Indonesian Journal of
Electrical
Engineer
ing and
Computer Science
V
o
l. 10
, No
. 3, Jun
e
20
18
, pp
. 10
07
~
1
012
ISSN: 2502-4752,
DOI: 10.115
91/ijeecs
.v10.i
3.pp1007-1012
1
007
Jo
urn
a
l
h
o
me
pa
ge
: http://iaescore.c
om/jo
urnals/index.php/ijeecs
Interference Temperature Measurements and Spectrum
Occupan
c
y Evalu
a
tion in t
h
e Context of Cognitive Radio
Pauls
o
n N.
Eb
erechukwu,
Daud
a S. Um
ar
,
Alias Mo
hd,
Kam
a
ludin
M. Y,
M.
Adib
b
i
n Sari
jari,
Roz
e
ha
A
.
Ra
shid
Faculty
of Electr
ical
Eng
i
neering
,
Universiti Tekn
ologi
Malay
s
ia,
81310 Skudai, J
ohor Bahru, Malay
s
ia
Article Info
A
B
STRAC
T
Article histo
r
y:
Received
Ja
n 15, 2018
Rev
i
sed
Mar
12
, 20
18
Accepted
Mar 28, 2018
This paper presents a refined
ra
dio s
p
ectru
m
m
eas
urem
ent platform
s
p
ecifi
cal
l
y
des
i
gned for s
p
ectr
u
m
occ
upancy
survey
s in th
e contex
t of
Cognitive r
a
dio. Cognitive rad
i
o
perm
its
the opportunistic usag
e of licensed
bands b
y
unlicensed users without ca
using h
a
rmful interfer
e
nce
to th
e
lic
ensed user. In this work, a stud
y
based on the measurement of the 800
MHz
t
o
2.
4 GHz
fre
que
ncy
ba
nd a
t
t
w
o di
ffe
r
ent
l
o
ca
t
i
ons i
n
side
Uni
v
e
r
siti
Teknologi Malay
s
ia (UTM)
,
Johor Bahr
u cam
pus, Malay
s
ia
is presented.
Two Tek
t
ronix
RSA306B spectrum an
aly
zer are set up
to condu
ct
simultaneous measurements at d
i
fferent
locations
for a 24 hours
period.
The
analy
s
is
conducted in
this work
is base
d on
th
e
real spectrum data
acquir
e
d
from environment in the experimental se
t up. Bus
y
and idle channels were
identif
ied
.
Th
e
channe
ls subjec
t
to ad
jac
e
nt-
c
ha
nnel in
terf
erenc
e
were
also
identif
ied
,
and t
h
e im
pac
t
of
th
e de
te
ction
thres
hold used
to d
e
t
ect
ch
anne
l
a
c
t
ivitie
s
wa
s a
l
so disc
usse
d.
The
consisten
c
y of the observed channel
occupation over a range of thres
holds and a sudden drop has good
chara
c
t
e
risti
c
s in determ
in
ing an appropriate th
reshold needed
in order to
avoid interf
eren
ce.
K
eyw
ords
:
Co
gn
itiv
e rad
i
o
Interfere
n
ce
tem
p
erature
Measurem
ent ca
m
p
aign
Spectrum
occupancy
Tekt
r
oni
x R
S
A
3
0
6
B
W
i
rel
e
ss c
o
m
m
uni
cat
i
ons
Copyright ©
201
8 Institut
e
o
f
Ad
vanced
Engin
eer
ing and S
c
i
e
nce.
All rights re
se
rve
d
.
Co
rresp
ond
i
ng
Autho
r
:
Ro
zeh
a A. Rash
id
,
Facu
lty of Electri
cal Engineering,
Un
i
v
ersiti Tekn
o
l
o
g
i
Malaysia,
8
131
0 Joho
r B
a
h
r
u
,
Joho
r, M
a
laysia.
Em
a
il: ro
zeh
a@u
t
m
.
m
y
1.
INTRODUCTION
A
s
w
i
r
e
less dev
i
ces and
w
i
r
e
l
e
ss co
mm
u
n
i
catio
n
techno
logy is b
e
in
g
adv
a
n
ced, th
er
e is an
in
cr
eased
scarcity of the
ra
dio s
p
ectrum
resources
.
In the
m
eanti
me, stud
ies show
th
at t
h
e
pres
ent s
p
ectrum
allocated
isn’t optim
a
lly utilised. Duri
ng the last fe
w years, cognitive radio (C
R
)
has appeare
d
as a technology that
pr
om
i
s
es t
o
pu
t
an en
d t
o
t
h
e pr
obl
em
of s
p
ect
r
u
m
s
carci
ty. Th
is con
c
ep
t relies
on the basic prem
is
e that
spectrum
is cu
rre
ntly underut
ilized. In
CR, t
h
e secondary
user (SU) is
permitted to sim
u
ltaneously access the
licen
sed
fr
equen
c
y b
a
nd
of
th
e p
r
im
ar
y user (PU) as long as the interfe
rence cause
d by the SU
to
th
e PU is
kept
bel
o
w a
pre
d
efi
n
ed t
h
r
e
sh
ol
d. T
h
i
s
ensu
res t
h
e s
p
e
c
t
r
um
i
s
dy
nam
i
cal
ly
ut
i
l
i
z
ed, he
nce, t
h
er
e i
s
a
sig
n
i
fican
t
imp
r
ov
em
en
t in
th
e sp
ectru
m
u
tilizat
io
n
rate.
