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
.
5
,
Octo
ber
201
9
, pp.
4020
~
40
26
IS
S
N:
20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v
9
i
5
.
pp4020
-
40
26
4020
Journ
al h
om
e
page
:
http:
//
ia
es
core
.c
om/
journa
ls
/i
ndex.
ph
p/IJECE
IoT
b
ase
d
c
ar
a
ccident
d
etection
an
d
n
otification
a
l
gorithm
for
g
en
eral
r
oad
a
ccident
s
Shiv
an
i S
ha
r
ma,
Sh
on
e
y
S
ebastian
Depa
rt
m
ent
o
f
C
om
pute
r
Scie
n
ce
,
CHRIS
T
(De
e
m
ed
to
b
e
Unive
rsit
y
)
,
Ind
ia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Oct
16
, 201
8
Re
vised
A
pr
18
, 2
01
9
Accepte
d
Apr
25
, 201
9
W
it
h
an
inc
r
eas
e
in
populatio
n,
the
r
e
is
an
inc
re
ase
in
the
num
ber
of
ac
c
ide
nts
th
at
h
a
ppen
eve
r
y
m
in
ute
.
The
se
ro
ad
ac
c
ide
nts
a
re
un
pre
dictable.
The
re
ar
e
situ
ations
where
m
o
st
of
th
e
a
ccid
ent
s
cou
ld
not
be
rep
or
te
d
prope
rl
y
to
n
ea
r
b
y
ambulan
ce
s
on
ti
m
e.
In
m
ost
of
the
c
ase
s,
t
her
e
is
th
e
unava
i
la
bi
li
t
y
of
emerge
nc
y
serv
ic
es
which
l
ac
k
i
n
providi
ng
the
f
irst
ai
d
and
ti
m
ely
servi
ce
which
c
an
l
ea
d
to
loss
of
li
fe
b
y
som
e
m
inut
es.
He
nce
,
the
r
e
i
s
a
nee
d
to
d
evelop
a
s
y
st
em
tha
t
c
at
ers
to
a
l
l
the
se
probl
ems
an
d
ca
n
eff
ectivel
y
func
ti
on
to
over
co
m
e
the
del
a
y
tim
e
ca
used
b
y
the
m
edi
cal
vehi
c
le
s.
The
pu
rpose
of
thi
s
pape
r
is
to
int
roduc
e
a
fra
m
ework
using
IoT,
which
hel
ps
in
det
e
ct
ing
ca
r
ac
c
ide
nts
and
n
oti
f
y
ing
the
m
i
m
m
edi
at
ely
.
Thi
s
ca
n
be
a
c
hie
ved
b
y
integ
rat
in
g
sm
art
sensors
with
a
m
ic
roc
ontroller
withi
n
th
e
c
ar
th
at
c
an
tr
igge
r
at
the
t
ime
of
an
a
cc
id
ent
.
The
oth
er
m
odule
s
li
ke
GP
S
and
GS
M
are
int
egr
at
ed
wi
th
th
e
sy
stem
to
ob
ta
in
the
lo
catio
n
coor
dinates
of
th
e
a
cc
i
dent
s
and
sending
i
t
to
reg
iste
red
num
ber
s
an
d
ne
ar
b
y
ambulanc
e
to
n
oti
f
y
the
m
about
the
ac
c
id
ent
t
o
obta
in
imm
edi
ate
hel
p
at
the
loc
a
ti
on
.
Ke
yw
or
d
s
:
Ardu
i
no
Ca
r
a
cci
de
nt
In
te
r
net
of
t
hings
No
ti
ficat
io
n
Sensors
Copyright
©
201
9
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights
reserv
ed
.
Corres
pond
in
g
Aut
h
or
:
Sh
iva
ni S
ha
rm
a,
Dep
a
rtm
ent o
f C
om
pu
te
r
Scie
nce,
CHRIST
(Dee
m
ed
to
be Un
i
ver
sit
y),
Ho
s
ur Roa
d, B
ang
al
or
e
, Kar
na
ta
ka,
560029
,
India
.
Em
a
il
:
sh
ivani.
sh
arm
a@cs.chri
stun
i
ver
sit
y.i
n
1.
INTROD
U
CTION
Nowa
days,
th
ere
is
an
incr
ease
in
the
nu
m
ber
of
acci
den
ts
that
ha
ppen
in
the
w
or
l
d.
As
the
popula
ti
on
is
increasin
g,
t
he
re
is
the
nu
m
ber
of
ca
rs
i
nc
reasin
g
on
t
he
ro
a
d
that
c
ontrib
utes
to
se
ver
e
acci
den
ts
t
hat
happe
n
daily
.
Aroun
d
80
per
cent
of
acci
de
nts
c
on
tri
bute
to
the
los
s
of
m
any
li
ves.
Mos
tl
y,
the
grow
i
ng
co
untrie
s
are
bein
g
ta
rg
et
ed
by
t
he
day
to
day
ro
a
d
acci
den
t
s.
Th
e
m
ajo
r
reason
is
t
he
la
ck
of
infr
a
struct
ur
e
,
la
ck
of
traf
fic
con
t
ro
l
a
n
d
ac
ci
den
t
m
anag
e
m
ent.
O
ut
of
al
l
the
dev
el
opin
g
c
ountries
,
India
ha
s
been
li
ste
d
as
t
he
co
untry
with
a
highe
r
num
ber
of
acci
den
t
s
[1
]
.
T
he
m
os
t
prom
inent
reaso
n
f
or
the
l
oss
of
a
li
fe
du
ri
ng
a
n
acci
den
t
is
the
un
a
vaila
bili
ty
of
im
m
ediat
e
h
el
p
that
can
sav
e
a
per
s
on
'
s
li
fe
by
a
few
seconds
.
The
m
om
ent
an
acci
den
t
has
occ
urre
d,
t
he
li
fe
of
al
l
pass
eng
e
rs
travell
i
ng
in
the
ve
hi
cl
e
is
at
sta
ke.
