Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
11
,
No.
1
,
Febr
uar
y
2021
, pp.
52
8
~
535
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v11
i
1
.
pp
528
-
535
528
Journ
al h
om
e
page
:
http:
//
ij
ece.i
aesc
or
e.c
om
Fin
gerpri
nt
p
ositioni
ng of
users d
evic
es
in
l
ong ter
m evolut
ion
ce
llu
lar
network
usin
g
K n
earest n
eighbou
r
a
l
gorith
m
Moham
med
J. Alh
asan
1
,
S
ar
mad M
uneer
A
b
dulhus
sei
n
2
,
Ali H
.
K
.
K
hwayyir
3
1
Depa
rtment of
Com
m
unic
at
ion Te
chn
ique
s E
ngi
nee
ring
,
Engi
ne
e
ring
T
ec
hni
cal
C
oll
eg
e/
Naj
af,
Al
-
Furat
Al
-
Aws
at
T
ec
hni
cal
Un
ive
rsit
y
,
Ir
aq
2,3
Depa
rt
m
ent of
om
m
unic
at
ion
,
Te
chn
ic
a
l
Inst
it
u
te
/
Na
ja
f, Al
-
Fur
at
Al
-
Aw
sat
Tec
hnic
a
l
Univer
si
t
y
,
Ir
aq
Art
ic
le
In
f
o
ABSTR
A
C
T
Art
ic
le
history:
Re
cei
ved
A
pr
25
, 20
20
Re
vised
Ju
l
8
,
20
20
Accepte
d
Aug 16
, 20
20
The
rap
id
expo
nent
i
al
growth
in
wire
le
ss
te
ch
nologi
es
and
th
e
nee
d
fo
r
publi
c
safe
t
y
h
as
le
d
to
inc
re
asing
demand
for
loc
at
ion
-
base
d
servic
es.
Te
rre
str
ia
l
cellu
la
r
ne
tworks
can
offe
r
a
cc
ep
table
positi
on
est
imati
on
for
users
tha
t
c
an
m
ee
t
the
st
a
tut
or
y
req
u
ire
m
ent
s
set
b
y
t
he
Feder
a
l
Com
m
unic
at
ions
Comm
ission
in
ca
se
of
ne
t
work
-
base
d
positi
oning
,
for
safe
t
y
reg
ul
at
io
ns.
In
thi
s
stud
y
,
the
proposed
rad
io
fre
qu
e
nc
y
p
at
t
ern
m
at
chi
ng
(RFP
M)
m
et
hod
is
implemente
d
and
te
sted
to
det
er
m
ine
a
user’s
loc
a
ti
on
eff
e
ctiv
ely
.
Th
e
RF
PM
m
et
hod
has
bee
n
te
sted
and
valid
at
ed
in
two
diffe
ren
t
environ
m
ent
.
The
evalu
at
ions
show
remarka
bl
e
result
s
e
spec
iall
y
in
the
Micro
cell
sce
nar
io
,
at
67%
of
positi
oning
err
or
15m
and
at
9
0%
31.
78m
for
Micro
ce
ll
s
ce
nar
io,
with
re
sults
of
75.
66
m
at
67%
and
141.
4m
at
90%
for
Mac
ro
cell
sc
ena
rio
.
Ke
yw
or
d
s
:
Fing
e
r
pr
int
Locati
on
-
base
d
se
rv
ic
es
LTE
l
ocali
zat
ion
Po
sit
io
ning
RFPM
This
is an
open
acc
ess arti
cl
e
un
der
the
CC
B
Y
-
SA
l
ic
ense
.
Corres
pond
in
g
Aut
h
or
:
Moh
am
m
ed
J. Alhasa
n
,
Dep
a
rtm
ent o
f C
omm
un
ic
at
io
n
Tec
hniq
ues En
gin
eeri
ng,
En
gin
eeri
ng T
echn
ic
al
C
ollege/
Na
j
af
,
Al
-
F
ur
at
Al
-
A
wsat Tec
hn
ic
al
Unive
rsity
,
54003,
Kufa,
Naj
a
f G
ov
e
rn
a
rate, Ira
q
.
Em
a
il
:
m
oh
a
m
m
edj
.alh
asan
@atu.e
du.iq
1.
INTROD
U
CTION
The
posit
ion
i
ng
has
been
c
onside
red
a
n
opti
on
al
feat
ur
e
in
the
sta
nd
a
rd
iz
at
io
n,
im
pl
e
m
entat
ion
,
and
ex
plo
it
at
io
n
of
e
xistent
c
el
lular
net
wor
ks
.
Ne
ver
t
heless,
the
big
cel
lular
com
m
un
ic
at
ion
in
fr
a
str
uctu
re
distrib
uted
a
r
ound
t
he
world
can
sti
ll
be
reused
for
lo
cal
iz
at
ion
pur
po
se
s,
co
ntributi
ng
a
n
a
dd
e
d
a
dvant
age
t
o
netw
ork
m
anag
em
ent
and
se
rv
ic
es
[1
]
.
I
n
pr
i
ncipal,
l
ocati
on
base
d
se
rvi
ces
ha
ve
be
en
dr
i
ven
by
t
wo
m
a
i
n
dem
and
s:
com
m
ercial
pu
r
po
s
es
and
em
erg
e
ncy
serv
ic
es
.
Com
m
ercial
us
ers
re
qu
i
re
accurate
an
d
res
pons
i
ve
locat
ion
pe
rform
ance,
f
or
pu
rposes
su
c
h
as
locat
ion
ba
se
d
a
dv
e
rtisi
ng
and
m
aps.
I
n
e
m
er
gen
cy
se
r
vices,
the
m
os
t
sign
i
ficant
s
ponsor
is
the
FCC
’s
E9
11
a
utho
rizat
ion
in
the
U
SA
,
w
hich
de
m
and
s
the
loc
at
ion
of
e
m
erg
ency
cal
le
rs
to
be
delivere
d
within
certi
fied
acc
ur
acy
lim
i
ts.
I
n
orde
r
t
o
m
eet
FCC
re
quire
m
ents,
sever
al
tec
hniq
ues ha
ve been
im
pr
ov
e
d
f
or th
e pur
poses
of posit
ion
i
ng [1
,
2].
The
FCC
requirem
ents
of
po
sit
io
ning
ac
cur
aci
es
are
50m
(6
7%
)
a
nd
150m
(90%
)
in
case
of
te
rm
inal
-
base
d
syst
e
m
s,
and
100m
(6
7%)
and
30
0m
(
90%)
in
case
of
net
work
-
ba
sed
syst
em
s
[1
,
2].
Wh
il
st
sat
el
li
te
te
chnolo
gy
ha
s
dem
on
strat
ed
it
can
m
eet
these
requirem
ents,
delive
rin
g
ver
y
go
od
loc
at
ion
e
s
t
i
m
a
t
i
o
n
i
n
o
p
e
n
e
n
v
i
r
o
n
m
e
n
t
s
.
I
n
u
r
b
a
n
a
n
d
i
n
d
o
o
r
e
n
v
i
r
o
n
m
e
n
t
s
,
h
o
w
e
v
e
r
,
t
h
e
p
e
r
f
o
r
m
a
n
c
e
o
f
p
o
s
i
t
i
o
n
i
n
g
c
a
n
b
e
v
e
r
y
p
o
o
r
d
u
e
t
o
b
l
o
c
k
i
n
g
o
f
s
a
t
e
l
l
i
t
e
s
i
g
n
a
l
s
b
y
b
u
i
l
d
i
n
g
s
a
s
w
e
l
l
a
s
m
u
l
t
i
p
a
t
h
prop
a
gatio
n.
Of
te
n,
posit
io
ning
with satel
li
te
syst
e
m
s in
these
envir
on
m
ents can be im
po
ssible [
3
,
4]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Fing
er
pr
i
nt po
sit
ion
in
g of u
se
rs d
evi
ces
in
l
ong t
erm
evol
ution
...
