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.
9
, No
.
6
,
Decem
ber
201
9
, p
p.
5412~
5419
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v9
i
6
.
pp5412
-
54
19
5412
Journ
al h
om
e
page
:
http:
//
ia
es
core
.c
om/
journa
ls
/i
ndex.
ph
p/IJECE
Impl
emen
tatio
n
of vo
yage d
ata rec
or
din
g devi
ce usi
ng a digi
tal
forensi
cs
-
based h
ash algo
rithm
Ki
-
Taek
Se
ong
1
, Gw
an
-
H
yung Ki
m
2
1
Dep
ar
t
m
ent
of
I
nform
at
ion
Secu
rity
,
Tongm
y
ong
Univ
ersity
,
Sout
h
Korea
2
Dep
ar
t
m
ent
of C
om
pute
r
Engi
n
ee
ring
,
Tongm
y
ong
Univ
ersity
,
South Kore
a
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
A
pr
1
9
, 201
9
Re
vised
Ju
l
5
,
201
9
Accepte
d
J
ul
1
7
, 2
01
9
Ide
nti
f
y
i
ng
the
c
ause
s
of
m
ari
ne
ac
c
ide
nts
is
difficult
be
ca
use
of
proble
m
s
in
sce
ne
pre
serv
at
i
on,
ree
na
ct
m
ent,
and
proc
uring
of
witne
ss
es.
Thanks
to
new
reg
ulations,
la
rg
er
vessels
ar
e
n
ow
req
uire
d
to
ca
rr
y
vo
y
a
ge
d
a
ta
re
cor
der
s
(VD
Rs)
and
aut
om
at
ic
id
ent
if
icati
on
s
y
s
te
m
s
(AIS
s).
How
eve
r,
the
cont
en
t
of
the
se
device
s,
which
is
cre
ated
,
stored,
and
m
a
nage
d
digitall
y
,
has
sec
uri
t
y
vulne
rab
il
i
ti
es
such
as
the
potentia
l
for
d
at
a
m
odifi
c
at
ion
.
Therefore,
whe
n
m
ana
ging
dig
it
a
l
rec
ords
it
is
im
porta
nt
to
gu
arante
e
reliab
il
i
t
y
.
To
th
is
end
,
we
suggest
a
dig
it
al
fore
nsi
cs
-
ba
sed
digi
t
al
rec
o
r
ds
m
igra
ti
on
m
ethod
using
a
hash
al
gor
it
hm
t
o
guar
an
tee
th
e i
nte
grity
and
au
th
ent
i
ci
t
y
of
d
igi
t
a
l
re
cor
ds.
Ke
yw
or
d
s
:
Au
t
hen
ti
ci
ty
Digital
for
e
ns
i
cs
Hash al
gorith
m
In
te
gr
it
y
NMEA
20
00
Vo
ya
ge data
r
e
corder
Copyright
©
201
9
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
Ki
-
Tae
k
Se
ong,
D
ep
ar
t
m
ent
of
Inform
at
ion
Se
cur
it
y,
To
ng
m
yong Univ
e
rsity
.
Sout
h Korea
.
Em
a
il
:
taichiboy
1
@
gm
ail.co
m
1.
INTROD
U
CTION
It
is
oft
en
di
ff
i
cult
to
i
den
ti
fy
the
ca
us
es
of m
arine
incide
nt
s,
not on
ly
b
ec
ause
is
it
dif
ficult
to
sec
ure
evide
nce
asso
ci
at
ed
with
the
locat
ion
an
d
case,
bu
t
al
so
beca
us
e
the
intangib
le
fac
tors
relat
ed
to
hu
m
an
resou
rces,
s
uc
h
as
m
ariner
s,
and
t
he
physi
cal
factor
s
relat
ed
to
th
e
e
nvir
on
m
ent,
su
c
h
as
ve
ssels,
harb
or
facil
it
ie
s,
and operati
onal
e
qu
i
pm
ent, ar
e inte
rtwine
d.
To
re
duce
thes
e
diff
ic
ulti
es,
r
ecentl
y
bu
il
t
sh
ips
a
re
eq
uipped
with
a
nd
op
e
rated
by
di
gital
dev
ic
es
enco
m
passing
the
auto
pilot
syst
e
m
,
the
vo
ya
ge
data
recorde
r
(
VD
R
),
the
autom
at
ic
i
den
ti
ficat
io
n
s
yst
e
m
(AIS),
an
d
CC
TV.
Since
m
arine
traff
ic
syst
e
m
s
hav
e
becom
e
m
or
e
co
m
plica
te
d,
m
arine
incidents
includi
ng
colli
sion
s
a
nd
sink
i
ng
ha
ve
i
ncr
ease
d;
how
ever,
beca
us
e
evide
nce
has
been
dif
ficult
to
secu
re,
m
aking
it
diff
i
c
ult
to
ide
ntify
the
cause
s
of
i
ncide
nts,
the
In
te
rn
at
io
na
l
Ma
ritim
e
Organ
iz
at
io
n
(
IM
O)
unde
r
the
U
N
has
i
m
po
sed
a
ne
w
regulat
io
n
r
equ
i
rin
g
the
in
sta
ll
at
ion
of
a
VD
R
an
d
an
AI
S
on
al
l
ves
sel
s
[1
]
.
The
VD
R
wa
s
dev
el
op
e
d
t
o
store
t
he
data
of
the
onboa
rd
di
gi
ta
l
dev
i
ces
to
help
in
identify
in
g
t
he
cause
of
a
m
arine
incident.
T
he
VD
R
data
i
nclud
e
the
sh
i
p’
s
locat
ion,
velo
ci
ty
,
and
hea
din
g
as
well
as
the
co
nversati
on
s
on
bo
a
r
d,
the
V
HF
ra
dio
c
om
m
un
ic
at
ion
s
on
the
s
hip’s
operati
on,
a
nd
t
he
inf
or
m
at
ion
disp
la
ye
d
on
the
rad
a
r
[2
]
.
Since
the
stored
el
ect
ronic
reco
r
ds
are
dig
it
al
data,
they
are
vu
l
ne
rab
le
to
integ
r
it
y
and
authe
ntici
ty
da
m
age
by
the
ph
ysi
cal
dete
rior
at
io
n
of
st
or
a
ge
de
vices
or
lo
gic
cha
ng
e
s
in
the
el
ect
ronic
recordi
ng
syst
e
m
.