It
all
dep
e
nds
on
res
pons
e
ti
m
e
tha
t
can
save
t
heir
li
ves
by
a
fe
w
m
inu
te
s
or
s
econds
.
Acc
ordin
g
to
the
sta
t
ist
ic
s,
r
ed
ucin
g
acci
de
nt
delay
tim
e
by
even
1
m
inu
te
can
save
6
per
cent
of
li
ve
s.
Hen
ce
,
this
r
esp
on
se
ti
m
e
i
s
ver
y
cru
ci
al
,
an
d
it
needs
to
be
re
du
ce
d
or
at
le
ast
ei
ther
i
m
pr
ov
e
d
to
sa
ve
their
li
ves
[
2].
To
co
ntri
bu
te
to
our
so
ci
et
y
and
reduce
the
nu
m
ber
of
acci
de
nts
happe
ning
in
our
day
to
day
li
fe,
the
re
are
se
ver
al
te
ch
niqu
es
an
d
m
echan
ism
s
that
can
dro
p
down
the
rate
of
acci
den
ts
an
d
can
sa
ve
l
ot
l
ives.
Li
ving
in
a
te
ch
w
or
l
d
t
hat
is
grow
i
ng
day
by
day
with
ne
w
te
chnol
og
ie
s,
we
can
a
pply
these
te
chn
i
qu
e
s
in
o
ur
s
oc
ie
ty
and
help
the
m
ov
e
rc
om
e su
ch
problem
s.
The
Visio
n
of
the
In
te
rnet
of
Things
(
Io
T
)
has
com
e
ou
t
to
reach
unex
pected
bo
unds
of
tod
ay
'
s
com
pu
ti
ng
world.
It
is
a
co
ncep
t
t
hat
not
on
ly
ca
n
im
p
act
hu
m
an'
s
lif
e
but
al
so
ho
w
they
f
unct
ion
[
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
Io
T
base
d
c
ar
accide
nt det
ect
ion
and n
otif
ic
ation al
gorit
hm for
gener
al r
oad accide
nts (
Sh
iv
an
i
Sharm
a)
4021
The
hea
rt
of
I
OT
is
sm
art
sens
or
s
wit
hout
wh
ic
h
it
wou
ld
no
t
ha
ve
e
xisted.
T
hese
sens
or
s
f
orm
a
vast
netw
ork
for
t
he
ir
com
m
un
ic
at
ion
.
They
ca
pt
ur
e
m
inu
te
de
ta
il
s
of
t
heir
s
urrou
nd
i
ngs
a
nd
pass
t
his
im
portant
inf
or
m
at
ion
to
each
oth
er
.
Ba
sed
on
the
re
cei
ved
in
form
at
i
on
,
rele
van
t
act
ion
s
are
pe
rfor
m
ed
acc
ordin
gly.
It
is
the
la
te
st
com
m
un
ic
at
ion
m
od
el
that
i
m
agines
the
pr
ox
im
at
e
fu
ture
,
in
wh
ic
h
obje
ct
s
of
day
to
day
li
fe
will
be
inco
rpor
at
e
d
with
m
i
cro
c
ontrolle
rs
for
dig
it
al
co
m
m
un
ic
at
ion
with
the
hel
p
of
a
pprop
riat
e
p
r
oto
c
ol
sta
cks
that
will
m
ake
them
ca
pab
le
of
c
omm
un
ic
at
io
n
with
on
e
a
nothe
r.
It
is
a
te
chnolo
gy
that
aim
s
to
i
m
par
t
intel
li
gen
ce to d
evices s
o
that t
hey can
s
m
art
ly
co
nn
ect
and p
er
form
the n
ecessa
ry act
ion
s
to
el
i
m
inate
h
um
an
la
bour.
It
gi
ve
s
an
im
age
of
the
f
uture
w
here
non
-
li
ving
obj
ect
s
will
be
com
m
un
ic
at
ing
with
eac
h
oth
er
a
nd
do
i
ng
t
he
nee
dful
w
ork
.
I
n
th
is
way,
hum
an
la
bour
will
be
el
i
m
inate
d
to
an
exte
nt
an
d
the
de
vices
wi
ll
be
perform
ing
n
ec
essary act
io
ns
.
The
sig
nifican
ce
of
acci
den
t
detect
io
n
a
nd
noti
ficat
ion
s
yst
e
m
is
ver
y
prom
inent
f
or
our
s
ociet
y.
Im
agine
a
sit
ua
ti
on
wh
e
re
a
n
acci
de
nt
ha
pp
e
ne
d,
it
is
i
m
m
ediat
ely
no
ti
fied
t
o
the
e
m
erg
ency
se
r
vices.
This
will
resul
t
in
the
rescu
e
of
inju
red
pe
op
le
in
vo
l
ved
in
the
acci
den
t.
As
the
I
nt
ern
et
of
Thi
n
gs
has
witnesse
d
fast
grow
t
h
these
da
ys,
it
has
the
powe
r
to
br
i
dge
these
two
sit
uations
[4
]
.
F
or
the
Io
T
par
a
di
gm
be
eff
ect
ive
,
it
sh
ou
l
d
ha
ve
the
capab
il
it
y
to
track
the
locat
io
n
of
the
obj
ect
s
(i.e.
cars
in
our
case)
wh
ic
h
can
serv
e
to be
us
e
fu
l
for
t
he
am
bu
la
nces
to reac
h
the
locati
on
on tim
e [5
]
.
2.
RELATE
D
W
ORK
Ca
r
acci
de
nts
that
ha
ppen
dai
ly
are
the
m
ajor
soc
ia
l
pro
blem
s
towards
w
hi
ch
seri
ou
s
act
ion
m
us
t
be
ta
ken
.
