
prop
agatio
n
delay compa
r
ed to
satellit
e system,
re
l
a
tively low o
peratio
nal
co
sts a
nd e
a
sy
to
mobilize in
emergency condition
s. HAPS also mi
nimize the problems
of multipath. As an
infrast
r
ucture that utilizes the
medium
of air, characteri
zation
of the channel is important,
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 14, No. 2, June 20
16 : 515 – 52
2
516
becau
se it greatly affect
the perform
ance of
the system. Me
anwhile, ch
a
nnel e
s
timation
techni
que
by usin
g pil
o
t signal i
s
u
s
e
d
to co
mpen
sate ch
ann
el
cha
r
a
c
teri
stic to the
syste
m
.
Integrating HAPS on LTE
provide a
pos
i
tive im
pact on the
world of
mobile c
o
mmunic
a
tions
today. The
s
e
two te
ch
nol
ogie
s
a
r
e
expecte
d to
a
n
swer the
n
eed
of safety, reliable
a
nd
revolution
ary
telecom
m
uni
cation te
chn
o
l
ogy with high
bit rates at lo
w co
st.
Figure 1. Net
w
ork
architecture of c
e
llular LTE via HAPS.
There is very few contributions deal wi
th LTE that deployed via HAPS whi
c
h has a
uniqu
e ge
om
etry ch
ann
el
comp
ared
with cellula
r te
rrestrial
chan
n
e
l. Uplin
k
and
downlin
k of
LTE
perfo
rman
ce
is discu
s
sed
in [6]. Downli
nk pe
rforman
c
e of multipl
e
acce
ss
usi
ng OF
DMA h
a
s
been di
scussed in [7]. However, tho
s
e
papers deal
with cellula
r commu
nicati
on for terrest
r
ial
system
which
is
ba
sed
on
terrest
r
ial to
wer.
The
cha
nnel
ch
ara
c
te
ristic mu
st b
e
differe
nt fro
m
that of HAPS channel
ch
aracteri
stic. In t
h
is
paper, SC-FDMA’s perf
or
mance on
HAPS with pil
o
t-
based chan
n
e
l estimation
will be anal
yzed [8-10]. In
our p
r
eviou
s
wo
rk, we h
a
ve studie
d
the
downlink
LTE c
h
arac
teris
t
ic
ov
er HAPS
c
hannel us
ing
channe
l es
timation algorithm [11].
Ho
wever, fo
r
the uplin
k, L
T
E uses
anot
her
multiple
acce
ss sche
me nam
ely S
C
-F
DMA to
save
the power transmit
so th
at it can sa
ve the
battery life of the UE termin
al. SC-F
DM
A’s
performance
on a HAPS channel is
evaluated based on a computer
sim
u
lation.
The result will
be a
nalyzed t
o
dete
r
min
e
t
he effe
ct of el
evati
on a
ngle
,
cha
nnel
ba
n
d
width, m
odu
lation type
an
d
Dop
p
ler frequ
ency on
syste
m
’s pe
rfor
m
a
nce. Charact
e
risti
c
of ch
a
nnel
is ta
ken
from re
sea
r
ch of
HAPS in Hokkaido, Japan [12].
HAPS i
s
usi
ng Ri
cean channel t
hat modeled
the condition of
Line of Sight (LOS
) and m
u
ltipath due to use
r
’s lo
cati
on and lan
d
m
a
rk
circum
sta
n
ce
s. K factor is
use
d
as a p
a
rameter to indi
cate LOS ratio.
The possible configuration
between cellular network
prov
ided by HAPS and terrestrial
tower
will be very interesting. However, we
have to be
carefully
desi
gned the net
work from
interferen
ce. The cell
cove
rage of HAP
S
must
be separated a
w
ay with enou
gh distan
ce f
r
om
cell
cove
rag
e
of terre
s
trial
BTS to avoi
d co-cha
nnel
interfe
r
en
ce.
An a
r
ea
call
ed bl
an
k
spo
t
of
terrestrial tower
will be cover
ed by HAPS. We found that the
perform
ance evaluation of LT
E
downlink over the HAPS channel ha
s not been much i
n
vestigated.
The rest of p
aper i
s
o
r
ga
n
i
zed
as follo
ws. Section
2 reviews SC-F
DMA on
LTE
while in
se
ction 3
we review
Hi
gh Altitude
Platform
s cha
nnel
model
an
d chara
c
te
risti
c
.
In Section
4,
we
explain simul
a
tion
model
for signal
transmi
ssi
on of
LTE
uplink over HAPS channel. Then
simulatio
n
re
sults
are
an
alyzed in S
e
ction 5.