The
store
d
data
shou
l
d
be
p
r
otect
ed
from
un
auth
ori
zed
acce
ss
an
d
from
delet
io
n
by
carele
ss
m
anag
e
m
ent
[3
]
.
T
his
pap
e
r
descr
i
be
s
a
m
et
ho
d
for
preve
ntin
g
ar
bitrary
delet
io
n
or
al
te
rati
on
o
f
dat
a
in
the
stora
ge
dev
ic
e
by
im
ple
m
enting
dig
i
ta
l
fo
rensi
cs
w
hen
recordi
ng
vo
ya
ge
data
in
t
he
stora
ge
de
vice.
In
ad
diti
on,
t
hi
s
pa
per
desc
ri
bes
t
he
im
plem
entat
ion
of
di
gital
foren
sic
s
f
or
ide
ntifyi
ng
a
nd
c
har
act
e
rizi
ng
the ty
pes
of
d
a
ta
that can b
e u
sed
as evid
e
nc
e fo
r
the in
vest
igati
on
and ass
essm
ent o
f
m
ar
ine inciden
ts a
s w
el
l
as in c
ollec
ti
ng, r
est
or
i
ng, a
nd an
al
yz
ing t
he
per
ti
ne
nt
data.
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
Impleme
nta
ti
on
of vo
y
age
da
t
a
rec
ordin
g de
vi
ce using
a di
gita
l f
or
e
ns
ic
s
-
ba
s
ed
has
h
..
.
(
Ki
-
Ta
ek
Seon
g)
5413
2.
RESEA
R
CH MET
HO
D
2.1.
NM
E
A 200
0
–
ba
sed
d
ata
re
cordin
g
de
vic
e
The
integ
rati
on
of
sh
i
pboard
syst
e
m
s,
sensing
a
nd
co
ntr
ol
within
syst
em
s,
sh
a
rin
g
of
inf
or
m
at
ion
,
and
c
ollec
ti
ng
of
data
are
occ
urrin
g
at
a
n
i
nc
reasin
g
rate
on
bo
a
r
d
vessel
s.
A
gen
e
ral
s
hi
pboard
c
onfi
gurati
on
is
sh
own
in
Fi
g
ure
1.
As
show
n
in
Fig
ure
1,
basic
de
vi
ces
includi
ng
sens
or
s
,
act
uator
s
,
an
d
com
pass
are
li
nk
ed
to
t
he
c
onve
ntion
al
N
MEA
-
0183
ne
twork
,
w
hile
ot
her
dev
ic
es
in
cl
ud
in
g
se
nsor
s,
act
uato
rs,
e
ng
i
ne,
fu
el
,
r
a
dar,
GPS, a
utopil
ot,
a
nd EC
DI
S
are
li
nk
e
d
t
o
the
inst
ru
m
ent n
et
w
orks (NME
A 2
000)
[4
,
5].
Figure
1. S
hipbo
a
r
d netw
ork
s and i
nterf
ace
s
2.2.
Voy
ag
e
d
ata
r
ecordin
g
The
vo
ya
ge
da
ta
recorde
r
(
VD
R
)
is
inten
ded
f
or
us
e
i
n
al
l
passenge
r
sh
ips
as
well
as
in
new
ly
const
ru
ct
e
d
ca
rgo
s
hip
s
of
3000
gross
to
nnage
a
nd
up.
T
he
S
-
V
DR
(sim
pl
ifie
d
V
DR
)
is
f
or
existi
ng
car
go
vessels
of
the
sam
e
cat
ego
ry
with
a
ph
ase
-
i
n
sche
du
le
re
quirin
g
instal
la
ti
on
first
in
ves
sel
s
of
20,
000
gros
s
tonnage
a
nd
up,
to
be
f
oll
ow
ed
by
oth
e
rs
of
3000
gro
ss
to
nn
a
ge
a
nd
up.
The
S
-
V
DR
is
no
t
require
d
to
store
the
sam
e
le
vel
of
detai
l
as
the
VD
R
bu
t
sh
ould
m
ai
ntain
a
store,
i
n
a
secu
re
a
nd
retrie
vab
le
form
,
of
in
f
or
m
at
ion
con
ce
rn
i
ng
t
he
po
sit
io
n,
m
ov
em
ent,
ph
ysi
cal
sta
tus,
co
m
m
and
,
a
nd
con
t
ro
l
of
a
sh
i
p
ove
r
the
pe
rio
d
le
a
ding
up
to
a
nd
fo
ll
ow
i
ng
a
n
incident.
Ta
ble
1
s
umm
arizes
the
tw
o
ty
pes
of
data
it
em
s
to
be
save
d on the
V
DR
[6
-
11]
.
Table
1.
Data i
tem
s to
be rec
orde
d
No.
Da
ta
to
be
rec
o
rd
e
d
VDR
S
-
VDR
In
terf
a
c
e
1
Da
te
a
n
d
tim
e
IEC
61
1
6
2
2
S
h
i
p
'
s p
o
si
t
ion a
n
d
d
a
tu
m
us
e
d
IEC
61
1
6
2
3
S
p
e
e
d
IEC
61
1
6
2
4
He
a
d
in
g
(f
r
o
m
c
o
mpass
)
IEC
61
1
6
2
St
e
p
,
S
y
n
c
h
ro
,
Analo
g
5
B
ri
d
g
e
a
u
d
i
o
(b
y
o
n
e
or
m
o
re
m
i
c
ro
p
h
o
n
e
s)
Aud
i
o
6
C
o
mm
u
n
ica
ti
o
n
s au
d
io
Aud
i
o
7
R
a
d
a
r, p
o
st
-
d
is
p
lay
sele
c
ti
o
n
R
,
G,
B
,
H,
V
R
,
G,
B
,
H
,
V
(i
f availabl
e
)
8
De
p
th
NM
E
A
/ IE
C
61
1
6
2
IEC
61
1
6
2
in
terf
a
c
e
(if
a
v
a
il
a
b
le)
"
"
"
"
"
"
"
9
Ma
in
a
lar
m
s
NM
E
A
/ IE
C
61
1
6
2
,
C
o
n
tac
t, A
n
a
lo
g
10
R
u
d
d
e
r o
rd
e
r and
re
sp
o
n
se
IEC
61
1
6
2
,
C
o
n
tac
t
,
Analo
g
11
E
n
g
in
e
o
rd
e
r and
re
sp
o
n
se
IEC
6
1
1
6
2
,
C
o
n
tac
t
,
Analo
g
12
Hul
l o
p
e
n
ings
s
tatu
s
IEC
61
1
6
2
,
C
o
n
tac
t
13
Wa
terti
g
h
t a
n
d
fi
re door
stat
u
s
IEC
61
1
6
2
,
C
o
n
tac
t
14
Ac
c
e
lera
ti
o
n
s and
h
u
ll
st
ress
e
s
IEC
61
1
6
2
,
C
o
n
tac
t
,
Analo
g
15
Win
d
s
p
e
e
d
a
n
d
di
r
e
c
ti
o
n
IEC
61
1
6
2
,
Analo
g
16
AIS
in
f
o
rm
a
ti
o
n
IEC
61
1
6
2
-
2
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.