On
e
of
t
he
so
l
ution
s
for
this
do
m
ai
n
i
s
the
In
te
r
net
of
Things
wh
ic
h
is
the
cur
re
nt
tren
d
in
te
ch
no
l
og
y.
Fo
r
this
pur
pos
e, m
any author
s h
a
ve w
orked
in this
dom
ai
n
by apply
ing t
hi
s tech
no
l
og
y.
Akrit
i
et
al
.,
[
6]
found
m
any
trade
offs
w
hile
work
i
ng
with
the
acci
de
ntal
m
anag
em
ent
syst
e
m
su
ch
as
high
c
os
t,
non
-
porta
bili
ty,
false
deliver
y
et
c.
The
sy
stem
faced
m
any
sho
rtcom
i
ng
s
due
t
o
la
ck
of
resou
rces.
I
n
their tec
hniq
ue, t
hey u
sed
seve
rity
scale
to
measur
e the im
pact o
f
an
acci
de
nt.
This r
e
duc
ed
loa
d
on
th
e
cl
oud
s
erv
e
r
by
30
pe
r
cent.
Chat
rapat
hi
et
al
.,
[7
]
desig
ne
d
a
fr
a
m
ewo
r
k
that
ha
s
two
com
ponen
t
s
.
First
one
is
acc
ident
detect
io
n
and
al
e
rting
s
yst
e
m
.
The
second
on
e
is
tra
f
fic
m
anag
e
m
e
nt
f
or
the
am
bu
la
nce.
The
e
ff
ic
ie
nt
routin
g
al
gorit
hm
is
us
ed
to
route
the
am
bu
la
nce.
T
he
te
chn
iq
ue
is
f
easi
ble
f
or
t
he
ro
a
d
j
unct
io
ns
with
sign
al
s.
Ho
we
ver,
it
is not a
ppli
cable t
o t
he se
gm
ents w
it
hout
sig
nals.
In
pap
e
r
[
8],
Ra
ut
and
Sach
dev
pro
posed
a
cal
l
no
ti
ficat
ion
syst
em
that
con
sist
s
of
XBee
W
iFi
Module
,
XBee
Sh
ie
ld,
GPS
Module
an
d
S
eeedu
i
no.
T
he
acci
den
t
is
de
te
ct
ed
us
in
g
on
ly
cras
h
se
ns
or
s
because
of
w
hich
it
gi
ves
le
ss
accu
rate
res
ul
ts.
Ali
an
d
Al
wan
[
9]
pro
po
sed
a
syst
e
m
t
hat
co
ns
ist
s
of
sever
al
cases t
o detec
t l
ow
s
pee
d
an
d hig
h
-
s
pee
d
car
acci
den
ts.
I
n
c
ase of h
i
gh
s
pe
ed,
if t
he
sm
artpho
ne'
s accel
erati
on
>
4G,
then
t
he
re
is
an
acci
den
t
wh
ic
h
is
ide
ntifie
d
by
the
sm
artph
one
'
s
app
li
cat
ion
.
Howe
ver,
it
lead
s
to
trigg
e
rin
g of fa
lse
alarm
in
few
cases
since
m
ob
il
es are th
e
sub
j
ect
.
In
pa
per
[
10
]
,
S.
R
et
al
.
disc
us
se
d
the
dri
ve
r'
s
beh
a
viour
by
analy
zi
ng
ey
e
blin
king
with
the
hel
p
of
IR
sens
ors.
T
he
head
m
ov
em
ents
of
the
dri
ver
a
re
m
on
it
or
ed
by
the
acc
el
ero
m
et
er
wh
ic
h
is
fixe
d
ont
o
the
fore
head
to
m
easur
e
the
a
ng
le
s
m
ade
by
the
hea
d.
T
his
te
ch
nique
is
no
t
feasible
since
it
w
ou
ld
be
un
c
om
fo
rtable
for
the
dri
ve
r
to
at
ta
ch
an
acce
le
ro
m
et
er
to
the
fore
he
ad
e
v
ery
ti
m
e
.
Mo
reover
,
dri
ver
beh
a
viou
r
is
the
only
factor
that
is
con
si
der
e
d
f
or
acci
den
t
detect
ion
.
Sandeep
et
a
l.,
[11]
intr
oduced
a
so
luti
on
f
or
t
he
acci
den
ts
tha
t
are
m
ajo
rly
cause
d
by
dri
nk
and
dr
i
ve
cas
e.
For
this
pur
po
s
e,
they
us
e
d
fe
w
sens
or
s
li
ke
to
uch
sen
sor,
he
artbeat
se
nsor,
al
cohol
se
nso
r
inter
face
d
wi
th
Ra
spber
ry
P
i.
In
their
w
ork,
t
hey
on
ly
c
onsidere
d
the
sit
uatio
n for the
drin
k
a
nd drive
cases.
In
pa
per
[12],
Pr
at
iks
ha
R
et
al
.,
de
velo
pe
d
a
syst
e
m
wh
ic
h
detect
s
a
n
ac
ci
den
t,
detect
s
the
co
ndit
i
on
of
t
he
car'
s
en
gin
e
a
bout
whic
h
us
e
r
is
in
f
or
m
ed
if
there
is
any
flam
e
or
sm
ok
e
is
de
te
ct
ed.
The
s
yst
e
m
eff
ect
ively
m
on
it
or
s
the
ov
e
r
al
l
abn
orm
alities
that
can
be
cause
d
in
a
car.
Howe
ver
,
the
syst
e
m
do
esn'
t
fo
cu
s
strongly
on
ac
ci
den
t
detect
io
n
par
t.
Kh
al
iq
et
al
.,
discusse
d
te
c
hn
i
qu
es
t
o
detect
the
ac
ci
den
t
by
us
in
g
a
fe
w
sens
or
s
a
nd
ot
her
hard
war
e
,
they
then
veri
fy
the
ge
ner
a
te
d
res
ults.