In
o
r
d
e
r to en
ab
le d
y
n
a
m
i
c
sp
ectru
m
u
tili
zatio
n
(DSU), t
h
ere is
a nee
d
to critically co
m
p
rehend the am
ou
n
t
of s
p
ect
r
u
m
used
by
t
h
e PU, identify
the pre
s
ence
or the a
b
senc
e of PU, therefore ha
ving a form
id
able SU access pol
i
cy. The
m
eas
urem
ent of spectrum
occupa
ncy level and a
n
alysing t
h
e interference tem
p
eratur
e
wh
ich
is th
e core aim
o
f
th
is p
a
p
e
r, has b
e
en
m
o
tivated by t
h
e ne
eds
stated above. T
h
is
problem
is
addresse
d usi
n
g s
p
ectrum
da
ta collected at an i
n
door
en
v
i
ron
m
en
t o
f
two spatially d
i
ffere
nt locatio
ns c
o
ncu
rre
ntly
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
502
-47
52
I
ndo
n
e
sian
J Elec Eng
& Com
p
Sci, V
o
l. 10
,
No
.
3
,
Jun
e
2
018
:
10
07
–
1
012
1
008
Recently, the DSU context has attract
ed a lot of i
n
terest.
Pre
v
iously, differe
n
t groups
have ca
rrie
d
o
u
t
m
easu
r
em
e
n
t stu
d
i
es at d
i
fferen
t
lo
cation
s
and
tim
e
[1
-6
]. Resu
lts from th
ese stu
d
i
es a lo
t o
f
wh
ite sp
ace
in
th
e sp
ectrum b
a
n
d
s
i
n
d
i
catin
g
th
at th
e
sp
ectru
m
is u
n
d
e
ru
tilized
. Mo
reo
v
e
r, limited
an
alysis h
a
s b
e
en
conducted
using s
u
c
h
m
eas
urem
ent data. A large
perc
e
n
t
a
ge
of t
h
ese
wo
rks a
r
e f
o
un
de
d o
n
t
h
e
o
ret
i
cal
anal
y
s
i
s
and c
h
an
nel
m
odel
s
. Whi
l
e
a l
a
rge am
ount
o
f
t
h
e
o
ret
i
cal
anal
y
s
i
s
and p
r
ot
oc
ol
s have
been
pr
o
pos
e
d
for sen
s
ing-b
a
sed
app
r
o
a
ch
,
litt
le is kn
own on
t
h
e ap
p
licab
ility o
f
t
h
e sch
e
m
e
s in
reality. Th
e resu
lts
o
f
the
measu
r
em
en
ts d
o
n
e
i
n
th
is
wo
rk
, can
b
e
u
tilized
to
v
a
lid
at
e d
i
fferen
t th
eo
retical an
alysi
s
and
sch
e
m
e
s with
in
th
is research
area. Th
erefore, th
e m
easu
r
e
m
en
t o
f
real n
e
t
w
ork
ac
tiv
ities, p
r
ov
id
es a
v
ital step
to
ward
an
accurate unde
rstandi
ng of DSU. This work
also
provide
s
a real sam
p
le
data of spatia
lly diverse s
p
e
c
tru
m
activity, whic
h can
be
use
d
t
o
e
num
erate the spectrum
utiliza
tion factor and access pa
ttern
of t
h
e
PUs.
In
ad
d
ition
,
th
e
resu
lts can
b
e
u
s
ed
to
ev
al
u
a
te th
e feasib
ility
o
f
d
i
fferen
t
p
r
o
p
o
s
ed
sch
e
m
e
s o
n
ch
ann
e
l prob
ing
and
use
r
det
ect
i
on a
n
d
,
t
o
st
udy
s
p
at
i
a
l
an
d t
e
m
poral
co
r
r
el
at
i
on
bet
w
e
e
n sy
nc
hr
o
n
i
z
ed se
nsi
n
g
uni
t
s
and
acros
s tim
e.
In t
h
i
s
wo
rk
, a
st
udy
o
f
base
d o
n
t
h
e m
easurem
ent
of t
h
e
80
0 M
H
z t
o
2.
4G
Hz f
r
eq
u
e
ncy
ba
nd at
two
d
i
fferen
t
lo
catio
n
s
inside Un
iv
ersiti Tek
n
o
l
og
i
Malaysia (UTM), Jo
hor Bah
r
u
cam
p
u
s
, Malaysi
a
is
prese
n
t
e
d
.
A s
e
t
up co
nsi
s
t
i
n
g
of t
w
o Tekt
r
o
ni
x R
S
A3
0
6
B
[7]
spect
r
u
m
anal
y
zer (S
A)
was use
d
t
o
ca
rry
o
u
t
m
easurem
ent
s
sim
u
l
t
a
neousl
y
at
t
h
e t
w
o
di
f
f
e
rent
l
o
cat
i
o
n
s
f
o
r a
2
4
h
o
u
r
s
pe
ri
o
d
.