9
, N
o.
6
,
Dece
m
ber
2
01
9
:
5412
-
5419
5414
The
i
nterf
ace
s
for
the
it
e
m
s
that
are
sa
ved
ne
ed
to
sat
isfy
I
nt
ern
at
io
nal
Ele
ct
ro
te
ch
nical
C
omm
issi
on
sta
nd
a
rd
IEC
61162.
Th
e
IE
C
61162
-
1
sta
nd
a
r
d
refe
rs
to
NMEA
0183,
and
IEC
61
162
-
2
re
fer
s
t
o
NME
A
2000. T
he
refore
, th
e
V
DR
ne
eds
t
o
sat
isfy t
he
inte
rf
ace
stand
a
r
d of
NME
A 200
0,
or IE
C 6
1162
-
2.
3.
FORE
NS
I
CS
-
BASED
DAT
A RECO
R
DI
NG
TE
CHNI
QUES
3.1.
Forensics
-
b
ase
d a
cci
de
nt i
n
vest
i
gatio
n p
r
ocess
Be
cause
of
the
uniq
ue
c
ha
rac
te
risti
cs
of
m
arine
acci
de
nts,
the
i
ns
ta
ll
at
ion
of
data
rec
ordi
ng
dev
ic
e
s
su
c
h
as
black
boxes
on
ai
r
planes
is
m
and
at
or
y.
H
oweve
r,
su
c
h
dig
it
al
data
are
vu
lne
rabl
e
to
m
od
ific
at
ion
a
s
they
are
rest
ored
a
nd
us
e
d
within
th
e
de
vice
pro
vid
e
d
by
the
dat
a
recordi
ng
de
vice
m
anu
fa
ct
ur
er
.
Fo
r
the
rec
ord
ed
data
to
ha
ve
le
gal
ef
fecti
ven
e
ss,
fore
ns
ic
te
chn
i
qu
e
s
s
hould
be
ap
plied,
a
nd
a
proc
ess
f
or
su
bm
it
ti
ng
su
c
h
data
as
le
ga
ll
y
bin
ding
ev
idence
in
th
e
even
t
of
a
m
a
rine
acci
de
nt
has
bee
n
propose
d.
As
show
n
in
Figure
2,
data
save
d
on
VD
R
are
sa
ved
usi
ng
a
di
gital
fo
ren
sic
s
m
et
ho
d,
an
d
the
pro
cess
of
pr
ese
r
ving
the
evide
nce
ha
pp
ens
w
he
n
the
s
hip
sta
rts
t
o
sink.
E
vid
e
nce
preser
vatio
n
re
f
ers
to
the
prev
entio
n
of
a
ny
m
anipula
ti
on
of
the
sa
ved
data,
wh
ic
h
is
an
esse
ntial
factor
for
th
e
data
to
be
c
on
side
red
a
s
ob
je
ct
ive
evide
ntial
data.
Des
pite
su
c
h
evide
nce
pres
erv
at
io
n,
t
he
da
ta
can
sti
ll
be
m
anipu
la
te
d
t
hro
ugh
the
i
nterf
ace
,
and th
us
it
is
ne
cessary t
o co
nfi
rm
w
hethe
r
t
he data
are
from
the tim
e o
f
a
cci
den
t.
Figure
2. Proce
ss for i
nvest
iga
ti
on
of m
arine casu
al
ti
es an
d
i
ncide
nts
us
in
g digit
al
for
e
ns
ic
techn
i
qu
e
s
3.2.
Forensics
-
Bas
ed Data
Rec
or
ding an
d
Rep
rodu
c
tio
n Tec
hniques
Re
cent
dig
it
al
dev
ic
es
us
e
d
in
sh
i
ps
a
dopt
the
sta
ndar
d
ne
twork
,
an
d
th
e
represe
ntati
ve
m
od
el
i
s
the
NME
A
2000.
T
hro
ugh
t
his
net
wor
k,
t
he
de
vices
tra
ns
m
it
the
voya
ge
-
relat
ed
data
to
the
NME
A
20
00
netw
ork,
w
hic
h
a
re
the
n
s
ave
d
by
the
V
DR,
centeri
ng
ar
ou
nd
the
data
s
um
m
arized
in
T
able
1.
As
m
entioned
befor
e
,
the
dat
a
nee
d
to
be
sa
ved
with
out
an
y
distor
ti
on
a
nd
m
us
t
be
a
ble
to
be
v
e
rified
a
s
the
data
ass
oc
ia
te
d
with a
giv
e
n pe
rio
d.
Fig
ur
e
3 s
hows
t
he data
re
cordin
g proce
ss prop
os
e
d
in
this pa
per.
Figure
3. F
or
e
ns
ic
s
-
base
d data rec
ordin
g pro
cess
In
ge
ner
al
,
t
he
vo
ya
ge
da
ta
are
tra
ns
m
i
tted
an
d
sa
ve
d
trans
par
e
n
tl
y
in
the
VD
R
.