I
n
their
ap
proac
h,
they
c
hec
ke
d
the
sever
it
y
of
an
acci
den
t
[
13]
.
Nam
rata
H
et
al
.,
detect
the
acci
den
t
by
a
detect
i
on
unit
that
is
fitt
ed
in
the
car.
The
a
uthors
i
m
ple
m
ented
this
unit
as
a
pu
s
h
on
switc
hes
w
hich
sen
ses
any
obsta
cl
es
and
tri
ggers
the
m
ic
ro
co
ntro
ll
e
r
(
AT89
52)
le
a
ding
to
im
m
ediat
el
y
turn
on
the
buzze
r.
H
oweve
r,
t
his
te
chn
i
qu
e
m
ay
not
wo
r
k
ever
y t
im
e b
ecau
se i
n
a
few s
it
uations
t
he dr
iver
m
ay
n
ot
be
ab
le
to
tu
rn on s
witc
h [14].
Yad
a
v
et
al
.,
[
15
]
detect
the
acci
den
t
an
d
noti
fy
the
cause
to
the
re
gister
ed
num
ber
.
I
n
their
w
ork
,
they
are
re
porting
a
n
a
cci
de
nt
to
a
s
pecific
nu
m
ber
a
nd
not
to
a
n
em
erg
ency
ser
vi
ce.
Howe
ver,
the
work
is
sti
ll
un
poli
sh
e
d
du
e
to
la
ck
of
res
ources
a
nd
im
ple
m
enta
ti
on
.
I
n
pap
e
r
[16],
Re
dd
y
a
nd
Ra
o
de
velo
ped
a
syst
e
m
wh
ic
h
is
us
e
d
to d
et
ec
t
cal
a
m
ities
li
ke
fire
in
t
he
ca
r
.
T
he
pro
po
se
d
m
et
ho
do
l
og
y deli
ver
s
g
oo
d
s
afety
.
It r
es
ults in
wa
r
ni
ng
s
that ca
n be
perform
ed
to tri
gg
e
r pr
e
ve
ntive m
easur
es
in
case
of s
uc
h
inci
den
ts
.
Kav
ya
a
nd
Ge
et
ha
sugg
e
ste
d
te
chn
iq
ues
to
m
ini
m
iz
e
the
delay
tim
e
cau
sed
by
am
bu
la
nce
to
reac
h
the
locat
io
n
of
the
acci
de
nt
an
d
rat
her
t
o
pro
vid
e
a sm
oo
th f
lo
w
of
em
erg
e
n
cy
ve
hicle
s
usi
ng
RF
Tech
no
log
y.
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.
5
,
Oct
ober
201
9
:
4
0
2
0
-
4
0
2
6
4022
They
ha
ve
a
ddresse
d
an
e
ff
ic
ie
nt
routin
g
al
gorithm
to
ro
ut
e
veh
ic
le
s
[
17]
.
Pall
avi
and
K.
Wagh
f
oc
use
d
on
intel
li
gen
tl
y
plann
i
ng
the
tra
nsporta
ti
on
syst
e
m
based
on
RF
te
chnolo
gy
to
re
du
ce
ov
e
rc
rowd
i
ng
of
ve
hicle
s
in locali
t
ie
s w
it
h
sm
art con
tro
l of
sig
nals and
a p
rope
r
pat
h
is plan
ned
with
the h
el
p of
a
n
andr
oid
appli
cat
ion
.
Au
t
hors
dev
el
op
e
d
a
real
-
ti
m
e
al
go
rithm
t
hat
m
akes
us
e
of
VANE
T
co
m
m
un
ic
at
ion
to
av
oid
ve
hicle
s
from
traff
ic
-
relat
ed c
ongestio
n. F
or s
om
e j
un
ct
i
ons,
it
is
not fea
sible [
18
]
.
Poor
a
ni
k
et
al
.,
[
19
]
discuss
e
d
ab
out
the
use
of
a
j
am
m
er
ci
rcu
it
w
hich
di
sables
the
key
pad.
It
use
s
i
m
age
processi
ng
te
c
hniq
ues
to
detect
dri
ve
r
be
ha
viour
a
nd
se
ns
ors
t
o
de
te
ct
the
acci
de
nt.
Kim
and
Jeo
ng
[20]
pro
pose
d
an
al
gorithm
for
detect
ing
cras
h
usi
ng
crash
pro
ba
bili
ti
es
data.
Th
e
pro
posed
al
gorithm
sh
owe
d
a
n
im
pr
ovem
ent o
ve
r M
ote
-
Ca
rlo
si
m
ula
ti
on
whic
h gav
e
ef
fecti
ve
r
es
ults o
f
thei
r
m
od
el
.
3.
PROP
OSE
D
METHO
D
The
sig
nifica
nc
e
of
def
i
ning
the
re
searc
h
pro
blem
is
to
ad
dr
ess
t
he
ga
ps
in
t
he
li
te
ratur
e.
The
pu
rpose
is
to
con
t
rib
ute
to
the
e
xisti
ng
work
t
o
en
ha
nc
e
the
qual
it
y
of
the
overall
fr
am
ewo
r
k
s
o
that
it
can
ben
e
fit
the
en
d
soc
ie
ty
in
fu
t
ur
e
.
T
his
c
an
be
achie
ve
d
by
ad
ding
m
or
e
f
unct
ion
al
it
ie
s
an
d
feat
ur
es
that
can im
pr
ove th
e wor
king
of th
e en
d
syst
em
.
Dh
a
nlak
sh
m
i
and
Le
ni
[21]
design
e
d
a
syst
e
m
that
m
o
nitor
s
the
c
onditi
on
of
the
car
durin
g
it
s
j
ou
rn
ey
.