We
focu
s ou
r an
alysis
on the
effect of user’
s
elevation angle, LTE spectrum ba
ndwidt
h, modulation type, and Do
ppler frequency shift. Finally,
con
c
lu
sio
n
s a
r
e drawn in Section VI.
2. Pilot-base
d SC-F
DM
A
in LTE-HAPS
Channel
SC-FDMA is
LTE’s
multiple acc
e
ss
sc
hemes
fo
r upli
n
k. SC
-FD
M
A can b
e
re
g
a
rde
d
a
s
DF
T-
spread o
r
tho
gonal fre
que
ncy division
multiple
acce
ss, where time dom
ai
n data symbol
s are
transfo
rme
d
to frequ
en
cy domain
by DFT befor
e go
ing thro
ugh
subcarrier m
a
pping p
r
o
c
e
s
s.
The only diffe
rent bet
wee
n
OFDMA a
nd
SC-F
DMA is
an additio
nal
DFT blo
c
k on
the transmitter
and IDFT bl
o
ck o
n
the re
ceiver.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
SC-F
DMA LT
E Perform
ance throug
h Hig
h
Altitude Platform
s Com
m
unication
s… (Iska
nda
r)
517
Figure 2. SC-FDMA re
so
urce gri
d
[8].
Re
sou
r
ce blo
ck
stru
ctu
r
e
on SC-FDMA can
be
de
scrib
ed a
s
in
Figure 2. A reso
urce
block ha
s du
ration of 0.5 ms and b
and
wi
dth of 180
kHz (12
sub
c
a
rri
ers). All the reso
urce blo
c
ks
con
s
titute of a resource g
r
i
d
. The numb
e
r of bloc
ks i
n
the reso
urce grid ra
nge
s from 6 to 100 for
1.4 MHz ch
a
nnel
s to
20
MHz
chan
nel
s, respective
l
y
.
Each uplin
k slot ca
rrie
s
seven
SC-FDMA
symbol
s. The
smalle
st ele
m
ent in
a re
source bl
ock is called
Re
so
u
r
ce Ele
m
ent
whi
c
h contain
s
a
sub
c
a
rri
er for the duration
of one SC-FDMA symbol [1
3-14].
Pilot sig
nal i
s
used
as refe
ren
c
e
si
gnal
that
is requi
re
d to
perfo
rm
cha
nnel
e
s
timation
at
the receiver
output [15]. P
ilot sig
nal i
s
i
n
se
rted
at
sp
ecific sym
bol
and
wh
en it
p
a
ss th
rou
gh t
h
e
cha
nnel, it
will be
p
r
o
c
essed
with
a metho
d
th
at estimate
s ch
ann
el co
ndition a
nd
then
comp
en
sate i
t
to anothe
r
symbol
s. Pilo
t signal
i
s
g
e
nerate
d
ba
se
d on Z
adoff-Chu
se
que
nce.
Zadoff-Chu
seq
uen
ce
co
mmonly refe
rre
d a
s
Co
nstant Am
pli
t
ude Ze
ro
Auto Co
rrela
t
ion
(CAZA
C
)
seq
uen
ce with th
e followin
g
eq
uation.
,
0
(1)
Whe
r
e
q i
s
Z
adoff-Chu
se
quen
ce
ro
ot index, Nzc
is seq
uen
ce
l
e
n
g
th
and
m = 0,1,….Nzc – 1.
Zadoff-Chu has constant amplitude, so
does it’s Nzc-point DFT and PAPR.
High Altitude
Platforms lo
cated in the strato
sp
here, at an altitude be
tween 1
7
and
22 km
above the earth surf
aces.
HAPS have a rapid roll
-out
capab
ility and the ability
to serve a large
numbe
r of
u
s
ers, usi
ng consi
derably
l
e
ss comm
uni
cation
s i
n
fra
s
tructu
re th
an
re
quired
by a
terres
t
rial net
work
.
HAPS loc
a
ted in the s
t
ratos
phere whic
h has
cons
tant
temperature
rise and
con
s
tant win
d
sp
eed rise.
There
is no weath
e
r
p
h
e
nomen
on o
ccurs in this la
yer be
cau
s
e
this
layer ha
s low content of water. That is also this
laye
r is stabl
e with only slight turbule
n
ce. No
clou
ds o
n
this layer thus all
o
w effective u
s
e of sol
a
r po
wer.