(
Mo
ham
med J.
Alh
as
an
)
529
Du
e
t
o
the
re
quirem
ents
for
public
safety
,
po
sit
io
ning
te
c
hnology
it
sel
f
is
one
of
the
vit
al
m
at
te
rs
to
be
ta
ckle
d.
Se
ver
al
syst
em
s
for
locat
io
n
de
te
rm
inati
on
i
n
LTE
net
wor
ks
hav
e
bee
n
i
m
pr
oved
t
o
ve
rify
a
us
e
r’
s
posit
ion
unde
r
s
pec
ifie
d
en
vir
onm
ents.
T
hi
s
unde
rlines
need
a
nd
i
nclu
des
se
arch
i
ng
f
or
t
he
best
avail
able te
ch
ni
qu
e t
hat c
ou
l
d be a
ppli
ed
t
o determ
ine the locati
on of a
u
s
er.
W
it
h
the
em
erg
ence of
LTE
, n
ew
e
ffo
rts
ap
pear
t
o
f
oc
us
on
em
po
we
rin
g
E9
11
a
nd
LBS
on
t
hese 4G
syst
e
m
s,
throu
gh
the
offe
rin
g
of
a
seam
less
co
nversi
on
betwee
n
2G/3G
an
d
LTE
po
sit
io
ning
se
rv
ic
e
s
.
Ce
ll
ular
netw
orks
ca
n
prov
i
de
an
acce
pta
ble
po
sit
io
ning
est
i
m
ation
that
sat
isfie
s
the
FCC
requirem
ents.
E
s
p
e
c
i
a
l
l
y
,
w
i
t
h
t
h
e
e
m
e
r
g
e
n
c
e
o
f
t
h
e
3
r
d
g
e
n
e
r
a
t
i
o
n
p
a
r
t
n
e
r
s
h
i
p
p
r
o
j
e
c
t
(
3
G
P
P
)
a
n
d
l
o
n
g
t
e
r
m
e
v
o
l
u
t
i
o
n
(
L
T
E
)
,
t
h
e
s
e
c
a
n
a
c
h
i
e
v
e
g
o
o
d
a
c
c
u
r
a
c
y
p
r
o
v
i
d
i
n
g
e
x
c
e
l
l
e
n
t
c
o
v
e
r
a
g
e
,
p
a
r
t
i
c
u
l
a
r
l
y
i
n
u
r
b
a
n
a
n
d
i
n
d
o
o
r
e
n
v
i
r
o
n
m
e
n
t
s
[
4
]
.
Ma
ny
cel
lular
netw
ork
posit
ion
in
g
te
ch
ni
qu
e
s
ta
ke
n
int
o
co
ns
i
der
at
io
n
in
Re
le
ase
9
of
(3GPP)
do
c
um
ent
su
c
h
as
obse
rv
e
d
tim
e
diff
ere
nc
e
of
a
rr
iv
al
(
O
TDOA
)
a
nd
e
nh
a
nce
d
cel
l
I
D
(E
-
C
ID).
Ba
sed
on
the
li
te
ratur
e
su
ch
as
[
5
,
6
]
,
a
m
on
g
oth
er
s,
the
highest
accuracy
can
be
pro
vid
e
d
by
global
po
sit
io
ning
syst
e
m
(G
P
S)
es
pecial
ly
,
in
li
ne
of
sig
ht
an
d
f
re
e
sp
ace
en
vir
on
m
ents.
H
ow
ever,
the
pe
r
f
or
m
ance
of
G
PS
is
sign
ific
a
ntly
r
e
du
ce
d
i
n dense
urb
a
n
a
nd in
door e
nviro
nm
e
nts.
Ce
ll
iden
ti
fication
(
C
ID) posit
ion
in
g,
is c
on
si
der
e
d
as a n
et
work
base
d
m
e
thod can
als
o
be
e
m
plo
ye
d
to
est
i
m
at
e
the
locat
ion
of
m
ob
il
e
sta
ti
on,
howe
ve
r,
with
v
ery
lim
i
te
d
accuracy.
The
ea
s
ie
st
case
exam
ple
is
wh
e
re
t
he
po
si
ti
on
of
the
m
ob
il
e
sta
ti
on
is
est
i
m
at
ed
to
be
the
l
ocati
on
of
the
base
sta
ti
on
[
7
]
.
I
n
cel
l
ID,
po
sit
io
ning
pe
r
form
ance
can
be
en
han
ce
d
by
m
easur
ing
s
pecific
netw
ork
char
act
e
risti
cs,
a
te
chn
iq
ue
known
as
enh
a
nce
d
cel
l
ID
(ECID).
A
cel
lular
network
ca
n
util
ise
the
ang
le
of
arr
ival
of
a
sig
nal
from
the
m
ob
il
e
sta
ti
on
to
deliv
er
directi
onal
i
nfor
m
at
ion
,
t
he
resea
rch
er
in
[
8
]
s
ta
te
d
t
hat
the
acc
ur
acy
is
im
pr
ov
e
d
es
pe
ci
al
l
y
in
LTE
r
ur
al
a
r
eas.
T
he
d
ra
w
ba
ck
with
this
m
et
ho
d
is
the
r
equ
i
rem
ent
fo
r
extra
ha
rdwa
r
e
an
d
hen
ce
it
can
be
exp
e
ns
i
ve.
T
he
auth
or
s
in
[
9
]
declare
d
in
the
ir
exp
e
rim
ents
that
the
accura
cy
of
Ce
ll
-
ID
i
s
no
t
sat
isfact
ory
as
a g
e
ner
al
s
olu
ti
on.
In
a
dd
it
io
n,
th
e
round
trip
ti
m
e
(RTT)
ca
n
be
us
e
d
to
est
i
m
a
te
the
dist
ance
betwe
en
the
m
ob
il
e
sta
ti
on
a
nd
bas
e stat
ion
as
ar
gued
in [
10
]
,
h
oweve
r,
due to
path
l
os
s a
nd s
hado
w
fa
ding
eff
ect
, t
he
acc
uracy
of
this
m
et
ho
d
ca
n
be
li
m
it
ed.
Anothe
r
posit
ion
i
ng
m
et
ho
d
can
be
us
e
d
in
LTE
syst
em
wh
ic
h
a
re
AoA
a
nd
ob
s
er
ved
ti
m
e
diff
e
ren
ce
of
arr
ival
(OTD
OA)
te
chn
i
qu
es.
The
A
oA
determ
ines
t
he
directi
on
of
rad
i
o
fr
e
qu
e
ncy
wa
ve
prop
a
gatio
n
and
i
nvolv
e
a
n
anten
na
arr
ay
at
the
side
of
t
he
inc
om
ing
wav
e
(
netw
ork
s
ide).
This
te
ch
nique
works
pa
rtic
ul
arly
well
under
conditi
ons
of
the
li
ne
-
of
-
sig
ht
(Lo
S
)
[
1
1
]
.
L
it
eratur
e
s
hows
that
there
are
m
any
eff
or
ts
ha
ve
been
ac
hieve
d
on
OT
DOA
posit
ion
i
ng
te
ch
no
l
og
y
in
LTE
syst
e
m
.
Fo
r
T
DoA
appr
oach
es
,
ti
m
e
diff
ere
nces
are
m
easur
ed
at
the
receiving
base
sta
ti
on
s
betw
een
th
e
tim
es
of
fligh
t
of
diff
e
re
nt r
a
dio
sign
al
s; t
his
is
util
iz
ed
in,
e.
g., L
oRa [1
2
,
13]
.