T
he
pa
ram
et
ers
that
are
a
ddresse
d
in
their
w
ork
are,
gas
le
aka
ge
w
hich
is
m
on
it
or
e
d
by
us
i
ng
an
MQ2
gas
se
nsor
,
veh
ic
le
s
pee
d
w
hich
is
rec
orded
by
hall
-
ef
fect
se
nsors
,
GP
S
a
nd
GS
M
m
odul
es
for
com
m
un
ic
at
ion
an
d
trac
king
locat
ion
of
ve
hicle
s.
H
owev
er,
f
or
an
acci
den
t
detect
ion
case,
only
sp
e
ed
ha
s
been
co
ns
ide
re
d
by
m
aking
us
e
of
hall
-
ef
fect
sens
ors.
More
ov
e
r,
Pin
and
W
a
ng
[
22]
pro
posed
a
veh
ic
le
colli
sion
detec
ti
on
al
gorithm
wh
ic
h
w
orks
well
fo
r
T
-
in
te
rsecti
on
r
oa
d
desi
gn.
The
par
am
et
ers
that
are
consi
der
e
d
f
or
the
design
of
the
al
go
rit
hm
are,
cu
rv
at
ur
e
area
of
T
-
inter
sect
ion
ju
nctio
ns
an
d
the
pr
e
dicte
d
tim
e
fo
r
th
e
tw
o
ca
rs
to
m
eet
at
the
ju
nctio
n.
W
e
feel
that
t
he
al
go
rithm
i
s
ef
f
ect
ive
f
or
the
s
pecific
ca
se
of
T
-
intersect
i
on
and
not
f
or
ge
ner
al
ro
a
d
acci
den
ts
.
T
her
e
fore,
the
re
is
a
ne
ed
f
or
m
od
ify
ing
t
he
existi
ng
w
or
k
done by a
uthor
s to
s
uppo
rt the
g
e
ner
al
ro
a
d a
cci
den
ts.
In
o
ur
a
ppr
oac
h,
we
are
a
ddr
essing
t
he
gaps
by
ad
ding
a
n
acce
le
ro
m
et
er,
vibrat
io
n
se
nsor
a
nd
m
os
t
i
m
po
rtantl
y
he
artrate
sens
or.
These
c
om
po
ne
nts
co
ntribute
to
the
hard
wa
re
set
up
of
t
he
syst
e
m
.
Also
,
we
would
li
ke
to
intr
oduce
a
n
al
gorithm
fo
r
ge
ner
al
ro
a
d
ac
ci
den
ts
that
is
a
ppr
opriat
e
f
or
this
ha
rdwa
re
set
up
.
We
ha
ve
c
on
sidere
d
a
fe
w
pa
ram
et
ers
wh
ic
h
are
help
fu
l
for
a
cci
d
ent
detect
ion
a
nd
noti
f
ic
at
ion
.
These p
aram
et
ers
are v
e
hicle
acce
le
rati
on
,
re
ta
rd
at
io
n,
c
ras
h
im
pact,
the
va
lue
of h
eart
ra
te
sens
or
(
em
bed
de
d
within
the
be
lt
)
an
d
inf
or
m
ation
of
acci
de
nt
locat
ion
w
hic
h
is
tracked
by
GP
S.
It
is
the
n
sent
to
em
erg
ency
serv
ic
es/
fam
il
y
m
e
m
ber
s b
y G
SM com
m
un
ic
at
ion
.
We
im
ple
m
ent
ed
the
syst
em
by
de
sig
ning
a
n
IO
T
ba
sed
c
ar.
T
he
ca
r
is
e
m
bed
de
d
i
n
Ardu
i
no
as
a
dev
el
op
m
ent
boar
d
w
hich
is
interface
d
with
dif
fer
e
nt
se
nsors
as
li
ste
d
a
bove.
It
is
co
nt
ro
ll
ed
via
Bl
ue
too
t
h
m
od
ule
HC05.
Also
,
t
he
car
is
te
ste
d
for
di
ff
e
ren
t
c
onditi
on
s
t
o
see
k
re
su
lt
s.
F
or
this
set
up,
the
al
go
rithm
op
e
rates
on
the
data
gather
e
d
by
acce
le
ro
m
ete
r
ADXL
345,
vibrat
ion
se
nso
r,
hea
rtrate
sen
so
r
,
GPS
and
GS
M
m
od
ule.
These
sensors
ha
ve
t
heir
co
nf
i
gurat
ion
s
a
nd
th
res
hold
ra
nge.
T
he
acce
le
ro
m
et
er's
input
range
c
an
be
2g
to
200g
(ne
gative
a
nd
pos
it
ive)
an
d
it
ca
n
var
y
e
ve
n
m
or
e
[
23
]
.
Wher
eas,
the
vi
br
at
i
on
sen
sor
has
on
ly
two
sta
te
s,
lo
w
an
d
hi
gh.
I
t
is
low
for
norm
al
cases.
On
e
xperienci
ng
a
la
r
ge
im
pact
force
f
rom
the
env
i
ronm
ent, it
b
ecom
es h
igh
[24]. Th
e
hear
t
r
at
e sens
or is the essentia
l co
m
po
nen
t si
nce
it
k
eeps track
of the
dr
i
ver
'
s
hear
t
be
at
s
durin
g
the
j
ou
rn
ey
.
N
orm
al
l
y,
the
hea
r
t
rate
for
a
person
is
betwe
en
75
-
170
bp
m
fo
r
t
he
gro
up of
peopl
e b
et
wee
n 2
0
a
nd 50 yea
rs [25]
.
Fi
gure
1
is t
he bloc
k diag
r
a
m
f
or the
syst
e
m
:
Figure
1. Bl
oc
k
diag
ram
o
f
pro
posed
syst
em
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
Io
T
base
d
c
ar
accide
nt det
ect
ion
and n
otif
ic
ation al
gorit
hm for
gener
al r
oad accide
nts (
Sh
iv
an
i
Sharm
a)
4023
The o
ver
al
l m
od
el
inclu
des
th
e
f
ollow
i
ng com
po
nen
ts
:
a.