Figure 3. Altitude vs HAPS
covera
ge
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 14, No. 2, June 20
16 : 515 – 52
2
518
HAPS coverage depends on several factors such
as altitude, user’
s
elevation angle,
and e
a
rth di
mensi
on [16]
. In Figure
3, we
can
see rel
a
tion
betwe
en altit
ude a
nd HA
PS
coverage. HA
PS
coverage is
expresse
d in
maxi
mum
diamete
r
of L
O
S com
m
uni
cation. Fi
gure
2
sho
w
s the di
ameter of L
O
S commu
nica
tion at altitude ran
g
ing fro
m
1 to 10
5
km as a fun
c
ti
on of
elevation an
g
l
e. The graph
was m
ade
with the assum
p
tion that the
prop
agation
is straight, so
it
c
an be s
a
id that the higher HAPS,
the broader the sc
op
e, but wit
h
a limitation that the c
o
verage
area i
s
sm
all
e
r than ea
rth
diamete
r
.
In ca
se of HA
PS chann
el, Ricean fadi
n
g
is a
g
ene
ral
ca
se of fadin
g
ch
annel m
o
del that
there
are two
com
pon
ents of si
gnal
arri
ve at the
re
ceiver. First
compon
ent a
r
rives at
re
ceiv
er
throug
h line
of sight (L
OS) path
while the second
come
s f
r
om
scattere
d sig
nal. In SPF
comm
uni
cati
on scena
rio, i
t
is proba
bly to get both compon
ents b
e
ca
use SPF is highly located
above the gro
und.
Con
s
e
quently
, it was found
that the Rice
an
fading ch
annel is a
n
a
ppro
p
ri
ate model for
the ca
se of SPF link with
K factor varie
s
dep
endi
ng
on the elevati
on angl
e and
the frequen
cy.
SPF cha
nnel
can
be
ch
ara
c
teri
zed
u
s
in
g Ri
cian
di
stri
bution a
s
foll
ows. Where
K
is
Rice fac
t
or,
(
t
) is the u
s
ers elevatio
n angle,
f
D
is Doppl
er shift from receiver moveme
nt, and
h
(i) i
s
th
e
scattered
co
mpone
nt. If the total po
we
r of scattere
d
signal i
s
de
n
o
ted by
2
and
power of L
O
S
sign
al rep
r
e
s
ented a
s
, then the total received po
wer
and
K
factor
are given by:
)
(
)
(
)
(
cos
(
t
h
K
e
K
K
t
x
t
D
f
j
1
1
1
2
(2)
2
(3)
K
σ
(4)
Then the Equ
a
tion (1
) ca
n be re
written a
s
follows:
H
H
H
(5)
Whe
r
e
is LO
S compo
nent
, and
is scattered
com
pon
ent.
Figure 4. SC-FDMA sim
u
la
tion model.
Table 1. SC-FDMA LTE HAPS Simulati
on Parameters
.
Specification Par
a
meters
Value
Channel band
w
i
dth (MHz)
1.4, 3, 5, 10, 1
5
,
and 20
Number of S
ubcarrier
1200
Number of
resou
r
ce block
25, 75, and 1
0
0
DFT size
1024 and 20
48
CP
108 and 144
Carrier f
r
eque
ncy (
G
Hz)
2.4
Signal constellati
on
QPSK and 16
-QAM
Channel model
AWGN and
Ricean fading channe
l
Doppler shift (HR
z
)
50, 100, and
15
Users-to-HAPS e
l
evation angle
10-to-9
0 degrees
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
SC-F
DMA LT
E Perform
ance throug
h Hig
h
Altitude Platform
s Com
m
unication
s… (Iska
nda
r)
519
3. SC-FDMA
in HAPS Channel Simulation Model
The
pro
c
e
s
si
ng of
SC-FDMA sig
nal i
s
transmi
tted v
e
ry
simila
r to
that of
OF
DMA. The
seq
uen
ce
of
bits tra
n
smitted for ea
ch
u
s
er, i
s
map
p
ed into
a
co
mplex con
s
te
llation of
sym
bols
s
u
c
h
as
BPS
K, QPSK or
M-QAM. Then different
trans
mitters (us
e
rs)
are ass
i
gned
different
Fouri
e
r
coeffi
cient
s. This
assignm
ent is ca
rri
ed out
in the mapp
ing and d
e
-mappin
g
blo
c
ks.
Pilot-ba
sed
cha
nnel
e
s
timation,
whi
c
h is u
s
ed
t
o
e
s
timate
the pe
rform
ance of
sig
nal
transmissio
n of SC-FDMA LTE on HAP
S
chann
el, wa
s evaluate
d
throug
h co
mp
uter sim
u
latio
n
s.