Th
e a
n
al
y
s
i
s
co
nd
uc
t
e
d i
s
base
d
on
t
h
e
re
al
spect
r
u
m
dat
a
acq
ui
re
d
fr
o
m
envi
ro
nm
ent i
n
t
h
e
e
xpe
ri
m
e
nt
al
set
u
p. B
u
sy
an
d i
d
l
e
c
h
a
nnel
s
were i
d
en
tified. Th
e ch
ann
e
ls
su
bj
ect
to adj
a
cen
t-ch
ann
e
l interferen
ce
were also
id
en
tified
during
th
e analysis
an
d d
i
scu
ssi
o
n
s. Th
e effect
of th
e d
e
tection th
resho
l
d
u
s
ed
to m
o
n
ito
r ch
ann
e
l activ
ities was also analysed.
Th
e rest of th
e
p
a
p
e
r is
o
r
g
a
n
i
zed
as
fo
llows. A brie
f
d
i
scussio
n
on
in
terferen
ce pro
b
l
em
s related
to
cog
n
itiv
e
radi
o i
s
pre
s
en
t
e
d i
n
Sect
i
o
n
2. Sect
i
o
n
3 p
r
esent
s
t
h
e se
ns
i
ng sy
st
em
and dat
a
col
l
ect
i
on m
e
t
hodol
ogy
. Dat
a
pr
ocessi
ng
an
d
res
u
l
t
s
anal
y
s
i
s
i
s
di
sc
usse
d i
n
Sect
i
o
n
4.
T
h
i
s
wo
r
k
i
s
c
o
nc
l
ude
d i
n
Sect
i
o
n
5.
2.
IN
TER
FE
R
E
N
C
E
PROBLEM
S R
ELA
T
E
D
TO C
OGN
ITIV
E
RADIO
A catego
r
izatio
n
o
f
p
o
t
en
tial in
terferen
ce fro
m
CR en
tit
i
e
s to
th
e PU i
s
d
i
scu
ssed
in
th
is sectio
n.
W
i
t
h
t
h
e i
n
t
r
o
duct
i
o
n
of
C
R
net
w
or
ks,
t
h
er
e are t
w
o
po
ssi
bl
e t
y
pes
of i
n
t
e
rfe
rence
fr
om
C
R
net
w
or
ks.
The
y
are the
interference
from
CR to
pri
m
ary
net
w
or
ks a
n
d
pri
m
ary
net
w
o
r
k
s
t
o
C
R
i
n
t
e
rfe
renc
e.
2.
1.
CR-Prim
a
ry I
n
terfere
nce
The term
interfere
nce tem
p
er
ature (
I
T) lim
it refers
to
th
e “wo
r
st case” in
terfering
situ
atio
n
in
a
specific fre
que
n
cy band and a
t
a precise geogra
phic loca
tion for prim
ary r
eceivers [8,
9]. That is, it represent
s
the m
a
xim
u
m
am
ount
of int
e
rfe
rence
that
the prim
ary
receiver ca
n t
o
lerate. T
h
e
IT
m
odel (ITM) tool i
s
useful in desc
ribing the CR prim
ary
interference. An ideal ITM shoul
d acco
unt for the c
u
m
u
lative RF
energy
fr
om
num
erou
s C
R
t
r
ansm
i
s
si
ons a
n
d set
s
a
m
a
xim
u
m
cap on thei
r a
g
gre
g
ate le
vel
.
C
R
users a
r
e
t
h
en
allo
wed to
u
s
e a frequ
e
n
c
y ban
d
pro
v
i
d
e
d th
eir activ
ities i
n
su
ch
band
will n
o
t
d
i
srup
t t
h
e
IT lim
i
t
s o
f
su
ch
b
a
nd
. Im
p
l
e
m
e
n
tin
g su
ch an
id
eal ITM
u
s
ually n
ecessita
tes real tim
e in
teractio
n
s
b
e
t
w
een CR tran
smit
ters
and prim
ary receivers a
n
d is t
h
ere
f
or
e e
x
tensively rega
rde
d
as
pra
c
tically
im
possible. Hence, se
ve
ral a
d
apte
d
ITM
[
1
0-
11]
h
a
ve
been
rec
o
m
m
e
nded as
m
o
re pract
i
cal
m
odel
s
for
t
h
e
C
R
-
p
r
i
m
ary
int
e
rfe
re
nce rec
e
i
v
ed a
t
prim
ary receivers.
[10], de
fined inte
rfere
n
ce as the
estimated
fraction of PUs
whos
e services
ha
ve bee
n
i
n
t
e
rr
upt
e
d
by
near
by
C
R
t
r
ansm
i
t
t
e
rs. Fact
ors s
u
ch as
C
R
si
gnal
m
o
dul
at
i
o
n, a
n
t
e
n
n
a gai
n
s, a
nd
po
w
e
r
co
n
t
r
o
l
w
e
r
e
co
n
s
i
d
er
ed
i
n
this m
o
d
e
l. Thoug
h, th
is m
o
d
e
l o
n
l
y acco
u
n
t
ed fo
r
t
h
e case wh
er
e th
e
PU ser
v
ices
were i
n
t
e
r
r
u
p
t
e
d by
ju
st
on
e C
R
user and i
t
di
d not
con
s
i
d
er t
h
e
cum
u
l
a
t
i
v
e effect
of several
C
R
tran
sm
issio
n
s
. In
[11
]
, th
e
cu
m
u
la
tive effect of the amassed interf
e
r
e
n
ce power
wa
s considere
d
and a
m
u
l
tifaceted stocha
stic m
odels was built to
describe t
h
e e
x
a
c
t PDF.