Ardu
i
no
:
This
is
the
cor
e
unit
of
the
entire
s
yst
e
m
as
it
cont
ro
ls
the
flo
w
of
in
form
at
ion
betwee
n
sens
ors.
It
is
basical
ly
a
dev
el
opm
ent
bo
ar
d
w
hich
gi
ves
the
flexibi
li
ty
of
wr
it
in
g
C
pr
ogram
s
for
the
sens
or
s
a
nd
la
te
r
they
can
be de
plo
ye
d i
n t
he fla
sh
m
e
m
or
y o
f
Ardu
i
no t
o
c
heck the
f
unct
ion
in
g o f se
ns
ors.
b.
Vibrat
ion
Se
nsor:
This
sens
or
can
rec
ognize
vibrat
ion
s
in
a
giv
e
n
a
rea.
It
ha
s
tw
o
values
as
lo
w
a
nd
high.
Usu
al
ly
,
it
rem
ai
ns
lo
w
f
or
t
he
scenari
os
where
vi
br
at
io
n
i
m
pact
is
no
t
that
powerfu
l.
It
at
ta
ins
high
va
lue
on r
ecei
ving
hig
h vi
br
at
io
ns
f
ro
m
the envir
onm
ent.
c.
Accele
r
om
et
er
:
It
is
a
ty
pe
of
sens
or
w
hich
is
design
e
d
to
m
easur
e
acce
le
rati
on
acc
ur
at
el
y.
It
m
easur
e
s
acce
le
rati
on
in
three
axis
wh
ic
h
are
x
-
directi
on,
an
d
z
-
directi
on.
The
x
-
a
xis
of
the
acce
le
r
o
m
et
er
giv
es
the
m
easur
e
of
posit
ive
acce
le
ra
ti
on
,
y
-
axis
gi
ves
t
he
m
easur
e
of
ne
gative
acce
le
rati
on
(
retard
at
i
on)
a
nd
z
-
axis i
nd
ic
at
e
the angle
of tu
r
nove
r of
t
he de
vice in
w
hich
it
is instal
le
d.
d.
Hear
t
r
at
e
sens
or
:
T
he
hear
t
r
at
e
sensor
is
ba
sed
on
t
he
pr
i
nciple
of
phot
ople
thysm
og
raphy.
It
is
desig
ne
d
to m
easur
e the
change i
n
th
e
volum
e o
f blo
od
. I
t
kee
ps
a t
rac
k of t
he per
son
’s heart
beat.
e.
Global
P
os
it
ion
in
g
Syst
em
(G
PS
):
A
globa
l
po
sit
io
ning
s
ens
or
is
a
rece
iver
wh
ic
h
give
s
posit
ion,
s
pe
ed
and
ti
m
ing
inf
or
m
at
ion
of
a
n
obj
ect
.
On
in
s
ta
ll
at
ion
of
this
sens
or
,
a
ny
de
vice
can
be
tra
cked
t
o
locat
e
it
s
po
sit
io
n.
f.
GS
M:
It
is
a
c
om
po
ne
nt
w
hi
ch
is
use
d
for
m
ob
il
e
to
m
ob
il
e
com
m
un
ic
a
ti
on
.
It
is
res
pons
ible
f
or
se
ndin
g
SMS to t
he des
ired n
um
ber
or
m
aking
a call
wh
e
ne
ve
r
instr
ucted.
g.
Ce
ntral
Ser
ve
r
:
On
ce
a
n
acci
den
t
is
detect
ed,
t
he
cent
ral
serv
e
r
is
im
mediat
el
y
infor
m
ed
ab
ou
t
it
.
It
is
respo
ns
ible
for
locati
ng nea
r
by
a
m
bu
la
nces
that can
r
eac
h
t
he
acci
de
nt l
oc
at
ion
.
3.1.
Prop
os
ed
al
gori
th
m
The
m
ai
n
fun
ct
ion
in
g
beh
i
nd
the
pro
pose
d
syst
em
is
t
he
ge
ne
rali
zed
acci
den
t
detect
ion
a
nd
no
ti
ficat
io
n
al
gorithm
that
tak
es
diff
e
re
nt
inputs
int
o
acc
ount
a
nd
base
d
on
t
hat
it
ge
ner
at
es
res
ults
that
ar
e
help
fu
l
for
det
erm
ining
the
s
ta
tus
of
the
propose
d
syst
em
.
T
o
gen
e
rate
intende
d
res
ult
s,
the
f
ollow
i
ng
a
re
consi
der
e
d:
-
Dep
l
oym
ent o
f
the
hard
war
e c
om
po
ne
nts in
e
ver
y ca
r.
-
Algorithm
w
or
ks
on
ly
for
t
he a
rea
wh
ic
h has
stron
g netw
orks.
-
It is only
appli
cable t
o ca
rs
.
-
Highway
jun
ct
ion
is
not c
on
si
der
e
d
-
On
ly
cases
for
po
s
sible c
rash
are c
on
si
der
e
d
.
-
Dr
i
ver
m
us
t
wear
s
eat
belt
each
ti
m
e
to
recor
d
the
he
artbeats
since
hear
trat
e
se
nsor
is
em
bedde
d
i
n
seat
belt.
Fo
ll
owin
g
a
re t
he
cases
that a
r
e co
ns
ide
red f
or a
n
acci
de
nt a
nd it
s ch
a
nces:
a.
Ca
se 1
:
Warni
ng to A
vo
i
d A
cci
den
t
In this case
, the
driver
is ale
rt
ed fo
r ov
e
rs
pe
edin
g
b.
Ca
se 2
:
Wh
e
n
t
he
ca
r
is st
at
ic
This
case
de
picts
a
scena
rio
for
a
po
s
sible
crash
w
he
n
th
e
car
is
at
rest
.
The
dri
ve
r
in
side
the
ca
r
cou
l
d be in
jure
d base
d on the
value give
n by
the h
ea
rt ra
te
s
ens
or
.
c.