Am
plit
ud
e
-
bas
ed
te
chn
i
qu
e
s
conve
rt
the
ob
t
ai
ned
sig
nal
intensit
y
to
a
dis
ta
nce
base
d
on
a
path
los
s
(P
L
)
m
od
el
f
or
distance
c
onve
rsion;
ho
wev
e
r,
the
co
ns
i
dered
set
ti
ng
dem
and
s
that
a
n
ef
fecti
ve
PL
m
od
el
be
accounte
d
f
or. K
no
wing
the net
work
t
opolog
y,
i.e., p
re
dicti
ng
t
he
distance
betwee
n
a
m
ob
il
e
us
e
r
an
d
a set
of
base s
ta
ti
ons
re
du
ce
s the
locat
ion
of all
sig
na
l par
am
et
ers
to
a trian
gula
ti
on or m
ulti
laterati
on
pro
blem
[
14
]
.
A
novel
OTDOA
posit
ion
in
g
schem
e
in
heter
og
e
ne
ou
s
LTE
-
a
dv
a
nce
d
syst
em
s
is
carried
out
.
This
schem
e
avo
i
ds
inter
fer
e
nce
w
hich
ca
n
gr
eat
ly
enh
a
nc
e
the
po
sit
io
ni
ng
acc
ur
acy
[
15]
.
In
[
1
6
]
the
stud
y
is
su
ggest
e
d
a
posit
ion
i
ng
te
c
hn
i
qu
e
dep
e
ndent
on
relat
ive
tim
e
diff
eren
ce
of
ar
riv
al
(RTDOA)
wh
e
re
m
e
a
s
u
r
e
m
e
n
t
s
w
e
r
e
e
m
p
l
o
y
e
d
i
n
a
f
e
m
t
o
c
e
l
l
s
c
l
u
s
t
e
r
.
H
o
w
e
v
e
r
,
t
h
e
n
e
e
d
s
f
o
r
s
y
n
c
h
r
o
n
i
z
a
t
i
o
n
o
f
t
i
m
e
a
n
d
a
t
l
e
a
s
t
t
h
r
e
e
b
a
s
e
s
t
a
t
i
o
n
s
t
o
b
e
a
v
a
i
l
a
b
l
e
a
r
e
t
h
e
m
a
i
n
d
r
a
w
b
a
c
k
s
o
f
t
h
i
s
t
e
c
h
n
i
q
u
e
,
a
l
t
h
o
u
g
h
i
t
p
r
o
v
i
d
e
s
g
o
o
d
a
c
c
u
r
a
c
y
.
The
Fin
gerp
rinting
po
sit
io
ni
ng
te
ch
nique
is
cur
re
ntly
a
ver
y
act
ive
top
ic
that
does
no
t
require
add
it
io
nal
hard
war
e
or
s
of
t
wa
re
in
-
net
wor
k
or
m
ob
il
e
sta
tio
n.
T
his
m
et
hod
is
discu
ssed
in
the
3GP
P
m
e
et
ing
unde
r
the
na
m
e
ra
dio
freq
ue
nc
y
patte
rn
m
atch
in
g
(RFPM)
to
be
dep
l
oyed
in
L
TE
releas
e
12
[1
7
]
.
I
n
[1
8
,
1
9
]
,
the
auth
ors
we
re
disc
us
sed
i
n
their
stu
dy
the
pri
nciple
of
fing
e
r
pr
inti
ng
po
sit
io
ning
m
et
hod
in
L
T
E
u
s
i
n
g
C
R
L
B
a
n
d
m
a
x
i
m
u
m
l
i
k
e
l
i
h
o
o
d
a
l
g
o
r
i
t
h
m
s
,
b
u
t
t
h
e
i
r
r
e
s
e
a
r
c
h
w
a
s
c
a
r
r
i
e
d
o
u
t
i
n
a
n
u
r
b
a
n
e
n
v
i
r
o
n
m
e
n
t
.
The
pro
po
s
ed
m
et
ho
d
i
n
this
pa
per
highli
ghts
fi
ng
e
r
pr
inti
ng
te
ch
nique
a
nd
will
be
im
p
lem
ented
t
o
fin
d
us
e
r’
s
pos
it
ion
in
LTE
ne
tworks
base
d
on
the
the
oret
ic
al
app
r
oac
h.
Firstl
y,
our
pro
po
s
ed
m
et
ho
d
will
be
perform
ed
us
i
ng
pro
gr
am
-
ba
sed
-
c
om
pu
t
er.
The
c
od
e
of
th
e
sugg
est
e
d
m
et
hod
will
be
wr
it
te
n
a
nd
ev
al
uate
d
in MAT
LAB e
nv
i
ronm
ent accordin
g
t
o
the
KNN
al
gorith
m
b
ased on
m
at
hem
a
ti
cal
eq
ua
ti
on
s.
Seco
nd
ly
,
data
colle
ct
ed
fro
m
ou
r
s
uggest
ed
m
et
ho
d
is
ba
sed
on
pat
h
l
os
s
pro
pag
at
i
on
a
nd
sh
a
dow
fad
i
ng
e
ff
ect
s
on
t
he
sig
na
l
transm
i
tt
ed.
Finall
y,
the
pro
posed
m
et
ho
d
will
be
te
ste
d
in
tw
o
di
ff
ere
nt
scenari
os
:
Ma
cro
cel
l
an
d
Mi
cro
cel
l
en
vi
ronm
ents,
unde
r
dif
fere
nt
c
onditi
ons.
This
arti
cl
e
is
arr
an
ged
a
s
fo
ll
ows:
the
m
otivati
on
a
nd
the
relat
ed
wor
ks
wer
e
d
e
scri
bed
in
Sect
io
n
1
,
w
hile
the
s
yst
e
m
m
a
the
m
at
ic
al
m
od
el
was
dr
i
ven
i
n
Sect
io
n
2
,
n
ext,
the
m
et
hodo
l
og
y
is
cl
arified
in
S
ect
ion
s
3
,
t
he
res
ul
ts
of
the
sim
ul
at
ion
are
descr
i
bed a
nd d
isc
us
se
d
i
n Sect
ion 4
, a
nd last
ly
, co
ncl
usi
on
s
are
sp
eci
fi
ed
in
Secti
on
5
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
11
, No
.
1,
Febr
uar
y
2021
:
52
8
-
53
5
530
2.
S
YS
TE
M MO
DEL
2.1.
S
ystem
c
onfig
urati
on
Creat
ing
t
he
fin
gerpr
i
nts
database
by
co
nv
e
ntio
nal
m
e
tho
d
a
re
relat
ively
sim
ple
pr
ocess.
The
receive
d
si
gn
al
(RS)
m
easur
em
ents
are
achieve
d
by
th
e
receive
r
at
ea
ch
re
fer
e
nce
point,
a
nd
after
certai
n
processi
ng
the
data
is
stored
in
the
datab
ase.
General
ly
,
the
m
or
e
refe
ren
ce
po
i
nts
that
are
sel
ected
i
n
the
c
ollec
te
d
or
trai
ning
ph
a
s
e,
the
bet
te
r
es
tim
a
ti
on
acc
uracy
can
be
acc
om
plished
in
t
he
posit
ion
in
g
ph
a
se
because
of
fin
ge
rprints
databa
se can
b
e
acq
ui
red m
or
e accu
r
a
te
ly
[
20
].
Com
pu
ti
ng
the
locat
ion
of
th
e
m
ob
il
e
sta
ti
on
duri
ng
t
he
po
sit
io
ning
phase
require
s
a
n
al
gorithm
.
Ther
e
a
re
sev
eral
a
lgo
rit
hms
in
order
to
est
i
m
at
e
the
us
er;
in
this
w
ork
the
K
nea
rest
neig
hbour
(K
N
N
)
al
gorithm
wil
l
be
us
e
d.
[
21
]
i
n
their
inv
e
sti
ga
ti
on
sta
te
d
th
at
the
distance
between
the
m
easur
ed
RSS
vector
[s
1
s
2
s
n
]
an
d
t
he
RSS
vect
or
in
the
databas
e
[S
1
S
2
S
n
]
can
be
com
pu
te
d.
The
ge
ner
al
iz
ed
distance
betwee
n
two vecto
rs
ca
n be calc
ulate
d by the
(
1)
:
=
(
∑
|
−
|
=
1
)
1
⁄
(1)
w
he
re
q=
1
is
r
efer
red to Ma
nhat
ta
n dist
ance
and E
uclidea
n dist
ance
whe
n q=
2.