2.
2.
Primary-CR I
n
terfere
nce
The i
n
t
e
rfe
re
n
ce fr
om
pri
m
ary
t
o
C
R
net
w
o
r
k
s
ca
n
be
di
rect
l
y
m
easured
by
C
R
re
cei
vers
wi
t
h
passi
ve
se
nsi
n
g t
ech
ni
q
u
es
.
B
a
sed
on
t
h
e
po
we
r s
p
ect
ral
de
nsi
t
y
(PS
D
) o
f
t
h
e i
n
t
e
rf
eri
n
g P
U
si
gn
al
s, t
h
e
spectra
ca
n be gene
rally
categorized
i
n
to three: (i) Black s
p
aces-spectra
band
being use
d
by
high-power PU
signals whic
h has
a high
probability
to
be
detected by the
CR receivers;
(ii)
Grey spaces refe
r to spectra band
whi
c
h has a l
o
w t
o
m
e
di
u
m
powe
r
PU
si
gnal
s
. T
h
ey
are very
w
eak
and as suc
h
m
a
y
not
be d
ecode
d
satisfactorily by the CR recei
vers
. T
h
us, ca
n ca
use a
sign
i
f
icant am
ount
of interfer
e
n
ce
to the
CR
network;
(iii) Wh
ite sp
aces are sp
ectra b
a
nd
s wh
ere
PU sign
als
h
a
v
e
in
sign
ificant p
o
w
er and
can
b
e
referred to
as
background noise. Desc
ribi
ng the
distributions
of wh
ite/gre
y/black s
p
aces
across
fre
que
n
cy, tim
e, and
space
dom
ains are
of great im
port
a
nce
for asse
s
s
ing the i
n
terference
face
d
by CR recei
vers.
To date,
suc
h
a
descri
pt
i
on
has
m
a
i
n
l
y
been s
h
o
w
n em
pi
ri
cal
l
y
by
an am
ount
of m
easure
m
ent
cam
p
ai
gns [
1
-
6
]
,
w
h
i
c
h sh
ow
t
h
at
t
h
e ra
di
o
spect
r
u
m
cons
i
s
t
s
of a hi
gh
perce
n
t
a
ge
of
whi
t
e
spa
ce.
A t
h
e
o
ret
i
cal
m
odel
was re
cent
l
y
Evaluation Warning : The document was created with Spire.PDF for Python.
In
d
onesi
a
n
J
E
l
ec En
g &
C
o
m
p
Sci
ISS
N
:
2
5
0
2
-
47
52
Interfere
n
ce Te
mper
at
ure Me
asureme
n
ts
and
Sp
ectr
u
m
Oc
cupancy… (Paulson N. E
b
ere
c
hukw
u
)
1
009
propose
d
in
[12] to desc
ribe t
h
e sp
atial
distributions of white/grey/black
spaces in t
h
e presence
of a ra
ndom
pri
m
ary
net
w
o
r
k wi
t
h
hom
ogene
o
u
s no
des
.
The
r
e,
i
t
was assu
m
e
d
th
at ev
ery activ
e p
r
im
ary tran
smit
ter
uni
quely de
fines a black s
p
ac
e area and a grey space area.
There is a
widesprea
d
perce
p
tion that blac
k
spaces
are
not
e
x
pl
oi
t
a
bl
e by
C
R
net
w
o
r
k
s
due
t
o
t
h
e
prese
n
ce
o
f
st
ro
ng
i
n
t
e
r
f
eri
n
g
p
r
i
m
ary
si
gnal
s
.
3.
MEASUREMENT METHODOLOGY
Method use
for s
p
ectrum
occupa
ncy m
e
a
s
urem
ent and
an
alysis is crucial in
th
e analysis o
f
th
e
m
easured
dat
a
. Va
ri
o
u
s m
e
tho
d
s
ha
ve
bee
n
di
scusse
d i
n
[1
2]
f
o
r c
o
n
d
u
ct
i
n
g
occ
u
pa
ncy
m
easurem
ent
s
i
n
sp
ectru
m
sen
s
in
g. A 24
h
our
p
e
ri
o
d
was
u
s
ed
each
fo
r
th
e two
l
o
catio
n
s
co
nsid
ered
i
n
Un
i
v
ersiti
Tekn
o
l
o
g
i
M
a
l
a
y
s
i
a
Joho
r cam
pus. The
s
e si
t
e
s have
heavy
wi
rel
e
ss traffics. Th
e
lo
catio
n
s
are th
e Un
iv
ersity PSZ
lib
r
a
r
y
w
ith
p
e
ak
tr
af
f
i
c
p
e
r
i
od
s
du
r
i
ng
t
h
e
m
o
r
n
in
g and
af
ter
noo
n hou
r
s
and
the
K
D
O
J
ho
stel w
ith af
t
e
r
noon
and
ni
ght
pea
k
peri
ods
.
3.
1.
Measurement setup
The Te
kt
r
o
ni
x
R
S
A
3
0
6
B
[7]
SA
was
use
d
fo
r t
h
e
spect
rum
occupancy
measur
em
ent. The
feature
s
and
co
nfi
g
u
r
at
i
on
of
t
h
e R
S
A3
0
6
B
S
A
are
sh
ow
n i
n
Ta
ble 1
.