As
s
how
n
in
Figure
3,
t
he
pro
posed
m
et
ho
d
r
ecei
ves
t
he
def
i
ned
secret
key
f
ro
m
each
data
it
em
and
us
es
t
he
data
s
aved
in
the
VD
R
a
nd
def
i
ned
secret
key
t
o
ge
ner
a
te
an
a
uth
e
nti
cat
ion
c
ode,
w
hich
is
the
n
s
aved
on
the
lo
gg
i
ng
dev
ic
e.
T
he
a
ut
hen
ti
cat
ion
co
de
i
s
a
un
i
qu
e
value
f
or
the
s
aved
data
an
d
is
us
e
d
t
o
ver
i
f
y
the
a
uth
e
ntici
ty
of
the
data
when
m
anipu
la
ti
on
ta
kes
place
within
t
he
V
D
R.
Fig
ur
e
4
s
hows
t
he
a
uthe
ntica
ti
on
pro
cess
f
or
the d
at
a
save
d i
n
the
V
DR
us
ing t
he
l
ogging
dev
ic
e.
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
Impleme
nta
ti
on
of vo
y
age
da
t
a
rec
ordin
g de
vi
ce using
a di
gita
l f
or
e
ns
ic
s
-
ba
s
ed
has
h
..
.
(
Ki
-
Ta
ek
Seon
g)
5415
Figure
4. F
or
e
ns
ic
s
-
ba
se
d V
DR dat
a aut
he
ntica
ti
on
proce
ss
Figure
4
s
hows
the
data
a
uth
e
ntica
ti
on
process
us
i
ng
the
l
oggi
ng
dev
ic
e
a
nd
V
DR
data
from
the
foren
sic
s
-
base
d
m
et
ho
d.
The
authe
ntica
ti
on
co
de,
ge
ner
at
e
d
w
hen
the
data
was
saved
,
is
sav
ed
on
the
lo
gg
i
ng
dev
ic
e.
Th
en
,
when
t
he
ne
w
sec
ret
ke
y
extracte
d
f
ro
m
the
rece
ived
V
DR
m
at
ches
the
aut
he
ntica
ti
on
value
as
well
as
the
da
ta
save
d
in
the
lo
ggin
g
da
ta
,
the
i
ntegr
it
y
of
the
V
D
R
data
is
consi
de
red
co
nf
irm
ed,
a
nd
w
he
n
the
inf
or
m
at
ion
do
e
s
not
m
atch
,
t
he
data
are
c
on
si
der
e
d
to
be
dam
aged
[12
-
27]
.
4.
IMPLEME
N
TATION
AN
D RESULTS
4.1.
Te
st
Be
d
Using
the
m
eth
od
propose
d
in
this
pa
pe
r
,
the
net
wor
k,
as
sh
ow
n
in
Fi
gure
5,
was
de
s
ign
e
d
with
the
sh
ip
boar
d
env
i
ronm
ent
in
m
ind
.
The
GPS
receiver
,
sho
wn
in
Fi
gure
5,
receives
the
s
ign
al
from
the
GP
S
and
pro
vid
es
a
var
ie
ty
of
da
ta
includi
ng
t
he
tw
o
-
dim
ension
al
locat
io
n
of
th
e
sh
i
p,
it
s
sp
ee
d,
it
s
ve
locit
y,
and
the
ti
m
e.
The
NME
A
20
00
is
res
pons
i
bl
e
fo
r
data
co
m
m
un
ic
at
ion
s,
us
in
g
a
receiv
er
an
d
transm
itter
that
sat
is
fy
the
sta
ndar
ds
of
t
he
C
AN
2.0
c
omm
un
ic
at
io
n
protoco
l,
an
d
the
CPU
m
anag
es
the
instal
le
d
ha
rdwar
e
and
im
ple
m
ent
s
pro
gram
s.
In
this
ar
rangem
e
nt,
first
the
G
P
S
recei
ver
tra
nsm
it
s
the
GPS sign
al
,
pa
rsing it
an
d
changin
g
it
to
the
NME
A
20
00
f
or
m
at
,
and
then
t
he
logge
r
us
es
this
m
essage
to
ge
ne
rate
the
secret
key
us
in
g
the
m
e
tho
d
gi
ven
i
n
Sect
io
n
4.2
belo
w,
creati
ng
t
he
a
uth
e
ntica
ti
on
ke
y.
Th
e
voya
ge
data
a
re
s
aved
i
n
the
V
DR,
w
hile
the
a
uth
e
ntica
ti
on
key
is
sa
ved
in
t
he
lo
ggin
g
dev
ic
e.
T
able
2
sho
ws
t
he
s
pecif
ic
at
io
ns
for
the test
b
e
d.
T
he
act
ual
test
bed is s
how
n
i
n Fi
gure
6.
Figure
5. NME
A 200
0 netw
ork
te
st
bed
Table
2.
Sp
eci
f
ic
at
ion
s fo
r
te
s
t bed
GPS r
e
ceiv
er
Log
g
er
Bo
ard n
a
m
e
Main role
(Related
hard
ware
)
Bo
ard n
a
m
e
Main role
(Related
hard
ware
)
CPU
b
o
ard
Cen
tral
p
rocess
in
g
un
it (
Ardu
in
o
UN
O):
Extern
al hardwa
re
co
n
trol
GPS
m
ess
ag
e parsi
n
g
Vo
y
ag
e data t
rans
m
i
ss
io
n
CPU b
o
ard
Cen
tral
p
rocess
in
g
un
it (
Ardu
in
o
UN
O):
Extern
al hardwa
re
co
n
trol
Secret
k
e
y
ext
racti
o
n
and
auth
en
ticati
o
n
cod
e
g
en
eration
,
sav
ed
in
log
g
in
g
dev
ice
Vo
y
ag
e data sav
ed
in VDR
GPS bo
ard
GPS sig
n
al pro
ces
sin
g
un
it (
Ardu
in
o
GPS
Sh
ield
):
GPS sig
n
al r
ec
eiv
in
g
Data log
g
er
Sav
in
g
un
it (
Ardu
in
o
SD
Sh
ield
):
VDR an
d
log
g
in
g
d
ev
ice r
o
le
CAN co
m
m
.
b
o
ard
Co
m
m
u
n
icatio
n
un
it (
Ardu
in
o
CAN
Sh
ield
):
CAN 2
.0 co
m
m
u
n
i
catio
n
tr
an
s
m
iss
io
n
p
rocess
in
g
CAN co
m
m
.
b
o
ard
Co
m
m
u
n
icatio
n
un
it (
Ardu
in
o
CAN
Sh
ield
):
CAN 2
.0 co
m
m
u
n
i
catio
n
tr
an
s
m
iss
io
n
p
rocess
in
g
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.