The
nea
rest
nei
ghbour
al
gorithm
can
be
de
fi
ned
a
s
t
he
po
i
nt
with
t
he
s
hort
est
distance
the
sig
nal
can
ta
ke
[
20
]
.
I
n
t
he
K
Nea
rest
Neig
hbours
,
t
he
a
ver
a
ges
of
K
points
c
oor
din
at
e
ca
n
be
util
iz
e
to
es
tim
a
t
e
the
locat
ion
of
m
ob
il
e
sta
ti
on
.
This
al
gorith
m
pr
ov
i
des
bet
te
r
est
i
m
a
ti
on
than
Near
est
N
ei
ghbours
al
gorithm
because
there
i
s
no
reas
on
t
o
sel
ect
on
ly
the
near
est
value
and
i
gnor
e
othe
r
nea
rb
y
valu
es.
Othe
r
al
gor
it
h
m
s
su
c
h
as
the
s
m
al
le
st
po
ly
gon
[
2
2
]
and
ne
ur
al
netw
orks
[2
3
]
are
ei
the
r
com
plica
te
d
or
do
not
nec
essaril
y
i
m
pr
ove the
ac
cur
acy
of the l
ocat
ion est
im
a
t
ion
.
In
order
to
c
ollec
t
database,
pro
pa
gation
m
od
el
s
can
be
util
iz
ed
to
ge
ner
at
e
arti
fici
al
fingerprints
f
or
the
RF
Patt
er
n
Ma
tc
hin
g
m
et
ho
d.
Ba
se
d
on
c
ertai
n
pr
op
a
gat
ion
m
od
el
the
m
easur
em
ents
can
be
cal
culat
ed
to
colle
ct
the
fing
er
pr
i
nts
data
b
ase.
Ge
ne
rall
y,
these
fing
e
rprints
that
ge
ner
at
e
d
arti
fici
al
ly
can
be
cal
le
d
syntheti
c
fin
ge
rprints
[
2
4
]
.
T
he
set
of
recei
ved
si
gnal
(RS
)
m
ap
fo
r
th
e
a
rea
un
der
st
udy
in
this
wor
k
will
be
base
d
on
COS
T
23
1
Wall
fisch
-
Ik
e
gam
i
m
od
el
. W
al
lfisc
h
-
Ik
e
gam
i
m
od
el
i
nclu
des
so
m
e
determ
inist
ic
aspects
to the em
pirical
m
od
el
s; t
hus, t
o
cha
racteri
ze
the pr
op
a
gatio
n
e
nv
i
ronm
ent m
or
e accur
at
el
y
[2
4
].
2.2.
Radio
freq
uen
cy
p
attern
m
atchi
n
g
RF
Patt
ern
Ma
tc
hin
g
or
fin
ge
rprintin
g
is
cu
r
ren
tl
y
a
ver
y
a
ct
ive
top
ic
in
posit
ion
i
ng,
sin
ce
it
can
be
perform
ed
eff
e
ct
ively
in
the
c
el
lular
net
work
where
it
is
dis
cusse
d
in
t
he
m
eet
ing
of
3G
PP
un
der
t
he
nam
e
of
r
a
d
i
o
f
r
e
q
u
e
n
c
y
p
a
t
t
e
r
n
m
a
t
c
h
i
n
g
(
R
F
P
M
)
[
1
7
]
.
R
F
P
M
c
o
n
s
i
s
t
s
o
f
t
w
o
p
h
a
s
e
s
:
t
r
a
i
n
i
n
g
p
h
a
s
e
a
n
d
p
o
s
i
t
i
o
n
i
n
g
p
h
a
s
e
.
In
t
he
trai
ning
ph
a
se,
s
om
et
i
m
es
ref
er
re
d
to
as
t
he
off
-
li
ne
phase,
t
he
ai
m
is
to
create
a
fin
gerpr
i
nt
da
ta
base.
This
requires
a
ref
ere
nce
po
int
(RP)
w
hich
first
m
us
t
b
e
cho
se
n
caref
ully
.
In
ge
neral
,
the
ob
ta
ine
d
dat
a
represe
nt
the
r
ecei
ved
sig
nal
m
easur
ed
by
t
he
m
ob
il
e
sta
ti
on.
L
ocati
ng
a
m
ob
il
e
sta
ti
on
at
a
re
fer
e
nce
point
requires
that
t
he
receiv
ed
si
gn
al
of
al
l
bas
e
sta
ti
on
s
m
us
t
be
m
easur
ed
.
The
cha
racter
ist
ic
featur
es
of
su
c
h
m
easur
e
m
ents
of
t
he
re
fer
e
nce
poi
nt
will
be
e
valuated
and
t
he
n
rec
orded
i
n
the
dat
abase.
This
it
e
rati
ve
process
w
il
l be
p
e
rfor
m
ed
at
a
no
t
her re
fer
e
nc
e point, a
nd s
o on unti
l al
l refe
ren
ce
points a
r
e v
isi
te
d [2
5
,
26]
.
In
the
posit
ion
i
ng
phase
or
on
-
li
ne
phase,
wh
e
n
t
he
m
ob
il
e
sta
ti
on
posit
ion
is
require
d,
the
m
easur
em
e
nts
of
each
rec
ei
ved
si
gn
al
will
ta
ke
place.
T
hen,
the
m
easur
em
ents
will
c
om
par
e
with
da
ta
in
the
ser
ve
r
database
acc
ordin
g
to
the
ap
pro
pri
at
e
m
at
ching
al
gorithm
.
The
outc
om
e
of
t
hi
s
process
is
li
kel
y
to
be
the
locati
on
of the m
ob
il
e s
ta
ti
on
[2
5
,
26]
. Th
e
F
ig
ur
e
1
il
lustrate
s the
pri
nciple
of the
two phases
.
Chan
nel
obsta
cl
es
su
c
h
as
in
te
rf
ere
nce
a
nd
noise
are
c
ha
ng
e
d
unpredict
ably
over
ti
m
e
beca
us
e
of
us
er
m
ov
em
en
t.
Pr
e
dicti
ng
th
e
relat
ion
s
hip
betwee
n
receiv
ed
sig
nals
a
nd
the
dista
nce
in
wireless
netw
orks
is
determ
ined
by
path
los
s
an
d
sign
al
fa
ding.
Path
loss
ca
n
be
seen
a
s
an
analy
ti
cal
m
o
del
that
can
es
tim
a
te
the
receive
d
sign
al
w
he
n
a
cl
ear
li
ne
of
sig
ht
path
is
avail
able
betwee
n
r
ecei
ver
an
d
tra
ns
m
itter.
Path
loss
is
base
d
on
car
rier
f
re
qu
e
ncy
a
nd
distance
bet
ween
both
tra
nsm
it
te
r
and
rec
ei
ver
.
Acc
ordi
ng
to
Fr
ii
s
fr
ee
sp
ace
equ
at
io
n, the
re
cei
ved
powe
r c
an be
ob
ta
i
ne
d from
the
(2)
[2
7
]:
=
(
4
)
2
(2)
w
he
re:
=
⁄
(3)
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Fing
er
pr
i
nt po
sit
ion
in
g of u
se
rs d
evi
ces
in
l
ong t
erm
evol
ution
...
(
Mo
ham
med J.
Alh
as
an
)
531
Pr
re
presents
t
he
recei
ved
po
wer,
Pt
is
the
t
ran
sm
issi
on
po
wer.
Gt
a
nd
G
r
are
t
he
ante
nna
gai
n
of
transm
itter
and
receiver.
is
th
e
wav
el
en
gt
h
of
the
sig
nal,
c
is
the
li
gh
t
sp
eed,
f
is
the
ca
rr
ie
r
f
re
qu
e
ncy
and
d
r
e
p
r
e
s
e
n
t
s
t
h
e
d
i
s
t
a
n
c
e
b
e
t
w
e
e
n
t
r
a
n
s
m
i
t
t
e
r
a
n
d
r
e
c
e
i
v
e
r
.