Th
e SA
o
f
f
e
r
s
ou
tstandin
g
f
eatur
es
with
±
3
ppm
frequency
accuracy and a dynam
i
c
range of
180 dB
m
.
The spect
rum
and spect
rogra
m
display were used
to vie
w
the
received si
gnal powe
r on t
h
e scanned fre
que
ncy a
nd
s
p
ectral dis
p
lay at each s
w
ee
p tim
e
,
respectively.
A sha
r
ed
data
for t
h
e spect
rogram
a
nd s
p
ect
ral
di
s
p
l
a
y
i
s
saved o
n
t
h
e Panas
o
ni
c Fz-G
1
To
ug
h
p
ad
f
o
r
of
fl
i
n
e
pr
ocess
i
ng.
T
h
e c
o
m
p
l
e
t
e
set
up
of
Fi
gu
re
1 i
s
use
d
i
n
t
h
e t
w
o
l
o
ca
t
i
ons
o
v
er t
h
e
ent
i
r
e
spectrum
of
9
kHz
to
6.2
GHz usi
n
g 64001 t
r
ace
poi
nts.
Tabl
e 1.
Features and
C
o
nf
igur
atio
n of
RSA306B
Spectrum
Analyzer [2]
Para
m
e
ter Value
RF fr
equency
r
a
ng
e
9 kHz to 6.
2 GHz
M
easur
e
m
ent range
+20 dB
m
to -
160 dB
m
Fr
equency
accur
a
cy
±3 pp
m
Bandwidth (RBW
/
V
BW
)
Auto
m
a
ticall
y
sele
cted by SA
Sweep ti
m
e
Auto
m
a
ticall
y
sele
cted by SA
Sweep type
Continuous
Refer
e
nce power
-
50 dB
m
Nu
m
b
er
of points
6400
1
Fi
gu
re 1.
RSA306B s
p
ectrum analyzer
with othe
r
access
o
ri
es
4.
DAT
A P
R
O
C
ESSIN
G
AN
D
RES
U
LTS
A
NAL
YSI
S
The 64001 tra
ce poi
nts selected from
the SA allo
w
for power m
easurement
at approxi
m
ately every
97
kHz i
n
terva
l
of the 9
kHz
to 6.2 GHz s
p
ectra s
p
an
. The highe
r
the trace point,
the
better the res
o
luti
on
wh
en
d
e
term
in
in
g
th
e
d
u
t
y cycle in
a b
a
nd
wid
t
h
ran
g
e
. Si
nce it is ex
p
ected
th
at no
ise floo
r
will d
i
ffer across
the spectrum
span, ave
r
a
g
e noise levels we
re take
n for
t
h
e ranges in
question. T
h
ese
noise le
vels are then
use
d
with the
energy detection
(ED)
m
e
thod
[13] to calculate the proba
b
ility of the presence or abse
nce of
signal in a particular band of interest. The ED m
e
thod c
o
m
p
ares the signal receive
d in a given fre
quency
ban
d
t
o
p
r
e
d
efi
n
ed
t
h
re
sh
ol
d
val
u
es
[
5
,
15]
.
A
10
dB
po
we
r ab
o
v
e t
h
e
ave
r
age
n
o
i
s
e l
e
ve
l
i
s
recom
m
ended
i
n
[16] as the
dec
i
sion thres
hol
d. Recei
ved
signal which is
below the
detect
ion thres
hol
d i
n
the
fre
quenc
y
band
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
502
-47
52
I
ndo
n
e
sian
J Elec Eng
& Com
p
Sci, V
o
l. 10
,
No
.
3
,
Jun
e
2
018
:
10
07
–
1
012
1
010
is said
to b
e
idle an
d can b
e
used
in cogn
itive rad
i
o
sc
en
ario
.
Th
e ratio
of th
e
freq
u
e
n
c
y b
i
n
s
wh
ere
en
erg
y
is
d
e
tected
to
the to
tal n
u
m
b
e
r o
f
b
i
n
s
in
the en
tire b
a
ndwid
th
o
f
in
terest is
th
e d
u
t
y cycle wh
ich
sig
n
i
fies
utilized channel. In calcula
ting the
duty cycle (in %) in each of
the
channels, the
total num
ber of the
occu
rre
nce
at
or
ab
o
v
e t
h
e
d
eci
si
on t
h
res
h
o
l
d i
s
di
vi
de
d
b
y
t
h
e t
o
t
a
l
num
ber
o
f
t
h
e
num
ber
t
i
m
e
sl
ot
s.
Thi
s
i
s
sho
w
n i
n
E
qua
t
i
on
(1
).
*100
%
Tslo
t
Ts
lo
t
DD
Dut
y
C
ycle
N
(
1
)
whe
r
e
Ts
l
o
t
DD
is the
num
ber
of tim
e slots
whe
r
e
receive
d si
gna
l powe
r is e
q
ual to
or
great
er tha
n
t
h
e
decisio
n
th
res
h
old a
n
d,
Ts
l
o
t
N
is the
entire tim
e slot
.
Decision threshol
d is the m
a
in
focal point in determ
ining t
h
e
true spectrum
occupa
ncy in the
fre
que
ncy
ban
d
o
f
inte
rest.