The
stru
cture
of simul
a
tion
model i
s
d
e
p
icted i
n
Fi
g
u
r
e 4. At the transmitte
r, the
se
rie
s
of bit
is
gene
rated
an
d converte
d f
r
om
se
rial to
parall
e
l, t
hen
modulate
d
int
o
symb
ol. Pilot sig
nal i
s
th
en
inse
rted at e
a
ch first sym
bol in all
sub
c
arri
e
r
. The
s
e modul
ated
symbol
s an
d
pilots pe
rform
M-
point Discret
e Fourie
r Tra
n
sform (DFT) to pr
odu
ce a freque
ncy domain rep
r
e
s
entatio
n of the
symbol
s. It then ma
ps ea
ch of
the
M-point DFT ou
tputs to on
e
of the orth
og
onal
sub
c
a
rri
ers
mappin
g
that can b
e
tran
smitted.
The receiver
side i
n
cl
ude
s one d
e
-m
ap
ping bl
ock, o
ne IDFT
blo
c
k, and
one
d
e
tection
block for
ea
ch user
sign
al to be re
ceiv
ed. Ju
st like
in OFDM, guard i
n
terval
s (calle
d cy
cl
ic
prefixes)
with
cycli
c
re
petition a
r
e
intro
d
u
ce
d b
e
twe
e
n
blo
c
ks of
symbols in vie
w
to
efficientl
y
eliminate int
e
r-symbol
int
e
rferen
ce fro
m
time
spre
ading
(cau
se
d by multi-p
a
th propa
gati
on)
among th
e bl
ocks. In SC-FDMA, multi
p
le a
c
cess
a
m
ong u
s
e
r
s is mad
e
po
ssi
b
le by assig
n
i
ng
different use
r
s different sets of no
n-overlap
p
i
ng
Fouri
e
r-coefficient
s (sub
-carri
ers). Thi
s
is
achi
eved
at the tra
n
smitter by in
se
rting
(p
rior to
IFF
T
)
silent fo
urier-co
efficient
s
(at po
sitio
n
s
assign
ed to o
t
her users), a
nd rem
o
ving t
hem on the receive
r
sid
e
after the FFT.
In this pap
er,
distrib
u
ted
method i
s
u
s
ed for
sub
c
a
r
rier m
appi
ng.
In this method, the
outputs a
r
e a
llocate/d eq
u
a
lly s
paced subcarrier
with
zero
s occ
u
p
y
ing the unused sub
c
a
rri
er in
betwe
en. T
h
e
n
IDFT
bl
ock
followe
d by
Cyclic P
r
efix
(CP) i
n
sertion.
Cycli
c
pr
efix is a
copy
of th
e
last p
a
rt
of symbol that
pl
ace
d
in
front
of sy
mb
ol th
at ca
n eli
m
in
ate Inter Symbol Inte
rference
(ISI). Then t
he si
gnal i
s
t
r
an
smitted through the
HAPS Ricean
channel. At the receiver,
the
oppo
site
set
of the op
eration i
s
pe
rform
ed. CP i
s
re
moved the
n
t
he
signal
is p
r
ocesse
d by t
he
DFT. Pilot si
gnal then b
e
i
ng extra
c
ted
to get
the chann
el co
ndi
tion. Chan
nel
conditio
n
th
en
being
comp
e
n
sate
d to the other sym
bol
s.
Table 2. Simulation pa
ram
e
ters of
chan
nel ban
dwidt
h
investigatio
n
Specification Par
a
meters
Sample#1
Samp
le#2 Sample#3 Sample#4
Sample#5 Sample#6
Channel band
w
i
dth (MHz)
1.4
3
5
10
15
20
Modulation QPSK
QPSK
QPSK QPSK
QPSK QPSK
Number of
resou
r
ce block
6
15
25
50
75
100
Number of sub
-
c
a
rrier
72
180
300
600
900
1200
CP
9
18 36 72
108
144
DFT
size
128 256 512
1024
1536
2048
Doppler
shift
(Hz
)
50 50 50 50
50 50
Bit
rate
(Mbps)
0.9 2.2 3.6 7.2
10.8
14.4
The re
ceived
sign
als
are d
e
-ma
ppe
d,
th
en
ID
FT
ope
ration i
s
p
e
rfo
r
med. Th
ese
receive
d
sign
als are d
e
modul
ated t
o
get th
e bit
strea
m
. Bit st
ream
in the
receive
r
i
s
the
n
compa
r
e
d
with
bit s
t
ream in
the trans
m
itter to get Bit
Error Ra
te (B
ER). In orde
r to kno
w
the
perfo
rman
ce
of
SC-F
DMA L
T
E transmitted in HAPS
cha
nnel
we
then propo
sed the follo
wing pa
ramete
r of
simulatio
n
a
s
in T
abel
1.