9
, N
o.
6
,
Dece
m
ber
2
01
9
:
5412
-
5419
5416
Figure
6. Test
bed
4.2.
Secret
key a
n
d g
e
nera
tio
n
of
au
t
hent
ic
at
i
on
c
od
e
The
in
form
at
i
on
t
hat
can
be
receive
d
f
rom
this
GP
S
re
cei
ver
is
as
f
ol
lows
:
tim
e,
data,
locat
ion,
sp
ee
d
(
knots
),
ang
le
,
al
ti
tud
e,
sat
el
li
te
s,
$P
G
TOP
data,
$G
PGGA
data,
a
nd
$GPRM
C
da
ta
.
Su
c
h
i
nform
at
ion
m
us
t
be
ex
pr
e
ssed
i
n
NME
A
0183
form
a
t
in
order
to
be
tr
ansm
it
te
d
thr
ough
t
he
N
MEA
2000
N
et
work.
Figure
7
s
hows
the actu
al
N
M
EA 0
183 form
at
f
or
receivin
g data.
Figure
7. Data
as p
a
rse
d
f
or
N
MEA
0183
for
m
at
As
al
l
of
the
i
nfor
m
at
ion
rec
ei
ved
from
the
receive
r
is
sti
ng
-
ty
pe
data,
no
disti
nguish
i
ng
is
need
e
d
wh
e
n
ge
ner
at
i
ng
the
secret
key.
T
he
refor
e
,
in
this
st
ud
y
,
the
G
PS
rec
ei
ver
ge
ner
at
e
d
the
sec
ret
ke
y
an
d
authe
ntica
ti
on
co
de
only
for
the
th
ree
m
essa
ges
(tim
e,
locat
ion
,
a
nd
$GP
RM
C)
that
fall
unde
r
N
os
. 1
a
nd
2
i
n
Table
1.
GP
R
MC
sta
nd
s
f
or
"Glob
al
P
o
sit
ion
in
g
Re
com
m
end
ed
Mi
ni
m
u
m
Sp
eci
fic
GNSS
Data,"
and
the
$GPRM
C
m
e
ssage
prov
i
de
s
su
ch
i
nform
at
ion
as
U
TC
tim
e
/date
,
sta
tus,
la
ti
tud
e/
longit
ude,
E/
W
/S/
N
ind
ic
at
or,
s
pee
d
ove
r
gro
und,
and
c
ourse
ov
er
gro
und.
T
he
fo
ll
owin
g
pro
cess
show
s
ho
w
the
a
uth
e
ntica
ti
on
cod
e
(
A
un
t
h_Cod
e
)
is
ge
nerat
ed
f
ro
m
the
receive
d
m
essage
(Msg
).
Fir
st,
t
he
sec
ret
ke
y
is
extracte
d
from
the m
essage to be sa
ve
d.
T
he
n, the a
uth
e
ntica
ti
on
c
od
e
is
ge
ner
at
e
d by co
m
bin
ing
this
w
it
h
the sec
ret
ke
y.
Secret
Key = F
un
ct
io
n(M
sg,
va
riables)
;
Au
nth_C
od
e
=
Functi
on(Ms
g + Sec
ret Key)
;
Ther
e
a
re
va
riou
s
m
et
ho
ds
f
or
gen
e
rati
ng
the
secret
key,
includi
ng
e
nc
odin
g
the
set
va
lue
or
us
in
g
the
strin
g
in
a
certai
n
locat
io
n.
Howe
ver
,
i
n
order
t
o
en
ha
nc
e
the
secu
rity
,
it
is
best
fo
r
it
to
be
re
cei
ve
d
fro
m
th
e
resp
ect
i
ve
m
essage,
an
d
it
is
bette
r
fo
r
it
to
be
of
a
fi
xe
d
le
ng
t
h
re
gard
le
ss
of
the
le
ng
th
of
the
m
essage
to
be
processe
d.
The
has
h
f
unc
ti
on
,
w
hich
is
widely
us
e
d
t
o
ver
ify
i
ntegri
ty
becau
se
of
it
s
char
act
eris
ti
c
of
creati
ng
a
val
ue
of
a
gi
ve
n
le
ng
t
h,
was
us
e
d
in
this
st
ud
y.
The
sec
ur
it
y
le
vel
was
en
ha
nc
ed
by
us
i
ng
t
he
has
h
functi
on
in
ge
ner
at
in
g
both
the
aut
hen
ti
cat
ion
c
ode
a
nd
the
sec
ret
key.
The
re
a
re
vari
ou
s
al
gorithm
s
tha
t
i
m
ple
m
ent
the
hash
f
unct
io
n,
includi
ng
MD
5
(Messa
ge
-
Di
gest
al
gorithm
),
S
HA
(
Secu
r
e
Hash
Algorit
hm
)
–
1/256,
a
nd
S
H
A
-
256/
224/51
2/
383.
M
D5
is
use
d
to
ve
rify
th
e
or
i
gin
al
c
opy
of
a
file
or
pro
gr
am
becau
se
of
it
s
32
-
bit
–
base
d
f
ast
processin
g
sp
eed
.
S
HA
-
1
increase
the
ha
sh
val
ue
in
orde
r
to
e
nh
a
nc
e
the
secu
rity
le
vel,
bu
t
it
is
kn
own
to
be
vu
lne
ra
bl
e
to
existi
ng
m
et
ho
ds
of
at
ta
ck.
Th
ere
fore,
the
nu
m
ber
of
ty
pes
of
cal
culat
ion
and the
has
h v
al
ue
we
re i
ncr
e
ased,
and
SHA
-
25
6/224
/
512/
383 we
re sugg
est
ed.
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
Impleme
nta
ti
on
of vo
y
age
da
t
a
rec
ordin
g de
vi
ce using
a di
gita
l f
or
e
ns
ic
s
-
ba
s
ed
has
h
..
.