W
e
c
a
n
a
l
s
o
w
r
i
t
e
t
h
e
e
q
u
a
t
i
o
n
i
n
d
B
u
n
i
t
a
s
f
o
l
l
o
w
i
n
g
:
=
+
+
−
(4)
w
he
re L
fs
can
be give
n by:
=
32
.
45
+
20
log
10
(
)
+
20
log
10
(
)
(5)
Ba
sed
on
a
co
m
bin
at
ion
of
Walfisch
a
nd
I
keg
am
i
-
Be
rto
ni
m
od
el
and
is
add
it
io
nally
dev
el
oped
by
COST
231
pro
j
ect
.
T
he
m
od
el
is
bu
il
t
on
nu
m
ero
us
a
naly
sis
an
d
sit
e
te
sts
and
it
is
su
it
a
ble
f
or
flat
su
bur
ba
n
and
urba
n
a
re
as
that
ha
ve
c
oor
din
at
ed
hei
gh
ts
of
buil
ding,
densel
y
locat
ed
buil
dings
and
la
rg
e
po
pula
ti
o
n
areas.
The
form
ula o
f
cal
cul
at
ing
the
Walfisch
-
I
keg
am
i
m
od
el
ca
n be
gi
ve
n by:
=
42
.
64
+
26
log
10
(
)
+
20
log
10
(
)
(6)
The
m
od
el
is
a
pp
li
cable
f
or
c
arr
ie
r
f
reque
nc
y
800M
Hz≤fc
≤20
00
M
Hz,
he
igh
t
of
e
Node
B
4m
≤
hb
≤
50
m
,
the
heig
ht
of
UE
1m
≤
hm
≤
3m
and
c
ov
e
ra
ge
a
rea
20
m
≤R≤5
km
[2
8
].
T
he
pr
e
di
ct
ed
rec
ei
ve
d
sign
a
l
betwee
n
tra
nsm
itter and
rece
iver base
d on s
hado
wing
bias
(ψ) ca
n be calc
ulate
d by [
2
9
]:
=
+
+
–
10
log
10
(
)
+
(7)
Figure
1. Tw
o ph
a
ses
of RFP
M posi
ti
on
in
g
3.
METHO
DOL
OGY
In
this
s
ubsect
ion,
the
te
sti
ng
env
i
ronm
ent
fo
r
est
im
ating
m
ob
il
e
sta
t
ion
will
be
base
d
on
MATL
AB
platfo
rm
.
Sever
al
scenari
os
will
be
exam
i
ned
unde
r
di
fferent
en
vir
on
m
ental
con
diti
on
s
i
n
orde
r
t
o
obta
in
acce
ptable
res
ults
of
m
ob
il
e
po
sit
io
ning.
The
acq
uire
d
r
esults
will
be
base
d
on
pat
h
loss
pro
pagat
ion
a
nd
log
-
norm
al
sha
dow
fad
i
ng
eff
ect
.
The
sta
ti
sti
cal
resu
lt
s
of
th
e
scena
rio
s
will
b
e
c
om
pu
te
d
the
or
i
gin
al
po
sit
io
n
of
m
ob
il
e
us
e
r
(
x
,
y
)
with
the
e
sti
m
at
ed
posit
ion
(`x,
`y)
a
nd
th
e
m
easur
e
m
ents
erro
r
wh
ic
h
c
an
be
def
i
ned in t
he
(
8)
:
=
√
(
−
`
)
2
+
(
−
`
)
2
(8)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
11
, No
.
1,
Febr
uar
y
2021
:
52
8
-
53
5
532
w
he
re
E
repres
ents
error
in
m
easur
em
ents
and
i
de
no
te
s
quantit
ie
s
rela
te
d
to
the
it
h
m
e
asur
em
ents.
By
us
in
g
Mi
cro
s
of
t
E
xc
el
we
can
obt
ai
n
these
m
easur
em
ents.
A
nothe
r
sta
ti
sti
c
will
be
com
pute
d
w
hich
re
presents
angular
er
ror
i
n
est
im
ation
m
easur
em
ents. E
rror i
n
a
ngle
s c
an be c
om
pu
te
d
acc
ordin
g
t
o t
he
(
9)
:
=
tan
(
−
−
)
(9)
w
he
re
(x
b
, y
b
)
represe
nts the
l
ocati
on of se
r
vi
ng
e
NB,
(x
,
y)
denotes to
the
locati
on of
UE
.
The
c
um
ulati
v
e
densi
ty
fu
nct
ion
(CDF)
will
be
com
pu
te
d
i
n
ord
er
to
dis
play
the
per
ce
ntil
e
values
of
67%,
90%,
95
% an
d 9
9%
to
see to
w
hat ex
t
ent ou
r
a
ppro
a
ch
ca
n
sat
isfy
FCC
r
e
qu
irem
ent.
3.1.
M
acr
o
ce
ll
e
nv
ir
onme
n
t
This
sce
nar
i
o
represe
nts
the
env
i
ronm
ent
of
m
acro
7
ce
ll
s,
on
e
-
ti
er
sy
stem
aro
un
d
the
ser
ving
eNodeB,
w
hic
h
is
in
the
centre.
W
e
al
so
a
ssu
m
e
that
eac
h
cel
l
has
a
hex
ag
onal
gr
i
d.
I
nter
-
sit
e
-
distan
ce
is
assum
ed
to
be
1000
m
et
er
with
a
tra
ns
m
issi
o
n
powe
r
of
46
dBm
and
ante
nna
gai
n
18dBi
f
or
ba
se
sta
ti
on
[29].
The
ass
um
ption
s
of
t
he
pa
ra
m
et
ers
are
bas
ed
on
3GPP
L
TE
sta
ndar
ds
[30
]
.
Ta
ble
1
pro
vid
es
sim
ulati
on
par
am
et
ers
in
detai
ls.
Table
1.
Ma
c
r
o
c
el
l
s
i
m
ulati
on
p
aram
et
ers
Para
m
eters
Valu
e
Bas
e statio
n
po
we
Tx an
ten
n
a gain
Rx
anten
n
a gain
Carrie
r
Fr
eq
u
en
cy
ISD
σs
h
ad
o
win
g
Nu
m
b
e
r
o
f
cells
Prop
ag
atio
n
M
o
d
el
4
6
dB
m
1
8
dB
i
0
dB
i
2
0
0
0
M
Hz
1
0
0
0
m
4
7
cells
W
alf
isch
-
Ikeg
a
m
i
m
o
d
el
The
sim
ulati
on
est
im
at
es
the
locat
io
n
of
m
ob
il
e
sta
ti
on
acco
rd
i
ng
to
K
near
est
nei
ghbo
ur
(
KNN
)
al
gorithm
as
discuss
e
d
it
s
pr
i
nciple
pr
e
viou
sly
.
Wh
en
the
m
easur
ed
RSS
from
UE
is
c
ol
le
ct
ed
the
sim
ulati
on
will
com
pu
te
the
diff
e
rence
betwee
n
m
easur
ed
RS
S
and
e
ac
h
ob
s
er
ved
RS
S
stored
i
n
da
ta
base.
If
th
e
m
at
ching
is
occ
urre
d,
the
pro
gram
will
return
t
he
re
fer
e
nce
po
i
nt
(
x,
y)
t
o
that
value
t
her
e
by,
the
est
i
m
ation
of
UE
is
occup
ie
d.
I
n
case
of
no
m
at
ching
,
oc
curred
wh
ic
h
is
m
or
e
pr
act
ic
al
ly
,
the
si
m
ul
at
ion
accor
ding
t
o
the
us
ed
al
gori
thm
co
m
par
es
the
m
ean
of
the
nea
rest
th
r
ee
neig
hbou
rs
co
ordinates
t
o
fi
nd
the b
e
st m
at
chi
ng of RS
S. Th
en,
retu
rn the
r
efere
nce
po
i
nt
that re
pr
ese
nts
the locati
on est
i
m
ating
of UE.