These
param
eters
a
r
e ba
sed
on
LTE
spe
c
ification and also HA
PS
cha
nnel p
a
ra
meter which i
s
de
rived fro
m
our p
r
ev
io
us expe
rime
n
t
[4]. Accordi
ng to the re
sults,
multipath fadi
ng are ob
served and sho
w
n that t
he fading d
epth
woul
d have to vary betwe
en
1
dB and
more
than 2
5
dB
depe
nding
o
n
the el
evat
io
n angl
e. Not
e
that in the m
easure
m
ent
we
use
d
a
n
o
m
n
i
dire
ctional
a
n
tenna.
We
t
hen
ch
arac
te
rize
the
st
rat
o
sp
heri
c
platf
o
rm
ch
ann
el
by
usin
g metho
d
of mome
nt to find Ri
ce
para
m
eter
(K
). Another propag
ati
on pa
ramete
r that
we
have fou
nd f
r
om the
data
o
f
mea
s
ureme
n
t is l
o
cal
m
e
an
re
ceived
p
o
we
r. Both K
factor an
d lo
cal
mean re
ceive
d
po
wer are evaluated un
der
the
va
ri
ati
on of elevatio
n angl
e.
Our evaluation sh
ow
that the K factor wo
uld hav
e to vary from 0.9 to
18.6
dB for a freq
uen
cy ca
rrie
r
of 1.2 GHz in
the
measurement
and 1.4 to
16.8 dB at freque
ncy
2.4
GHz. Stand
ard deviatio
n
of local mean
received
po
wer i
s
fo
und
t
o
de
crea
se
a
s
el
evation
a
ngle i
n
crea
se
indi
cating
little multip
ath i
n
high el
evation angl
e. Up
to this poi
nt we h
a
ve
descri
bed t
he chan
nel
cha
r
a
c
teri
stic in
strato
sph
e
ri
c platform
com
m
unication.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 14, No. 2, June 20
16 : 515 – 52
2
520
4. Performan
ce Analy
s
is
4.1. Users El
ev
ation Angle Analy
s
is
Simulation par
a
meters
of
the SC-
F
DMA LT
E on HAPS ar
e
s
u
mmar
iz
ed in Table 1 and
the re
sult
s is sho
w
n
in Fi
gure
5. Thi
s
figure
sh
ows that the
hi
gher t
he el
e
v
ation angl
e, the
performance
of SC-F
DMA
will be
better.
This i
s
co
nsi
s
tent with the notion that t
he greater the
LOS si
gnal
p
o
we
r
receive
d
by the
re
cei
v
er, t
hen th
e
better the
pe
rforman
c
e
of
SC-F
DMA. Hi
gh
elevation ang
le mean
s hig
h
K factor, be
cau
s
e
K facto
r
is the ratio
betwe
en average LOS po
wer
and average
multipath sig
nal’s p
o
we
r. Whe
n
the
ele
v
ation angle i
s
high, the probability of LOS
comm
uni
cati
on lin
k is al
so high. It me
ans m
o
re LO
S signal
will
be re
ceive
d
by the re
ceiv
er.
Perform
a
n
c
e
differen
c
e
s
d
ue to chan
ge
s in elevation
angle will be
more sig
n
ificant when Eb/
N
o
is more than
6 dB. Eb/No repre
s
e
n
ts the
ratio bet
wee
n
sign
al with
noise, so wh
en Eb/No i
s
less
than 6 dB, influen
ce of the elevation ang
le is less
cle
a
r
ly visible be
cause the sign
al to noise ratio
is too small. Whe
n
Eb/No
is mo
re than
6 dB, signifi
cant differen
c
e
s
will b
e
sho
w
n bet
wee
n
10
0
-
40
0
elevation
angle
and
5
0
0
-90
0
el
evation angle. Therefore
com
m
unication
will be optimum if
elevation
angle bet
ween the tran
smitter and HAPS
i
s
more
than
40
0
a
n
d
Eb
/N
o is
mo
r
e
th
an
6
dB. At that condition, we ca
n obtain BE
R differe
n
c
e a
b
out 0.018
2 to
0.0905
46 at
same elevatio
n
angle.
4.2. Chann
e
l Band
w
i
d
t
h
Analy
s
is
Another
re
sul
t
of our i
n
vestigation on
SC
-F
DMA
sch
e
me b
a
sed o
n
pilot-aided
cha
nne
l
estimation on a HAPS channel is an effect of cha
nnel bandwidth. Our
simula
tion is perform
e
d at
6 chan
nel b
and
width, wh
ich are 1.4 MHz,
3 MHz, 5 MHz, 10 MHz, 15 M
H
z a
nd 20
MHz.