This is
beca
use
hig
h
an
d
low t
h
res
h
olds re
sult in
un
de
restim
ation an
d
overestim
at
ion respectively,
of the ch
annel
availability. While too hi
gh t
h
re
sholds m
i
ss
occupied
bands with
a low energy level, extrem
el
y low t
h
reshol
ds are affected
by
noise resulting to fa
lse
unavailability of the
fre
que
ncy
cha
nnel. T
h
ese ar
e sho
w
n in Fi
gu
res 2 a
nd
3
for the tw
o location
s
o
n
U
T
M
Joh
o
r ca
m
pus. In
Figu
re
2,
a
v
er
age
d
u
ty
cy
cle (
ADC
)
of
1
0
0
% a
n
d
0%
is rec
o
r
d
e
d
fo
r l
o
w
d
ecision
th
resh
old
s
(-
10
4
dB
m
to -1
1
0
dB
m
)
and
hig
h
decisi
o
n
thres
h
olds
(-
86
dB
m
to
-70
dBm
)
, respectively in all the
channels. E
x
ce
pt the
W
i
m
a
x and di
gital
m
u
lti
media service (DMS) which
has a
sharp transition in the
ADC
,
the rest of the
bands
have a
g
r
ad
ual
increase i
n
A
D
C
f
r
om
high
to lo
w decisi
on t
h
res
h
olds
.
GSM
90
0 a
n
d
GSM
18
0
0
h
a
ve th
e
slowest tra
n
sition i
n
A
D
C
fr
o
m
high to lo
w
thres
hol
d va
l
u
es in Fig
u
re
3
whe
n
c
o
m
p
ared to ot
her c
h
a
nnels
.
This
variatio
n i
n
the
tw
o Fi
gu
r
e
s co
uld
be
attr
ibut
ed to
different pea
k
pe
ri
od in the t
w
o locations.
Figure 2.
Average duty
cy
cle
as a function of the
decision thres
h
old
for
differe
n
t system
s in UTM PSZ
Lib
r
ar
y:
CDM
A
8
0
0
(
830
-880
MHz)
,
GSM 9
0
0
(8
80
-
9
6
0
M
H
z),
DMS (1
452
-1
492
MHz)
, GSM
1
800
(
171
0-
1
880
MHz)
, IMT-
200
0 (1
915
-22
00
M
H
z),
W
i
m
a
x
(
230
0-
239
0 M
H
z) and
I
S
M
ban
d
(24
50
-24
83
.5
MHz)
Figure 3.
Average duty
cy
cle
as a function of the
decision thres
h
old
for
differe
n
t system
s in UTM
KDOJ
Ho
stel: CDMA
80
0 (83
0-8
80
M
H
z),
GSM
9
0
0
(8
80
-96
0
MHz)
, DMS (14
52-
149
2 MHz)
,
GSM
1
800
(
171
0-
188
0
MHz)
,
I
M
T-
200
0
(1
915
-22
00
MHz)
,
W
i
m
a
x
(
230
0-
239
0 M
H
z) and
I
S
M
ban
d
(
245
0-
248
3.
5 MHz)
M
a
nual selection
o
f
decisi
on
thres
hol
d by
vi
sualiz
ing a
n
d s
e
lecting a m
i
d-poi
nt bet
w
een
the sig
n
al
and
n
o
ise leve
l was use
d
in
[1
6]
. T
h
ese m
e
tho
d
s
have
a
setback
of
real
-tim
e deploy
m
e
nt [
14]
. Si
nc
e the
energy detection
m
e
thod de
pends
on
t
h
e
num
ber
of sa
m
p
les (that is, trace
poi
nts)
and decisi
on thres
h
old,
64001 trac
e points were
use
d
for the whole spectra span
. The fourth colum
n
in Tab
le 2 sho
w
s t
h
e trac
e poi
nt
for
eac
h of
the channel bandwidth.
The
ADC
for the tw
o l
o
cations
using t
h
e E
D
m
e
thod
is
also illustrate
d i
n
Fi
gure
4
usi
n
g di
ffe
rent
threshol
d val
u
es. The
ADC chart form
of the two sites in
Table 2 is shown in
Figure
4.
I
n
all th
e fr
eq
u
e
n
c
y
bands c
onsi
d
ered, the PSZ library has less spectrum
occ
u
p
a
n
c
y w
h
en
co
mp
a
r
ed
to
th
e KDOJ
Hostel location.
This
c
oul
d be attributed
to
m
o
st han
d
h
eld d
e
vices be
ing
tur
n
ed
o
f
f
.
E
x
cep
t fo
r th
e D
M
S
b
a
nd
an
d
W
i
ma
x
in
the PSZ library which ha
s zero
utilized spectrum
s, other bands in
t
h
e locations
have uti
lized spectrums. The
-11
0
-1
05
-1
00
-9
5
-
9
0
-8
5
-
80
-
7
5
-
70
0
20
40
60
80
100
A
verag
e
D
u
t
y
C
ycl
e (%
)
De
cision Th
reshold (
d
Bm
)
CD
MA
800
GSM
9
0
0
DM
S
G
S
M 18
00
IM
T
-
200
0
Wim
a
x
2.4
G
H
z
IS
M
low
t
h
re
shold
high
t
h
re
shold
-1
1
0
-1
0
5
-1
0
0
-9
5
-
9
0
-8
5
-
8
0
-7
5
-
7
0
0
20
40
60
80
10
0
A
ver
age D
u
t
y
C
ycl
e (
%
)
Decis
i
on Thres
hold (dBm)
CDMA
8
0
0
GS
M
9
0
0
DMS
G
S
M
18
00
IMT-
20
00
Wi
m
a
x
2.4 GHz
ISM
lo
w
th
r
e
sh
o
l
d
high
th
r
e
sh
o
l
d
Evaluation Warning : The document was created with Spire.PDF for Python.