3.2.
Micr
o
cel
l enviro
nmen
t
In
Mi
cro
cel
l
scena
rio
we
al
so
will
co
ns
ide
r
t
he
sa
m
e
cel
ls
sit
e
in
t
he
pr
e
vi
ou
s
sce
nar
i
o.
Howe
ver,
the
cov
e
ra
ge
area
of
the
cel
ls
wi
ll
be
le
ss
than
m
acro
cel
l
to
prov
i
de
an
othe
r
te
st
appro
ac
h
f
or
fin
gerpr
i
nting
m
et
ho
d.
The
pa
ram
et
ers
of
th
e
si
m
ulati
on
and
cel
l
la
yout
will
assum
e
to
hav
e
slo
w
fa
di
ng
or
sh
a
dowing
e
ffec
t.
The
as
sum
pt
ion
s
of
t
he
par
am
et
ers
are
ba
sed
on
3GPP
LT
E
st
and
a
r
ds
[
30
]
.
Table
2
il
lustrate
s the
pa
ram
et
ers
of
t
he
prop
os
ed
sce
nar
i
o.
In
this
sce
na
rio,
the
tra
ns
m
is
sion
powe
r
of
the
eN
odeB
is
set
to
40dBm
si
nce
the
m
ic
ro
cel
l
req
uir
e
s
le
ss
powe
r
to
av
oi
d
inter
fer
e
nce.
A
ntenn
a
gai
n
of
the
tran
sm
i
tt
e
r
is
ass
um
ed
to
be
15
dB
i
an
d
inter
-
sit
e
-
distance
(I
S
D)
is
250
m
et
er
with
carrier
f
reque
nc
y
of
2000
M
H
z.
T
he
s
hado
w
fa
ding
is
lo
g
-
norm
a
l
distrib
ut
ion.
A
gain
the
sim
u
la
ti
on
will
est
i
m
at
e
the
loc
at
ion
of
m
ob
il
e
sta
ti
on
accor
ding
to
K
near
es
t
neig
hbour
(
KNN)
al
gorithm
in
the
sam
e p
rocedure t
hat exp
la
ined
in
m
acro
cell
sce
nar
i
o.
Table
2.
Mi
cr
o c
el
l
s
i
m
ulati
on
p
aram
et
ers
Para
m
eters
Valu
e
Bas
e statio
n
p
o
we
Tx an
ten
n
a gain
Rx
anten
n
a gain
Carrie
r
Fr
eq
u
en
cy
ISD
σs
h
ad
o
win
g
Nu
m
b
e
r
o
f
cells
Prop
ag
atio
n
M
o
d
el
4
0
d
B
m
1
5
d
Bi
0
d
Bi
2000
MHz
250
m
4
7
cells
W
alf
isch
-
Ikeg
a
m
i
m
o
d
el
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Fing
er
pr
i
nt po
sit
ion
in
g of u
se
rs d
evi
ces
in
l
ong t
erm
evol
ution
...
(
Mo
ham
med J.
Alh
as
an
)
533
4.
RESU
LT
S
A
ND
DI
SCUS
S
ION
S
The
ac
quire
d
resu
lt
s
s
howe
d
that
i
n
m
acro
cel
l
scena
rio,
m
ean
value
of
the
e
stim
at
i
on
er
ror
i
s
65.54
m
and
st
and
a
r
d
de
viati
on
is
43.
73
m
.
The
m
ean
an
d
sta
nda
rd
de
viati
on
value
s
giv
e
i
ndic
at
ion
tha
t
est
i
m
ation
erro
r
values
in
thi
s
scenari
o
are
acce
ptable
an
d
can
achieve
F
CC
con
diti
ons.
Figu
re
2
is
show
n
that
w
hen
the
distance
e
rror
is
increase
d
t
he
pro
ba
bili
ty
of
getti
ng
e
rror
is
al
so
inc
reas
ed.
T
his
m
eans
that
m
ob
il
e
us
er
posit
ion
locat
es
ne
ar
f
r
om
the
ba
se
sta
ti
on
c
an
be
s
poor
m
ore
accu
ratel
y.
Af
te
r
a
s
pecifi
c
valu
e
the
CD
F
giv
es
co
ns
ta
nt
va
l
ue
s
an
d
t
his
is
be
cause
at
that
point
t
he
e
rror
be
com
es
at
m
axi
m
u
m
value.
A
s
w
e
can
see
in
this scenari
o
the p
osi
ti
on
in
g
er
ror
at
67
%
is def
in
it
el
y
satisfied
the
FCC
re
qu
i
r
e
m
ent
of
posit
ion
i
ng
accuracy
as
w
el
l
as
90
%,
95
%
and
99
%
a
r
e
al
so
achieve
d
the
re
q
ui
rem
ent
of
FCC
.
T
urnin
g
to
m
ic
r
o
cel
l
Scena
rio,
a
nd
accor
ding
to
t
he
FCC
re
quir
e
m
ent
fo
r
posi
ti
on
in
g
acc
ur
a
cy
our
ap
proac
h
gi
ves
bette
r
resu
lt
s
than
pr
e
vious
scenari
o.
Where
m
ean
or
aver
a
ge
value
o
f
the
e
stim
at
ion
er
ror
is
16.75
m
and
s
ta
nd
a
rd
dev
ia
ti
on is
11.
15
m
as sho
wn
in
F
ig
ure
3.
Figure
2
.
CDF
of
m
acro
c
el
l
s
cenari
o
Figure
3
.
CDF
of
m
i
cro
c
el
l
s
cenari
o
Com
par
ing
our
r
esults t
o
the l
it
eratur
e
s
uch
a
s [
18]
sta
te
d
in
their r
esea
rch
that t
heir
res
ults for urb
a
n
area
sce
nar
i
o
wer
e
at
67
%
of
posit
io
ning
erro
r
was
38.
4
m
and
at
95
%
an
d
99%
w
ere
85
m
an
d
161.4
m
resp
ect
ively
.
Wh
e
reas,
the
researc
hers
in
[1
9]
decla
red
in
their
inv
e
sti
gation
in
CR
LB
on
RF
Patt
ern
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
11
, No
.
1,
Febr
uar
y
2021
:
52
8
-
53
5
534
Ma
tc
hin
g
m
eth
od
in
LTE
sy
stem
fo
r
urba
n
area
scenari
o
the
resu
lt
s
we
r
e
at
67
%
eq
ua
ls
49
.
6m
and
at
95
%
was
93.4
m
.
Th
eref
or
e
,
acco
r
di
ng
to
the
li
te
r
at
ur
e,
we
ca
n
deduce
t
hat
ou
r
re
s
ults
in
m
i
cro
cel
l
en
vir
onm
en
t
eff
ect
ively
im
pr
ove
d
the
accu
racy o
f UE
pos
it
ion
ing.
In
gen
e
ral,
ba
s
ed
on
the
res
ul
ts
ob
ta
ine
d
us
i
ng
KNN
al
gori
thm
,
the
si
m
ul
at
ion
sho
ws
th
at
m
ic
ro
cell
scenari
o
ca
n
pro
vid
e
bette
r
a
ccur
acy
for
posit
ion
in
g
tha
n
m
acro
cel
l
sc
enar
i
o.
By
stu
dy
ing
the
be
havi
our
of
cum
ulat
ive
fun
ct
ion
as
il
lustr
at
ed
in
F
ig
ur
e
4,
we
ca
n
see
that
w
hen
the
distance
betwe
en
m
ob
il
e
us
e
r
an
d
serv
i
ng b
a
se st
at
ion
decr
ea
se,
the
pro
bab
il
it
y o
f gett
ing
e
rro
r
will
also
d
e
cr
ease;
thu
s
, we
get b
et
te
r
accu
racy.
Figure
4
.