Simulation p
a
ram
e
ters a
r
e sum
m
ari
z
e
d
in T
abl
e 2. Figure 6
sho
w
s
that the greate
r
the
band
width
of
the chann
el, then th
e pe
rfo
r
man
c
e
w
ill
b
e
worse. BE
R d
e
crea
se
s
rangi
ng from
1.4
MHz to 20 M
H
z. Thi
s
is be
cau
s
e the g
r
e
a
ter t
he ch
an
nel band
widt
h, the greater the noise po
we
r
contri
bute to
the ch
ann
el. SC-F
DMA i
s
a multiple
acce
ss
sche
me
that ha
s lo
w
pea
k to ave
r
age
power ratio (PAPR).
Based
on s
i
mulation
results
,
performance
will be decr
eas
e
d from
half to 12
times wh
en
comp
ared wi
th 1.4 MHz
band
width p
e
rform
a
n
c
e. Larg
e
ch
ann
el band
width
with
poor p
e
rfo
r
m
ance can b
e
solved with in
cre
a
si
ng
the
power tran
sm
it on t
he user
equipm
ent si
de.
A com
p
romise bet
ween el
evation angle (coverage
) and
the channel
band
width usage will
l
e
ad
us to an advantage wi
reless
system brought by HAPS.
Figure 5. SC-FDMA sch
e
m
e
perfo
rman
ce
throug
h HAP
S
chann
el as
a function of
elevation an
g
l
e
Figure 6. SC-FDMA sch
e
m
e
perfo
rman
ce
throug
h HAP
S
chann
el as
a function of
Cha
nnel ba
n
d
width
4.3. Modulati
on T
y
pe Analy
s
is
We
co
nsi
der mod
u
lation
type dep
end
ency i
n
o
u
r
observation
of SC-FDMA
sche
me
performance
in a HAPS
channel to look at t
he possibility of user equipm
ent hardware power
efficiency. Simulation perf
ormed
at two different modulations
of QPSK and 16-QAM to represent
low level and
high level mo
dulation resp
ectively. Figure 7 sh
ows the simulatio
n
result
s usin
g 4
0
0
and 90
0
elev
ation a
ngle i
n
ord
e
r t
o
kno
w
thei
r p
e
rfo
r
mance
in
a g
ood and bad
con
d
ition
s
of the
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
SC-F
DMA LT
E Perform
ance throug
h Hig
h
Altitude Platform
s Com
m
unication
s… (Iska
nda
r)
521
cha
nnel. T
h
e
re
sult i
s
th
a
t
there
is a
big differen
c
e bet
wee
n
p
e
rform
a
n
c
e
o
f
SC-F
DMA
with
QPSK modul
ation and
wit
h
the 16-QA
M
modulation.
This is
because
QPSK only use 2
bits per
symbol while the 16-QAM
uses 4 bits
per
symbol
, so the 16-QA
M
will be more susceptibl
e
to
noise du
rin
g
tran
smi
ssi
on.
In the 16
-QA
M
mod
u
lation
, con
s
tellation
of ea
ch
point
is
clo
s
e
r
to t
h
e
other
point than QPSK, s
o
that t
he noise would be
more lik
e
ly t
o
occ
u
r. Average dis
t
anc
e
of
points
on QPSK c
ons
tellation is
2
√
2 whi
l
e 16-QAM a
v
erage di
stan
ce is 2
√
10. If that distan
ce
is
compared on dB, we
will get 7 dB differences. T
h
is
means that t
o
get the same BER, 16-QAM
requi
res approximately 7 dB from the QPSK need.
From the sim
u
lation results,
for QPSK with
90
0
elevation
angle with E
b
/N
0
= 0 dB,
we get 0.206
7 of BER. So
, 16-QAM will
need E
b
/N
0
= 7
dB to get the
same BER. Howeve
r, beca
u
se in the
si
mulation, E
b
/N
0
is increa
si
ng by 2 dB, then
th
e
c
l
os
es
t E
b
/N
0
is 8 dB which h
a
ve 0.2
277 BER Diff
eren
ce b
e
twe
en the simul
a
tion results wi
th
the calcul
atio
n be
cau
s
e
th
e sim
u
lation
s ca
rri
ed o
u
t with Eb/No i
n
crease pe
r
2 d
B
, so the
re
sults
are not very a
c
curate.