Ca
s
e 3
:
Wh
e
n
t
he
ca
r
is st
at
ic
,
and the
drive
r i
s not inside
This
case
de
pi
ct
s
a
sit
uatio
n
wh
e
n
t
he
ca
r
is
at
rest,
but
the
dri
ver
is no
t
i
nsi
de.
T
his
is
al
so
a
case for
an
acci
de
nt,
bu
t for s
uc
h
cases
, em
erg
ency se
rv
ic
es
nee
d no
t
b
e i
nfor
m
ed.
d.
Ca
se 4
:
Wh
e
n
t
h
e ca
r
is m
ov
in
g
It
is
the
m
os
t
com
m
on
case
wh
e
n
a
m
ov
in
g
car
gets
hit
by
an
oth
e
r
ve
hicle
.
For
s
uc
h
sit
uatio
ns
,
e
m
erg
ency
ser
vices
m
us
t
be
prom
pted
f
or
r
escue.
It
is
the
m
os
t
co
m
m
on
case
w
he
n
a
m
ov
i
ng
car
gets
hit
by
ano
t
her
ve
hicle
.
F
or
su
c
h
sit
uations,
em
ergency
ser
vices
m
us
t
be
prom
pted
f
or
resc
ue
.
Ta
ble
1
de
pi
ct
s
the
above
sta
te
d
ca
ses.
Table
1.
Senso
r
Ra
nges
to det
erm
ine an
A
cci
den
t
Vib
ration
Sen
so
r
Acceler
o
m
ete
r
(
m
/s
^2)
Hear
t
-
rate
Sen
so
r(
b
p
m
)
Inf
erence
Low
Peak
Value
No
r
m
alRan
g
e
Ov
er
Sp
eed
Hig
h
0
Peak
Value
Acciden
t
Hig
h
Hig
h
0
Peak
Value
0
Peak
Value
Acciden
t
Acciden
t
To
desi
gn
t
he
al
gorithm
,
the
peak
values
of
acce
le
ro
m
et
er
and
hear
t
rate
sensor
are
ke
pt
in
m
ind.
Accor
ding
to
the
data
sh
eet
s
for
sens
or,
the
peak
value
f
or
hear
t
rate
sen
s
or
is
170
bp
m
and
a
bove.
T
he
peak
value o
f
acce
le
ro
m
et
er is b
et
ween
-
150 t
o
-
200
i
n
case
of
retard
at
i
on.
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.
5
,
Oct
ober
201
9
:
4
0
2
0
-
4
0
2
6
4024
Let
v
ib
r,
a
cc a
nd h
ea
rtrate a
r
e the
values
of
the v
i
br
at
io
n s
ens
or
,
accel
er
om
et
er an
d hea
r
t beat se
ns
or.
Let
p
t1
d
e
note
peak val
ue fo
r heart
rate se
nsor
and
pt2 de
note
p
ea
k val
ue a
ccel
ero
m
et
er.
1.
Rece
i
ve
in
put f
ro
m
the se
nsors
.
2.
Process
sens
or in
pu
t
\
*
W
a
r
ning to
avo
i
d
acci
den
t
*
\
3.
W
hile
(
vibr
== LO
W)
{
If
(acc ==
pt2
)
{
“Alert f
or
ov
e
r
sp
ee
ding
of v
e
hicle
”
}
}
/*
Wh
e
n
ca
r ha
s ei
ther hig
h va
lue for
r
et
ar
dat
ion
or
wh
e
n
ca
r
is
no
t m
ov
in
g an
d
sti
ll
ther
e
is a cras
h */
4.
If
(
hear
t
rate
> =pt1
)
{
/*f
or
m
ov
in
g
c
ar (su
dden
cras
h resu
lt
s i
n
reta
rd
at
io
n) ||
stat
ic
car
*/
If
(acc
==
p
t
2
||
acc==
0
)
{
“Sen
d
locat
i
on to
cen
tr
alised server
”
“Sen
d
SMS
/E
m
erg
ency cal
l
to f
am
il
y”
}
}
5.
Exit
4.
RESU
LT
S
A
ND AN
ALYSIS
The
syst
em
is
sim
ulate
d
us
ing
A
r
du
i
no
I
DE
as
a
t
oo
l
t
o
gen
e
rate
te
st
res
ults
f
or
ea
ch
se
nsor
by
giv
in
g
a
n
i
nput
value
to
it
.
T
o
us
e
t
his
to
ol,
i
t
is
require
d
to
us
e
a
pro
gram
m
able
ci
rcu
it
c
omm
on
ly
kn
own
as
a
m
ic
ro
co
ntr
oller whic
h
is t
he sole com
ponent
f
or
t
his to
ol. The c
od
e is
wri
tt
en
fo
r
se
nso
r
s in
C p
rogr
am
m
ing
la
nguag
e
i
n
A
rduin
o
I
DE
a
nd
it
is
up
l
oad
e
d
in
the
flas
h
m
e
m
or
y
of
the
m
ic
r
ocontr
oller
to
te
st
the
s
ens
or
.
The dat
a
gen
e
r
at
ed
by t
he
se
nsor
can
b
e
an
al
yz
ed
in t
he o
utp
ut
scree
n of t
he Ar
duin
o
I
D
E.
4.1.
Simul
at
i
on
f
or
st
at
ic
c
ar ac
ci
dent
In
this
case
,
there
can
be
a
possi
ble
crash
w
he
n
t
he
ca
r
i
s
at
rest
an
d
dri
ve
r
is
i
ns
ide
.
The
acce
le
ro
m
et
er
will
gi
ve
va
lues
lo
w
valu
e
or
m
os
tl
y
0m/
s^2.
T
he
vibra
ti
on
se
nsor
wil
l
switc
h
from
l
ow
to
high,
t
he
m
ome
nt
it
ex
pe
rien
ces
a
c
rash
with
la
r
ge
r
im
pact
.