(
Ki
-
Ta
ek
Seon
g)
5417
As
s
how
n
in
Figure
8,
th
e
secret
key
gen
e
rated
us
in
g
the
m
essage
“Date:
27
/
2/
2016
”
was
“0x7
48
e
06…”,
and
w
he
n
thi
s
was
com
bin
ed
with
“
Date:
27
/2/2
016”
thr
ough
has
hing,
the
a
uth
e
ntica
ti
on
cod
e
“0
x328
21…”
was
ge
nerat
ed.
T
hen
the
m
essage
was
s
aved
in
the
V
DR,
a
nd
t
he
a
uth
e
ntica
ti
on
ke
y
was
save
d
in
the
loggin
g
de
vi
ce.
On
ce
t
he
VD
R
m
essage
is
m
od
ifie
d,
t
he
a
uthe
ntica
ti
on
c
ode
in
the
authe
ntica
ti
on
proce
ss
wi
ll
no
t
m
at
ch;
t
her
e
fore
aut
he
ntici
ty
is
gu
ar
anteed
in
the
even
t
that
a
m
at
ch
is
ob
ta
ine
d.
Fig
ure
9
sho
ws
th
e
process
f
or
ge
n
erati
ng
a
n
a
uth
e
ntica
ti
on
co
de
t
hat
us
es
t
he
S
HA
-
1
a
nd
MD5
al
gorith
m
s.
Figure
8. Ge
ne
rati
ng au
t
hen
ti
cat
ion
c
ode f
or
SHA
-
256
(a)
(b)
Figure
9. Ge
ne
rati
ng au
t
hen
ti
cat
ion
c
ode f
or
SHA
-
1 (a)
a
nd MD5 (
b)
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.
9
, N
o.
6
,
Dece
m
ber
2
01
9
:
5412
-
5419
5418
Table
3
s
umm
arizes
the
te
st
res
ults.
T
he
s
ecur
it
y
of
t
he
has
h
al
gorit
hm
show
s
the
f
re
qu
e
ncy
of
cl
ashin
g,
in
w
hich
the
sam
e
ou
tp
ut
val
ue
occurs
f
or
diff
e
ren
t
in
pu
t
v
al
ues
.
The
l
ow
e
r
the
f
re
quency,
the
hi
gher
t
he
secu
rity
le
vel,
an
d
th
us
the
SHA
-
25
6
m
e
thod
is
t
he
m
os
t
sta
ble
,
bu
t
it
requires
a
long
e
r
processi
ng tim
e, as sh
own
i
n Table
3.
Table
3.
Su
m
m
ary o
f results
Alg
o
rith
m
(Outp
u
t bits
)
Inp
u
t
Key
valu
e
/
Execu
tio
n
ti
m
e us
Au
n
th
_
Code v
alu
e
/
Execu
tio
n
ti
m
e us
Secu
rity
MD5
(12
8
)
String
1
0
x
4
d
ae9… / 1
5
0
4
0
x
e1
4
b
8
… /
1
3
,520
Low
String
2
0
x
0
0
0
7
8
… /
1
3
,524
0
x
2
3
2
7
c… /
1
3
,524
String
3
0
x
a2
3
d
a… /
1
3
,52
4
0
x
0
9
d
b
1
… /
1
3
,524
SHA
-
1
(16
0
)
String
1
0
x
8
f
f7c6
… /
3
5
,984
0
x
5
eea8… / 57
,200
Mediu
m
String
2
0
x
4
9
9
5
2
… /
5
7
,200
0
x
6
8
5
a2
… /
5
7
,200
String
3
0
x
8
b
c
2
c… /
5
7
,20
0
5
e8
6
9
2
a… /
5
7
,20
0
SHA
-
256
(25
6
)
String
1
0
x
7
4
8
e0
… /
6
3
,008
0
x
3
2
8
2
1
6
2
… /
8
2
,16
0
Hig
h
String
2
0
x
8
f
0
8
6
b
… /
8
2
,1
6
0
0
x
4
5
d
5
2
… /
8
2
,160
String
3
0
x
c8
e8
6
… /
8
2
,160
0
x
7
b
b
0
0
… /
8
2
,160
No
te:
String
1 = “
Data: 2
7
/2
/2
0
1
6
”
String
2 = “
$GPG
GA,02
5
5
3
8
.000,3
5
0
7
.1156
,N,
1
2
9
0
6
.15
1
2
,E
,1,0
4
,9.0
8
,11
3
.3,M,
2
5
.4,M,
,
*
6
A”
String
3 = “
Locati
o
n
: 35
0
7
.11
4
9
N,
1
2
9
0
6
.1
5
1
3
E”
5.
CONCL
US
I
O
N
In
this
st
ud
y,
a
vo
ya
ge
da
ta
reco
r
ding
m
et
ho
d
us
in
g
di
gital
fo
re
ns
ic
s
was
pr
opos
e
d
a
nd
i
m
ple
m
ented.
The
m
essage
received
f
r
om
t
he
GPS
recei
ve
r
is
change
d
into
the
NMEA
0183
f
or
m
at
t
hro
ugh
par
si
ng.
T
he
ti
m
e
and
locat
io
n
in
form
at
ion
for
the
VD
R
r
ecord
it
em
is
us
ed
to
ge
ner
at
e
a
secret
key;
this
is
us
e
d
to
ge
nerat
e
the
auth
en
ti
cat
ion
co
de
f
or
t
he
m
essage
an
d
is
the
n
save
d.
A
ha
sh
functi
on
is
use
d
t
o
ge
ne
rate
a
sec
r
et
key
for
eac
h
of
the
save
d
m
essages.
In
gen
e
rati
ng
the
authe
ntica
ti
on
cod
e
,
the
secre
t
key
and
the
m
essage
are
com
bine
d,
an
d
the
has
h
f
un
ct
io
n
is
use
d
in
orde
r
to
gu
a
ran
te
e
the
un
i
qu
e
c
har
act
erist
ic
s
of
the
m
essage
and
it
s
auth
entic
it
y.
In
t
his
way,
the
has
h
functi
on
wa
s
us
ed
bo
t
h
to
enh
a
nce
the
le
vel
of
secur
it
y
and
t
o
uniq
uely
identify
it
.
Thr
ee
widely
us
ed
ty
pes
of
hash
al
gorithm
s,
S
HA
-
1,
SHA
-
256,
an
d
MD5, we
re im
plem
ented.