CDF
for
po
sit
io
ning
e
rro
r
5.
CONCL
US
I
O
N
In
this
w
ork,
r
adio
f
reque
ncy
patte
r
n
m
at
ching
(RFPM
)
m
et
ho
d
has
be
en
pro
posed
to
determ
ine
a
us
er
’s
posit
ion
i
n
LTE
syst
e
m
s
us
ing
KNN
al
go
rith
m
.
Ou
r
stu
dy
adopte
d
a
th
eor
et
ic
al
ap
proach
t
o
i
m
ple
m
ent the
m
et
ho
d
ba
sed on
a co
m
pu
te
r pro
gr
am
. D
epe
nd
i
ng
on
path
loss propa
gatio
n
an
d
sh
a
dow
f
adin
g
eff
ect
s
on
t
he
sign
al
,
t
he
dat
abase
of
the
s
uggeste
d
m
e
th
od
was
c
ollec
te
d
us
i
ng
a
pro
gr
am
-
base
d
-
c
om
pu
te
r.
Our pr
opose
d
m
et
ho
d was
test
ed
a
nd ev
al
ua
te
d
into
tw
o
s
cenari
os
unde
r diffe
re
nt con
diti
on
s.
The
obta
ine
d
r
esults
from
the
si
m
ulati
on
sh
ow
e
d
that
the
accuracy
was
i
m
pr
ov
e
d
in
both
scenari
os
and
m
et
the
FCC
req
ui
rem
e
nts.
T
he
stu
dy
has
dem
on
str
at
ed
that
the
RFPM
m
e
tho
d
has
s
how
n
ve
ry
go
od
perform
ance
in
re
du
ci
ng
posit
ion
in
g
er
ror
i
n
both
scena
rio
s
,
15m
at
67%
and
31.
78m
at
90%
of
posit
io
ning
error f
or Mi
cro cel
l scena
rio
a
nd 75.6
6m
(
67%)
a
nd 14
1.4m
(90
%
) for
Ma
cro cel
l scena
rio.
REFERE
NCE
S
[1]
J.
A.
del
Per
al
-
Rosado,
et
al.
,
“
Surve
y
of
C
el
lu
l
ar
Mobile
R
adi
o
Loc
a
l
izati
on
Methods:
From
1G
to
5G,
”
in
I
E
EE
Comm
unic
ati
ons Surve
ys
&
Tuto
rials
,
vol
.
20
,
no
.
2
,
pp
.
1124
-
11
48,
2018
.
[2]
F
ede
ral
Com
muni
cations
Com
m
ission
,
“
W
ire
le
ss
E911
Loca
ti
on
Acc
u
racy
Requi
rement
s
,
”
2011.
[Onlin
e
]
.
Avail
ab
le
:
htt
p
:/
/
hra
unfoss
.
fcc.go
v/e
docs_
pub
li
c
/
at
t
ac
hm
at
ch
/DA
-
11
-
920A1.pdf
.
[3]
J.
Trogh,
et
al
.
,
“
Outdoor
loc
at
i
on
tra
ck
ing
of
m
obil
e
device
s
i
n
ce
l
lul
ar
net
wo
rks,”
Eurasip
Jo
ur
nal
On
Wirel
e
ss
Comm
unic
ati
ons
a
nd
Ne
tworki
ng
,
2019
.
[4]
S
.
A
.
Zulkifl
ie
,
et
al.
,
“
D
y
n
a
m
ic
navi
g
at
ion
indoor
m
ap
u
sing
W
i
-
Fi
fin
ger
printing
m
obil
e
technolog
y,
”
Inte
rnational
Jo
urnal
of El
e
ct
ri
c
al
and
Comput
er
Engi
n
ee
ring
,
v
o
l.
9
,
n
o
.
2
,
pp
.
73
9
-
746,
2020
.
[5]
Li,
e
t
al
.
,
“
High
-
pre
ci
sion
posi
ti
oning
al
gor
it
h
m
s
base
d
on
G
PS
,
”
2nd
Inte
rnational
Confe
re
nce
on
Adv
anc
e
d
Computer
Control
,
vo
l
.
5,
pp.
36
4
-
368,
2010
.
[6]
F.
Bene
d
et
to
,
et
al
.
,
“
Enha
nc
ed
TOA
-
base
d
indo
or
-
positi
oning
algorithm
for
m
obil
e
LTE
c
ellula
r
s
y
stems
,
”
201
1
8th
Workshop o
n
Posit
ion
ing, N
avi
gati
on
and
C
omm
unic
ati
on
,
Dresden,
pp
.
13
7
-
142,
2011
.
[7]
S.
Çola
k,
et
al
.,
“
Anal
y
zi
ng
c
el
l
phone
locat
ion
dat
a
for
urba
n
tra
ve
l:
cur
ren
t
m
et
hods,
li
m
it
at
ions
,
an
d
o
p
p
o
r
t
u
n
i
t
i
e
s
,
”
T
r
a
n
s
p
o
r
t
a
t
i
o
n
R
e
s
e
a
r
c
h
R
e
c
o
r
d
:
J
o
u
r
n
a
l
o
f
t
h
e
T
r
a
n
s
p
o
r
t
a
t
i
o
n
Res
e
a
r
c
h
B
o
a
r
d
,
p
p
.
1
2
6
-
1
3
5
,
2
0
1
5
.
[8]
A.
Kanga
s
and
T.
W
igr
en,
“
An
gle
of
Arriva
l
L
oca
l
iz
a
ti
on
in
{
LT
E}
Us
ing
{
MIM
O}
Pre
-
Coder
Inde
x
Fe
edbac
k,
”
in
IE
EE Comm
unic
ati
ons
Letters
,
vol
.
17
,
no
.
8
,
p
p.
1584
-
1587
,
2
013.
[9]
E.
Tre
visan
i
and
A.
Vit
al
e
tt
i
,
“
C
el
l
-
ID
loc
a
ti
on
t
ec
hniqu
e,
li
m
it
s
and
b
en
efi
ts:
an
expe
riment
al
stu
d
y
,
”
S
ix
th
IE
EE
Workshop on
M
obil
e
Computing
Syste
ms
and
Ap
pli
cations
,
W
ind
ermere
,
Cum
bria,
UK
,
pp
.
51
-
60
,
2004.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Fing
er
pr
i
nt po
sit
ion
in
g of u
se
rs d
evi
ces
in
l
ong t
erm
evol
ution
...
(
Mo
ham
med J.
Alh
as
an
)
535
[10]
Borokows
ki,
et
al
.
,
“
Perform
anc
e
of
ce
l
l
ID+RT
T
h
y
brid
p
ositi
oning
m
et
h
od
for
UM
TS
rad
io
net
works
,
”
in
Proc
ee
dings
o
f
European
Wire
le
ss
Confe
ren
ce
,
pp.
487
-
492,
20
04
.
[11]
L.
Gaz
za
h
,
e
t
a
l
.,
“
Sel
ec
t
ive
h
y
brid
RS
S/AO
A
weight
ing
a
l
gorit
hm
for
NLOS
int
ra
cell
l
oca
l
iz
a
ti
on
,
”
IE
E
E
Wirel
ess Comm
unic
ati
ons
and
Ne
tworki
ng
Conf
er
enc
e
(
WCNC)
,
pp.
2546
-
2551
,
2
014.
[12]
F.
Adela
n
ta
do,
et
a
l.
,
“
Under
st
andi
ng
th
e
L
imits
of
LoRa
W
AN
,
”
in
IEEE
Co
mm
unic
ati
ons
Magazine
,
vo
l.
55,
no.
9
,
pp
.
34
-
40
,
2017.
[13]
D
.
Jos
e
and
S
.
Sebastian
,
“
Tay
l
o
r
serie
s
m
et
hod
in
TDOA
appr
oac
h
for
indoor
p
ositi
oning
s
y
s
tem
,
”
Inte
rnationa
l
Journal
of
Elec
t
rical
and
Computer
Eng
ine
ering
(
IJE
CE)
,
v
ol. 9,
n
o.