In
d
onesi
a
n
J
E
l
ec En
g &
C
o
m
p
Sci
ISS
N
:
2
5
0
2
-
47
52
Interfere
n
ce Te
mper
at
ure Me
asureme
n
ts
and
Sp
ectr
u
m
Oc
cupancy… (Paulson N. E
b
ere
c
hukw
u
)
1
011
t
h
ree hi
ghest
s
p
ect
r
u
m
occup
a
ncy
obse
r
ved
are i
n
t
h
e GS
M
180
0, I
S
M
2.
4 G
H
z and
GSM
9
00 i
n
t
h
at
orde
r
for th
e KDOJ Ho
stel.
A si
milar tren
d is
o
b
s
erv
e
d in
t
h
e PSZ
libra
ry
location. It ca
n
be sai
d
that
a non-
u
n
i
form
sp
ectru
m
u
tilizat
io
n
i
s
ob
tain
ed
from
th
e m
easu
r
emen
t resu
lts.
Tabl
e 2.
Avera
g
e
duty cycle for differe
n
t
ba
nds
Services
Fr
equency
range
(MHz
)
Bandwidth
(MHz
)
No. of
trace
points in
bandwidth
Average duty c
y
cl
e (%)
L
o
cation
PSZ Librar
y
KDOJ
Hostel
CDM
A
800
830 – 88
0
50
517
9.
86
13.
73
GSM
900
880 – 96
0
80
827
14.
75
45.
71
L-Band Digital
Multi
m
e
dia Servic
e
(DMS
)
1452 – 1
492
40
414
0
0
GSM
1800
1710 -
188
0
170
1756
21.
81
66.
74
I
M
T-
2000
1915 – 2
200
285
2943
9.
41
31.
36
W
i
m
a
x
2300 – 2
390
90
930
0
32.
26
I
S
M
band 2.
4
2450 – 2
483.
5
33.
5
399
21.
30
62.
82
Fi
gu
re
4.
A
v
er
age S
p
ect
r
u
m
Occu
pa
ncy
by
Fre
que
ncy
B
a
n
d
s
The P
S
D
w
h
i
c
h i
s
av
era
g
e
d
f
o
r
t
h
e
24
ho
u
r
s m
easure
m
ent
peri
o
d
i
s
sh
ow
n i
n
Fi
gu
re
5. T
h
e
fre
que
ncy
occ
upa
ncy
i
n
t
h
e
s
e fi
g
u
r
es ca
n
easi
l
y
be c
o
r
r
e
l
at
ed wi
t
h
t
h
e
ADC
i
n
Fi
g
u
r
e
4
f
o
r
t
h
e
sel
ect
ed
b
a
nd
s of
in
ter
e
st.
Fi
gu
re
5.
24
h
o
u
r
s a
v
e
r
age
d
po
we
r s
p
ect
ral
den
s
i
t
y
(PS
D
)
fo
r
fre
que
ncy
r
a
nge
8
3
0
M
H
z
t
o
2.
48
3
5
M
H
z:
(a)
PSZ lib
rary
(b
)
K
DOJ
H
o
stel
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN:
2
502
-47
52
I
ndo
n
e
sian
J Elec Eng
& Com
p
Sci,
Vo
l. 10
,
No
.
3
,
Jun
e
2
018
:
10
07
–
1
012
1
012
5.
CO
NCL
USI
O
N
Eve
n
th
ou
g
h
several s
p
ectr
u
m
m
easurem
ent cam
paigns
have
bee
n
c
o
m
p
leted in th
e conte
x
t o
f
cognitive radio, there is a defici
ency of m
u
tual and appropriat
e evaluation methodology. These
m
easurem
ents are use
f
ul in investiga
tin
g iss
u
es
of s
p
ectr
u
m
sensing
f
o
r
DS
U, incl
udi
n
g
P
U
sig
n
al de
tection,
adj
ace
nt c
h
annel interfere
n
c
e
, receive
r se
nsitivity,
and
policy pe
rform
a
nce with l
o
cal and c
o
operative
sensin
g.
T
h
is
wo
rk
has
pre
s
e
n
ted
an
am
ple and
in
-de
p
th
d
i
scussio
n
of
se
veral im
porta
n
t
aspects t
h
at
n
eed t
o
be ca
refully taken int
o
acc
ount whe
n
assessing spe
c
tr
um
occupa
ncy so
as to a
v
oid inaccurate
results a
nd
properly cha
r
a
c
terize the activ
ity
of P
U
net
w
o
r
k
s
. M
a
ny
i
ssues a
r
e y
e
t
to be i
nve
stigated,
the
s
e includes the
spatial correlat
i
on am
ong se
n
s
ing
devices at
diver
s
e dist
an
ces and
h
o
w t
o
ch
o
o
se a sui
t
able thres
hol
d
that
differentiates low-power
activities from
noi
se.