The
act
ual
dat
a
save
d
in
the
VD
R
ca
n
be
in
string
unit
s
or
file
un
it
s.
Wh
ereas
irre
gula
r,
short
pieces
of
in
form
at
ion
are
transm
it
te
d
in
the
fo
rm
of
a
m
essage,
analog
sig
nals,
in
cl
ud
in
g
voic
e
r
ecordin
gs
an
d
rad
a
r
i
m
ages,
can
be
saved
in
file
un
it
s.
T
he
adv
a
ntage
of
this
m
et
ho
d
is
t
hat
it
can
easi
ly
be
app
li
ed
t
o
the
tra
d
it
ion
al
VD
R
since
it
c
an
be
a
ppli
ed
t
o
the
af
orem
entioned
file
s.
T
he
us
e
of
f
oren
sic
s
f
or
V
DR
da
ta
to
be
us
e
d
as
e
vi
den
ti
al
da
ta
in
the
e
ven
t
of
m
arine
acci
de
nts
ca
n
e
nh
a
nc
e
t
he
obj
ect
i
vity
of
the
e
vi
den
ce
,
and it
is our h
ope that t
he pr
opose
d rec
ordin
g de
vice
will
be widely
us
e
d.
REFERE
NCE
S
[1]
Inte
rna
ti
ona
l
Marit
ime
Organ
izati
on
(IMO
),
“
I
nte
rna
ti
ona
l
Co
nvent
ion
for
the
Safe
t
y
of
Li
fe
at
Sea
(SO
LAS),
”
Chapt
er
V,
In
te
r
nat
ion
al
Mar
it
im
e
Organi
zation,
London,
1974
.
[2]
Ministr
y
o
f
La
n
d,
“
Tra
nsport
an
d
Marit
ime
Affa
irs,
Notice
2012
-
075:
Vess
el
Eq
uipment
Standa
r
ds,
”
Arti
cl
e
108
.
7
(Vo
y
ag
e
Da
ta
R
ec
orde
r)
,
2012
.
[3]
Ministr
y
of
La
n
d,
“
Tra
nsport
a
nd
Marit
ime
Af
fai
rs,
Int
ern
a
ti
o
nal
Mari
ti
m
e
O
rga
nizati
on
55
th
Marit
ime
Saf
e
t
y
Com
m
it
te
e
Fina
l
Report,
”
pp.
2
9
-
32,
2009
.
[4]
J.
H.
Huh,
et
al.
,
“
Design
of
N
MEA2000
CAN
Bus
Inte
gra
te
d
Network
Sy
st
e
m
and
Its
Te
st
Bed:
Sett
ing
Up
the
PLC
S
y
stem
in
Be
twee
n
Bri
dge
—
Bow
Room
Sect
ion
on
a
Conta
iner
Ship
as
a
B
ac
kbone
S
y
stem,”
Lect
ur
e
Note
s
i
n
Elec
trical
Eng
ine
ering
,
Springer
,
vol
.
2
016,
pp
.
191
-
20
4,
2015
.
[5]
A.
Ninom
i
y
a,
“
About
the out
l
in
e
of
NM
EA
200
0®,
”
Tok
y
o
Uni
ver
sit
y
o
f
Merc
a
nti
le Mari
n
e
,
pp.
52
-
54,
2011
.
[6]
S.
C.
Aus
ti
n
and
P.
A.
W
il
son,
“
Marit
ime
vo
y
a
g
e
da
ta
rec
ord
er
s
tud
y
b
y
the
Eur
opea
n
m
aritime
dat
a
m
ana
gemen
t
proje
c
t,
”
Int
ernati
onal Journal
o
f
Mariti
me
Engi
n
ee
ring
,
vol
.
151
,
pp.
13
-
24,
2009
.
[7]
Uchij
ima,
et
al.,
“
Deve
lopment
o
f
VD
R
(Voy
age
Data
Re
cor
der
)
,
”
Nihon
Musen
Giho,
JR
C
Revi
e
w
,
JR
C
A
uthors
,
vol.
67
,
pp
.
42
-
4
5,
2016
.
[8]
Hs
u,
et
al.,
“
Co
nstruct
ing
an
Ef
fic
i
ent
St
at
e
Space
Quer
y
S
y
s
tem
for
the
Vo
y
a
ge
Dat
a
Re
cor
d
er,
”
Frontiers
i
n
Arti
ficial Int
el
l
ig
enc
e
and
Appl
i
c
ati
ons
,
IOS
Pres
s
,
vol.
2015,
pp.
294
-
305,
20
15
.
[9]
J.
Kang,
et
al.
,
“
Deve
lopment
of
Remote
Alar
m
Module
with
Pl
a
y
ba
ck
fun
ct
ions
in
Vo
y
age
Data
Rec
orde
r
,
”
SIC
E
-
A
nnual
C
on
fe
re
nce
,
vol
.
2009
,
p
p.
3I06
-
4
,
2009
.
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
Impleme
nta
ti
on
of vo
y
age
da
t
a
rec
ordin
g de
vi
ce using
a di
gita
l f
or
e
ns
ic
s
-
ba
s
ed
has
h
..
.
(
Ki
-
Ta
ek
Seon
g)
5419
[10]
C
.
Jung,
et
al.
,
“
The
d
eve
lopmen
t
of
E
the
rn
et
b
ase
d
rad
ar
and
EC
DIS
image
proc
essing
for
vo
y
ag
e
da
ta
r
ec
ord
er
,
”
Proce
ed
ings
of
the
14th
Inte
rn
ati
onal
Confe
re
nce
on
Control,
Aut
omation
and
Syste
ms
,
Gye
onggi
-
do,
Kore
a,
Re
public
o
f
,
pp.
963
-
966
,
2014
.
[11]
Marit
ime
navi
g
a
ti
on
and
rad
i
oco
m
m
unic
at
ion
eq
uipment
and
s
y
s
te
m
s
-
Shipborn
e
vo
y
age
da
ta
re
cor
der
(VD
R)
-
Part1:
Perform
a
nce
r
equi
r
ement
s,
m
et
hods of tes
ti
ng
and
r
equi
re
d
te
st
result
s
,
I
E
C
61966
-
1,
Ma
y
,
2013
.