Figure 7. SC-FDMA sch
e
m
e
perfo
rman
ce
throug
h HAP
S
chann
el as
a function of
modulatio
n type usi
ng 40
0
and
90
0
eleva
t
ion
angle
Figure 8. SC-FDMA sch
e
m
e
perfo
rman
ce
throug
h HAP
S
chann
el as
a function of
Dop
p
ler frequ
ency u
s
ing 7
0
0
elevation a
ngle
4.4. Doppler
Freque
nc
y
Analy
s
is
The last investigation in SC-FDMA perform
ance via HAPS
channel is an effect of
Doppler
shift
to investigate the mobili
ty char
a
c
te
ri
stic of the
use
r
eq
uipm
ent. Simulation
perfo
rmed
wit
h
paramete
r
sho
w
n in T
a
b
l
e 1, but wi
th
4 different Do
ppler frequ
en
cie
s
whi
c
h i
s
0,
5, 10
and
15
0 Hz. Fi
gure
8 sho
w
s the
simulatio
n
re
sult. It shows that the
pe
rforma
nce i
s
b
e
tte
r
whe
n
the
Do
ppler freq
uen
cy is
sm
aller.
Dop
p
ler fr
eq
u
ency
of HAP
S
is le
ss influ
ential on
grea
ter
elevation a
n
g
l
e. This is du
e to the
Do
pp
ler E
ffect i
s
in
fluenced by t
r
ansmitte
r mo
vement toward
or away from
HAPS. The
greater the
angle, the e
ffect will be less signifi
cant, j
u
st like what
we
get from elev
ation angle a
nalysi
s
. In 2.4 GHz
frequ
ency, the sp
eed of user
equipm
ent for the
Dop
p
ler
shift
of 150
Hz will
be
aroun
d 2
00
km/h.
We
again
will
hav
e a
trad
e-off betwe
en
HA
PS
coverage
whi
c
h i
s
dete
r
mi
ned by an
ele
v
ation angle
and the m
obil
i
ty of user eq
uipment. With
out
Dop
p
ler com
pen
sation te
chniqu
e, LTE
with ten
s
Mb
ps of
data transmi
ssion
rate in a
HA
PS
system
woul
d have mod
e
r
ate pe
rform
a
nce. It is
req
u
ired
com
p
e
n
satio
n
tech
n
i
que to impro
v
e
the performa
n
ce in a high
bit rate transmissi
on while
user i
s
moving with a very high spee
d. This
is our
chall
e
nge for future i
n
ve
stigation of LTE deployed via HAPS.
5. Conclu
sion
We have i
n
vestigate
d
and
prop
osed an
anal
ysi
s
of SC-F
DMA LT
E signal p
e
rf
orma
n
c
e
transmitted
via HAPS
channel in which it’s
fadi
ng follow Ri
cean di
stri
bution
based
on
experim
ental
data
colle
ctio
n. Simulation
wa
s ca
rri
ed
out to
evalua
te an
effect
o
f
use
r
elevati
on
angle,
LTE chann
el ba
nd
width, mo
dul
ation type, an
d
Dopple
r
shi
ft effect. We f
ound
that a
s
use
r
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 14, No. 2, June 20
16 : 515 – 52
2
522
elevation an
g
l
e increa
se, S
C
-F
DMA LTE
system’
s
pe
rforman
c
e i
s
a
l
so in
crea
se.
The in
creasi
n
g
of the
cha
n
nel b
and
widt
h cau
s
e
s
th
e large
r
n
o
ise
ban
dwidt
h
, therefo
r
e
the
system
’s
perfo
rman
ce
decrea
s
e
d
when the ch
an
nel band
widt
h is increa
se
d. Modulation
types which have
fewer bits
p
e
r
symbol
h
a
ve a b
e
tter perfo
rma
n
ce
. Finally, with Do
pple
r
shift of 70 Hz the
particula
r system perfo
rma
n
ce of SC-F
DMA LTE rea
c
h an una
ccep
table perfo
rm
ance.
Referen
ces
[1]
Stefania S, Iss
a
m T
,
Matthe
w
B. LT
E the Long T
e
rm Evolu
t
ion from T
heo
r
y
to Practic
e
.
John W
i
l
e
y
and So
ns Ltd. 200
9.
[2] Morray
Rumney
.
Ed
i
t
o
r
s
.
L
T
E and the
Evoluti
on to
4G W
i
reless
Desig
n
a
nd
Measur
emen
t
Chal
le
nges. U
K
: Agilent T
e
chnol
ogi
es. 200
8.
[3]
3GPP T
S
36.211 v
8
.9.0, 3
G
PP.
T
e
chnic
a
l Sp
ecif
icati
o
n Group
Ra
di
o Access N
e
tw
o
r
k. Evolv
e
d
Univers
a
l T
e
rrestrial R
adi
o Access (E-UT
R
A). Ph
y
s
ic
al C
han
nel
and Mo
dul
ation (R
el
ea
se 8). 2009.