REFERE
NC
ES
[1]
F. H. Sanders,
"
B
roadband spectrum su
rveys in Denver, CO, San Di
ego, CA, and Los Angeles, CA: Methodolog
y
,
analysis, and comparative results,"
i
n
Proc
.
IEE
E
Int
e
rna
t
i
o
na
l Sy
mposi
u
m on E
l
e
c
t
r
oma
gne
tic
Compa
t
i
b
i
lity
(EMC 1998), vo
l. 2
,
Aug
.
1998
,
pp. 988-993
.
[2]
M. A. McHen
r
y
et al.,
"Spectru
m occupan
c
y
measurem
ents," S
h
ared Spectrum
Compan
y
,
Tech. Rep., Jan
2004
-
Aug 2005,
available at: h
ttp://w
ww.
sharedspectr
u
m.com/measurements.
[3]
A. Petrin and
P. G.
Ste
ffe
s,
"
A
n
a
lysis and comparison of spectrum meas
urements performed in urban and rural
areas to determine th
e to
tal am
ount of spectru
m
usage,"
in
Proc. In
tern
ation
a
l S
y
mposium on Advanced
Radio
Techno
logies (I
SART
2005),
M
a
r. 2005
, pp
. 9-1
2
.
[4]
R. I. C. Ch
iang,
G. B.
Rowe,
and K.
W.
Sowerby
,
"A quantitative
analysis of spec
tral occupan
cy m
e
asurements for
cognitive radio
,
" in Proc. IEEE 65th Vehicular Tec
hnolog
y
Conf
erence (VTC 2007 Spring),
Apr. 2007, pp. 3016-
3020.
[5]
P
.
E. N
., B
.
A
d
e
d
eji K
,
K
.
M
.
Y
.
,
P
.
J
.
J
.
, J
.
B
.
D
i
n, and S
.
K
.
S
.
Y
.
,
"
Spectrum Occupancy M
e
as
urement: A Case for
Cognitive Radio
Network
in Lagos, Nigeria,"
ARPN
Journal of Engineer
ing
and
Applied Sciences, vo
l. 12,
pp.
951-955, Febru
a
r
y
2017.
[6]
M. H. Islam
e
t
al.,
"
Spe
ctrum survey in
Sing
apor
e: Occupan
cy measurement
s
and analyses,"
in Proc. 3rd
Interna
tiona
l Co
nferenc
e
on Co
gnitive
Radio
Oriented
W
i
rele
ss Networks and Com
m
unicati
ons (CrownCom
2008),
May
200
8, pp
. 1-7
.
[7]
Tektron
i
x. (20
16, 5th May
)
.
RSA306
B U
S
B Rea
l
T
i
m
e
Spec
t
rum A
nalyzer Datasheet
. Avai
labl
e:
http://www.
te
k.com/da
ta
she
e
t/rsa306b-usb-re
a
l
-time
-spe
c
t
rum-a
n
a
l
y
z
e
r-0
[8]
J.Mitola, III
, “C
ognitive rad
i
o,”
Li
centi
a
te th
esis, Ro
y
a
l
Institu
te
of
Technolog
y
,
Stockholm
,
Sweden, 1999
[9]
A. Ghas
em
i and E. S
.
S
ous
a,
“Capacity of fad
i
ng channels under
spectrum sha
r
ing constraints,”
in P
r
oc. IEEE
ICC06, Istanbu
l, Turkey
, June 20
06, pp
. 4373-43
78.
[10]
T. K. Phan, S. A. Voroby
ov
, N.
D. Sidiropoulos, and C. Tellambura, “Sp
ectrum sharing in wireless networks via
QoS-aware seco
ndar
y
m
u
ltic
ast
beam
form
ing,
”
IEEE T
r
ans. Sig
.
Process.
,
vol.
57, no. 6
,
pp. 2
323–2335, June
2009.
[11]
X. Hong, C
.
-X.
Wang, H.-H. Chen,
and Y. Zhan
g,
“Secondar
y
s
p
ectrum access netw
orks: recent development
o
n
the spa
tia
l m
ode
ls,”
I
E
EE Veh. Technol. Mag
.
,
vo
l. 4
,
no
. 2
,
pp
. 36
–43, June 2009.
[12]
R.
J. Ma
the
s
on,
"
S
trategies
for Spectrum usage Measurements,"
in
IEEE International S
y
mposium on
Electromagnetic Compatib
ility
, 1
988, pp
. 235-24
1.
[13]
R. A.
W
itte
,
Sp
ectrum and n
e
two
r
k measurements
: Scitech
Publ., 2
014.
[14] M.
López
‐
Benítez and F. Casad
e
vall, "Methodo
logical Aspects o
f
Spectrum Occ
upancy
Ev
aluation in the Con
t
ex
t
of Cognitive R
a
dio,"
Europ
e
an Transactions
on Telecommunica
tions,
vol. 21
, pp
. 680-693, 2010.
[15]
R. Bureau, "Spectrum Monitorin
g
," ed
: In
ternatio
nal Telecommunica
tion Union
(I
TU), 2011
.
[16]
M. A. McHenr
y
and
K. S
t
eadm
a
n, "S
pectrum
Occupanc
y M
eas
ur
em
ents
,"
Shared Spectrum Comp
any Report,
pp
.
1-55, 2005
.
Evaluation Warning : The document was created with Spire.PDF for Python.