[12]
M
y
eong
H
.
B
.
a
nd
Sangji
n
L.
,
“
A
new
i
nvest
igation
m
et
hodo
log
y
of
m
ari
ne
c
asua
lt
i
es
and
in
ci
d
ent
s
using
d
igi
t
a
l
fore
nsic
techniq
ues,
”
Journal
of
The
Korea
Instit
ute
of
Informati
on
Sec
urity
&
C
ryptol
ogy
,
vol.
2
3,
pp.
515
-
530
,
2013
.
[13]
M.
Piccinelli
an
d
P.
Gubian
,
“
Modern
ships
vo
y
a
ge
da
ta
rec
ord
er
s:
A
fore
nsics
p
e
rspec
ti
v
e
on
the
Costa
Concord
ia
shipwrec
k,
”
Dig
it
al Inve
st
igat
ion
,
vol
.
10
,
pp
.
41
-
49,
2013
.
[14]
A.
Frieze
and
T
.
Johanss
on,
“
On
the
insert
ion
t
ime
o
f
ran
dom
walk
cu
ckoo
ha
shing,”
John
W
i
ley
&
Sons
,
Lt
d
,
vol.
54
,
pp
.
721
-
729,
2019
.
[15]
Z.
Han,
et
al.
,
“A
novel
routi
ng
al
gorit
hm
for
IoT
cl
oud
base
d
on
hash
offset
tre
e,
”
Fu
ture
gene
r
ati
ons
compute
r
systems
,
El
s
ev
i
e
r Sc
ie
n
ce B.V
.
,
A
ms
te
rdam
,
vol.
8
6,
pp
.
456
-
463
,
201
8.
[16]
S.
L.
Garf
inke
l a
nd
M.
McCar
rin
,
“
Hash
-
base
d
carving:
Sear
chi
ng
m
edi
a
for
complete
files
and
file
fra
gm
ent
s
with
sec
tor
h
ashing a
nd
hashdb,
”
Digi
t
al
Inv
esti
gat
ion
,
vol
.
14
,
pp
.
95
-
105,
2015
.
[17]
M
.
G.
Noblet
t,
et
al.
,
“
Rec
ove
r
ing
and
Exa
m
in
ing
Com
pute
r
F
ore
nsic
Evi
d
ence,
”
Journal
in
Forensic
Scienc
e
Comm
unic
ati
ons
,
v
ol
.
2
,
pp
.
1
-
13
,
2000
.
[18]
W
.
G.
Henr
ique,
“
Anti
Forensics
:
Making
computer
for
ensic
s h
ar
d,
”
Cod
e
Br
ea
k
e
rs III,
Sao
Paulo
,
Brazil
,
2006
.
[19]
M.
C.
Stamm
and
K.
J.
R.
Li
u
,
“
Anti
-
fore
nsics
o
f
digi
t
al
imag
e
c
om
pre
ss
ion,
”
IE
EE
Tr
ans.
Inf.
F
orensics
Sec
urit
y
,
vol
.
6
,
pp
.
1050
-
1065,
2011
.
[20]
M.
C.
Stamm
,
e
t
al.
,
“
Anti
-
for
en
sics
of
JP
EG
compress
ion,
”
Proc.
IE
EE
In
t.
Co
nf.
A
cousti
cs
,
Sp
ee
ch
,
and
Signa
l
Proce
ss
ing
,
pp
.
1694
-
1697
,
201
0
.
[21]
M.
C.
Stamm
a
nd
K.
J.
R.
Li
u
,
“
W
ave
le
t
-
bas
ed
image
compres
sion
ant
i
-
for
ensi
cs,
”
Proc.
I
EEE
Int.
Con
f.
Imag
e
Proce
ss
in
g
,
pp
.
173
7
-
1740
,
201
0
.
[22]
M.
C.
Stamm
,
e
t
al.
,
“
Forensics
vs.
anti
-
fore
nsi
cs:
A
decision
a
n
d
game
the
or
etic
fra
m
ework,”
Proc.
I
EE
E
Int
.
Conf.
Ac
ousti
cs,
Spee
ch
,
and
Sig
nal
Proc
essing,
Kyot
o,
Japan
,
p
p.
1749
-
1752
,
2
012
.
[23]
M.
Chen
and
W.
Hu,
“
Resea
rch
on
Bat
SLAM
Al
gorit
hm
for
UAV
Based
on
Aud
io
Perc
ept
u
al
Ha
sh
Closed
-
Loop
Dete
c
ti
on,
”
In
te
r
nati
onal journal of pat
t
ern
re
cog
nit
ion
and
arti
fi
c
ial
in
telli
g
ence
,
vol.
33
,
2019
.
[24]
G.
Coll
om
,
et
al.,
“
Fast
Mesh
-
to
-
Mesh
Remaps
Us
ing
Hash
Alg
orit
hm
s,”
SIAM
Journal
on
Scie
nti
fic
Computin
g
,
vol.
40
,
pp
.
C45
0
-
C450,
2018
.
[25]
X.
W
ang
and
J.
Yu,
“
Studie
s
on
An
Online
Densit
y
Sensi
tive
Hash
Algori
thm,”
M
INIMIC
RO
S
Y
STEMS
-
SHEN
Y
ANG
,
v
o
l.
39
,
pp
.
1068
-
1
073,
2018
.
[26]
Y.
Qin,
et
al.
,
“
Multi
-
stage
IPv6
Routi
ng
Looku
p
Algorit
hm
Based
on
Hash
Ta
bl
e
and
Mult
ibi
t
T
rie
,
”
MINIMICR
O
S
Y
STEMS
-
SHE
N
Y
ANG
,
v
ol.
39,
pp.
893
-
898,
20
18
.
[27]
M
.
A
.
Abdulha
y
oglu
and
B
.
Thi
j
s
,
“
Us
e
of
loc
al
ity
sensitive
hashi
ng
(LSH) a
lgori
t
hm
to
m
at
ch
W
eb
of
Scie
nce
and
Scopus,”
Sc
ie
n
to
metric
s
,
Ak
ad
emiai
Kiado
R
t.
,
vo
l.
116
,
pp
.
1229
-
1245,
2018
.
Evaluation Warning : The document was created with Spire.PDF for Python.