[4]
G Wu, R Miura, Y Hase.
A
Broad
ban
d
W
i
reless Acce
ss System Usi
ng Stratosp
her
ic Platfor
m
s
.
Procee
din
g
of Globa
l T
e
lecommunicati
ons
C
onfer
ence. S
an F
r
ansisc
o
. 200
0; 1: 225-2
30.
[5]
D Grace, M M
ohor. Bro
a
d
b
a
nd C
o
mmun
i
ca
tions vi
a
Hi
gh
Altitude P
l
atfor
m
s. John W
i
l
e
y & S
ons
Ltd.
201
1.
[6]
A
w
s Yo
nis, MF
L Abdul
lah.
Upli
nk an
d Do
w
n
li
nk
of LT
E Rele
ase 10 i
n
Cell
ul
ar Co
mmunicati
ons
.
Internatio
na
l Journ
a
l of Infor
m
atics
a
nd C
o
mmu
n
icati
on T
e
chn
o
lo
gy (IJ-ICT
)
. 2012; 1(1)
: 43-53.
[7]
Nasar
udd
in
N,
Meli
nd
a M, E
l
lsa F
S
. A M
o
d
e
l to
Investi
gat
e Perform
anc
e
of Ortho
gon
al
F
r
equ
enc
y
Cod
e
Div
isio
n
Multipl
e
xi
ng.
T
E
LKOMNIKA Indo
nesi
an J
o
u
r
nal
of Electric
al En
gin
eer
ing
.
201
2; 10(
3):
579-
585.
[8]
H
y
un
g G M
y
u
ng, Davi
d J Goodm
an. Sin
g
l
e
Carrier F
D
M
A
, A Ne
w
Air
Interface for Lon
g T
e
rm
Evoluti
on. UK: John W
i
l
e
y
an
d Sons, Ltd. 20
08.
[9]
MM Ran
a
. A Pilot Bas
ed R
L
S Ch
an
nel E
s
timation for
L
T
E SC-F
DMA in Hi
gh
Do
pp
ler Spr
ead
.
IJCSIS
. 2010; 8(6).
[10]
MM Rana. C
han
nel Estim
a
tion Al
gorithm
s,
Compl
e
xitie
s
and
LT
E Impleme
n
tatio
n
Chal
le
nges
.
IJCSIS.
2010; 8(8).
[11]
MRK Aziz, Iska
ndar.
Channel estimation
for LT
E downlink in High A
l
titude
Platform
s
(HA
P
s) system
s
.
ICoICT
2013. 2013: 18
2-1
86.
[12]
Iskandar, S
h
ig
eru Sh
imam
oto.
Cha
n
n
e
l c
haracter
i
zatio
n
and
p
e
rformance
eva
l
uati
o
n of mo
bil
e
communic
a
tio
n
emplo
y
i
ng str
a
tosph
e
ric pl
at
forms.
IEICE
T
r
ansactio
n
s o
n
Co
mmu
n
icat
ions
. 20
06;
89(3): 93
7-9
4
4
.
[13]
Che-K
ang S
u
n
.
Chann
el Esti
mation a
nd Eq
ualiz
at
io
n of SC-F
DMA for LT
E Uplink S
y
st
em. Nation
al
Centra
l Univ
er
sit
y
, T
a
i
w
a
n
. 2
009.
[14]
Bahattin Kar
a
k
a
ya, Huse
yi
n Arslan, Haka
n
Ali Cirpa
n.
Chan
nel Esti
mat
i
on for LT
E
U
p
link
i
n
Hig
h
Dop
p
ler Spr
e
a
d
. In Proceed
in
gs of IEEE WC
NC. Las Veg
a
s
.
2008; 1: 112
6
-
113
0.
[15] B
y
u
ng J
a
n
g
Jeon
g, H
y
un
K
y
u
Ch
ung.
Pilot Structure
s
for the Up
l
i
nk Si
ngl
e Ca
rrier F
D
MA
Transmission System
s
. V
ehi
cular
T
e
chnol
o
g
y
C
onfere
n
ce
(VT
C
Spring).
Sin
gap
ore.
20
08: 1:
25
52-
255
6.
[16]
Iskandar, S
h
i
g
eru S
h
imam
oto.
On th
e D
o
w
n
link P
e
rfor
ma
nce
of Str
a
tosph
e
ric
Pla
tform Mo
bi
le
Co
mmun
icati
o
ns Ch
ann
el
. In
Procee
din
g
s o
f
Global T
e
leco
mmunicati
on
C
onfere
n
ce. Sa
n
F
r
ansisco.
200
6; 1: 1-5.
Evaluation Warning : The document was created with Spire.PDF for Python.