5
,
pp
.
3927
-
3933,
2019
.
[14]
V.
Os
a,
et
al
.
,
“
Loc
a
li
z
at
ion
in
W
ire
le
ss
Networks:
The
Potentia
l
of
Triangulatio
n
Te
chn
ique
s
,
”
Wirel
ess
Pe
rs
on
al
Comm
unic
ati
ons
,
vol
.
68
,
pp
.
152
5
-
1538,
2013
.
[15]
T.
Zha
ng
,
et
a
l.
,
“
A
novel
OTD
OA
positi
oning
sche
m
e
in
het
er
ogene
ous
LT
E
-
a
dvanc
ed
s
y
st
ems
,
”
in
Proceedi
ng
s
o
f
t
h
e
3
r
d
I
E
E
E
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
N
e
t
w
o
r
k
I
n
f
r
a
s
t
r
u
c
t
u
r
e
a
n
d
D
i
g
i
t
a
l
C
o
n
t
e
n
t
(
I
C
-
N
I
D
C
'
1
2
)
,
p
p
.
1
0
6
-
1
1
0
,
2
0
1
2
.
[16]
M.
A.
Am
in
an
d
S.
Fis
che
r
,
“
Pos
it
ion
locati
o
n
of
LTE
f
emtoce
l
ls
depl
o
y
e
d
in
a
c
luste
r,
”
2012
Inte
rnatio
na
l
Confe
renc
e
on
Locali
zation
and
GNSS
,
Starnberg
,
pp
.
1
-
6
,
2012
.
[17]
3GP
P
TR
36.
80
9
V0.3.
0.
“
Evolved
Univer
sal
T
err
estrial
Rad
io
Acc
ess
(E
-
UTRA):
RF
Patt
er
n
Matc
h
ing
lo
ca
t
i
on
m
et
hod
in the
L
TE
,
”
2012
.
[Onl
i
ne]
.
Available:
h
tt
p://ww
w.3gpp.org/ftp/
Spe
cs/a
r
chi
ve
/36_seri
es/
36.
809/
.
[18]
Xiao,
e
t
al.
,
“
Cr
amer
-
Rao
Lowe
r
Bound
on
RF
pat
t
ern
m
at
chi
n
g
m
et
hod
with
vel
oc
ity
in
LT
E
s
y
stem,
”
in
8t
h
Inte
rnational
Co
nfe
renc
e
on
Wir
el
ess and
M
obile
Comm
unic
ati
on
s
,
Venice, Ital
y
,
pp.
94
-
97
,
2012
.
[19]
J
.
Zhu,
e
t
al.
,
“
Maximum
li
kel
i
hood
sche
m
e
for
finge
rprin
ti
ng
positi
oning
in
LT
E
s
y
s
te
m
,
”
2
012
IEE
E
14t
h
Inte
rnational
Co
nfe
renc
e
on
Co
mm
unic
ati
on
Te
chnol
o
gy
,
Ch
eng
du,
pp
.
428
-
432
,
2012.
[20]
B.
Li,
et
al
.
,
“
I
ndoor
positi
onin
g
t
ec
hn
ique
s
ba
sed
on
wire
l
ess
LAN
,
”
in
1st
IE
EE
Int
ernati
ona
l
Confe
ren
ce
o
n
Wirel
ess B
roadb
and
and
Ultra
W
ide
band
Comm
u
nic
ati
ons
,
S
y
den
y
,
Aus
tra
l
ia
,
pp
.
1
-
7,
2006
.
[21]
P.
Bahl
and
V.
N.
Padm
ana
b
ha
n,
“
RAD
AR:
an
in
-
buil
ding
RF
-
base
d
us
er
locat
ion
and
tr
ac
king
s
y
stem
,
”
Proc
.
IEE
E
INFOCOM
2000.
Confe
renc
e
on
Computer
Comm
unic
ati
ons.
19
th
Annual
Joi
nt
Confe
r
enc
e
of
th
e
IE
E
E
Computer
and
C
omm
unic
ati
ons
Soci
e
ti
es
(
Cat. N
o.
00CH37064)
,
Te
l
Aviv,
Israe
l,
v
ol.
2
,
pp
.
775
-
7
84,
2000
.
[22]
D.
Pand
y
a,
e
t
a
l
.,
“
Indoor
loc
a
tion
esti
m
at
ion
u
sing
m
ult
ipl
e
wi
rel
ess
t
ec
hnolog
ie
s
,
”
14th
I
EE
E
Proceedi
ngs
on
Pe
rs
onal,
Indoo
r and
Mo
bile
Ra
dio
Comm
unic
at
ions
(
PIM
RC
20
03
)
,
vol.
3
,
pp
.
2
208
-
2212
,
2013
.
[23]
A.
Bat
t
it
i
,
e
t
a
l
.,
“
Loc
at
ion
-
awa
r
e
computing:
a
n
eur
al
n
et
work
m
odel
for
d
et
erm
i
ning
locati
on
in
wire
le
ss
LANs
,
”
Te
chn
ic
a
l
R
epor
t,
2002
.
[Onl
ine
]
.
Avai
la
b
le
:
htt
p
:
//
epr
in
ts.bi
b
li
o.
u
nit
n.
i
t/
233/1
/83.pdf
.
[24]
G.
W
ölfe
,
et
al
.,
“
Enha
nce
d
Loc
a
li
z
at
ion
Techni
q
ue
withi
n
Urb
an
and
Indoor
Env
i
ronm
ent
s
base
d
on
Acc
ur
a
te
and
Fast
Propaga
t
ion
Models
,
”
It
aly
,
Firenc
e
,
F.
,
pp
.
1
-
7,
2002
.
[25]
B.
Li,
et
al
.,
“
Method
for
y
i
el
ding
a
d
at
ab
a
se
o
f
loc
ation
finge
rprin
ts
in
W
LAN
,
”
in
IE
E
Proceedi
ngs
-
Comm
unic
ati
ons
, v
ol
.
152
,
no
.
5
,
pp.
580
-
586
,
20
05.
[26]
Firdaus,
et
al.
,
“
Fingerpri
nt
ind
oor
positi
oning
base
d
on
user
orie
nt
at
io
ns
an
d
m
ini
m
um
computat
ion
t
ime
,
”
Inte
rnational
Jo
urnal
of El
e
ct
ri
c
al
and
Comput
er
Engi
n
ee
ring
,
v
ol
.
17
,
n
o.
4
,
pp
.
1
740
-
1749,
2019
.
[27]
S.
Saunder
s
an
d
Aragón
-
Za
v
ala,
“
Antenna
s
a
nd
propa
gation
for
wire
le
ss
co
m
m
unic
at
ion
s
y
stems
(2nd
ed.
)
,
”
Chic
hest
er,
England
,
John
W
iley &
S
on
s Ltd
,
200
7.
[28]
A.
Goldsm
it
h,
“
W
ire
le
ss
Com
muni
cations,
”
Ca
mbr
idge
Unive
rs
it
y
Press
,
2004
.
[29]
H.
H
.
A
.
Toska
l
a,
“
L
TE
for
UM
TS
-
OF
DMA
and
SC
-
FDM
A Ba
s
ed
Rad
io
Ac
ce
ss
,
”
W
iley
,
2009
.
[30]
3GP
P
TS
36.
101
V
12.
0.
0
.
,
“
E
volve
d
Univer
sa
l
Te
rr
estrial
Ra
dio
Ac
ce
ss
(E
-
UTRA):
Us
er
E
quipment
-
R
adio
Tra
nsm
ission a
n
d
Recept
ion
,
”
20
13.
[Onlin
e]
.
Av
ai
l
abl
e
:
ht
tp:
/
/www
.
3gpp.
org/ft
p
/
Specs/a
r
chi
ve
/3
6_serie
s/36.1
01
/
.
Evaluation Warning : The document was created with Spire.PDF for Python.