Ta
ble
2
dep
ic
ts
the
nu
m
erical
values
re
spo
ns
ible
for
an
acci
de
nt
.
From
the
Figu
re
2
,
it
can
be
inferred
that
hear
t
rate
sen
s
or
giv
es
peak
value
f
or
hear
t
beats
wh
e
n
the
acce
le
rati
on
is
0.
This
m
eans
th
at
dr
iver
is
no
t
in
healt
hy
con
diti
on.
Als
o,
wh
e
n
bot
h
the
sensor
giv
es
0 val
ue
t
hen that m
eans th
e
dr
i
ver
is
not insi
de
the
ca
r
a
nd that is
w
hy the
hea
rt r
at
e senso
r
is
0.
Table
2.
Senso
r
rea
dings
for s
ta
ti
c car accide
nt
Vib
ration
Sen
so
r
Acceler
o
m
ete
r
(
m
/s
^2)
Hear
t
-
rate
Sen
so
r(
b
p
m
)
Inf
erence
1
0
190
Acciden
t
1
1
1
0
0
0
170
195
185
Acciden
t
Acciden
t
Acciden
t
1
0
0
Crash
Figure
2
.
Grap
h betwee
n
acce
le
ro
m
et
er an
d
hear
trat
e
sen
s
or
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
Io
T
base
d
c
ar
accide
nt det
ect
ion
and n
otif
ic
ation al
gorit
hm for
gener
al r
oad accide
nts (
Sh
iv
an
i
Sharm
a)
4025
4.2.
Simul
at
i
on
f
or
mo
vin
g
c
ar acci
den
t
This
is
a
sp
eci
fic
case
depi
ct
ed
fo
r
a
m
ov
i
ng
car
.
Wh
en
the
car
m
eet
s
with
an
acci
den
t
the
acce
le
ro
m
et
er
will
ex
per
ie
nc
e
a
ce
rtai
n
a
m
ou
nt
of
retard
at
io
n
(
neg
at
i
ve
acce
le
r
at
io
n).
At
this
m
om
ent,
vibrat
ion
se
nsor
s
witc
hes
f
r
om
low
to
hi
gh
sta
te
.
There
are
sit
uations
wh
e
re
the
dri
ve
r
gets
in
j
ure
d
du
e
to
i
m
pact
of
cras
h
beca
us
e
of
wh
ic
h
there
will
be
a
dr
ast
ic
change
in
the
dr
i
ver’s
hea
rtb
eat
.
Table
3
re
pr
ese
nts
the abo
ve
sta
te
d
sce
nar
i
o.
Table
3.
Senso
r
rea
dings
for
m
ov
ing
ca
r
acc
ident
Vib
ration
Sen
so
r
Acceler
o
m
ete
r
(
m
/s
^2)
Hear
t
-
rate
Sen
so
r(
b
p
m
)
Inf
erence
0
130
100
Ov
er
Sp
eed
1
-
150
190
Acciden
t
1
1
1
-
180
-
170
-
200
170
195
185
Acciden
t
Acciden
t
Acciden
t
1
-
195
200
Acciden
t
Figure
3
repre
sents
the
gra
ph
bet
ween
acc
el
ero
m
et
er
and
hear
trat
e
se
nsor
.
Acc
ordin
g
to
the
grap
h,
wh
e
n
t
he
ca
r
e
xp
e
riences
hi
gh
reta
rd
at
io
n
due
to
cras
h,
th
e
dr
i
ver’s
hea
rt
beats
raise
up
dr
ast
ic
al
ly
ind
i
cat
ing
the cau
se
of
a
n i
njury
. T
his situat
ion cal
ls f
or an
im
m
ediat
e
help f
or
a
n
am
bu
la
nce
.
Figure
3
.
Grap
h betw
ee
n
acce
le
ro
m
et
er an
d
hear
t
rate se
nsor
No
te
that
the
values
of
hear
t
rate
sens
or
va
ry
accor
ding
to
the
age
of
a
per
s
on
an
d
it
has
not
bee
n
us
e
d
f
or
sim
ulati
on
.
On
ly
vi
br
at
io
n
se
nsor
and
acce
le
ro
m
et
er
ha
ve
bee
n
con
si
der
e
d
for
te
sti
ng
.
Howe
ver,
in
the
bot
h
the
ta
bles
on
ly
th
ose
cases
a
re
de
picte
d
t
hat
re
qu
i
res
t
he
nee
d
for
a
wa
r
nin
g
to
the
dri
ve
r
or
e
m
erg
ency cal
l
to
a
n
am
bu
la
nc
e.
5.
CONCL
US
I
O
N
The
m
ai
n
idea
of
t
his
pa
pe
r
is
to
noti
fy
the
co
nce
rn
e
d
auth
or
it
ie
s
ab
out
an
acci
den
t
on
ly
if
t
he
passe
ng
e
rs
are
injur
e
d.
T
he
pr
opos
e
d
fr
am
ew
ork
is i
ntend
e
d t
o
so
lve the sa
m
e b
y i
nco
rporat
ing
m
or
e featur
es
in
the
al
re
ady
existi
ng
w
ork
done
by
the
au
thors.
With
t
he
ad
dit
ion
of
a
bove
discu
ssed
functi
onal
it
ie
s,
this
syst
e
m
can
reso
lve
m
os
t
of
the
acci
de
nt
scenari
os
by
det
ect
ing
acci
de
nts
on
ti
m
e
and
trigg
e
rin
g
im
mediat
e
help
f
r
om
e
m
e
rg
e
ncy
ser
vice
s
without
wast
ing
a
ny
tim
e.
More
ov
e
r,
the
dr
i
ver’s
healt
h
is
being
trac
ke
d
by
hear
t
rate
se
nsor
(em
bed
de
d
in
seat
belt)
w
hi
ch
ser
ves
as
t
he
ad
de
d
ad
va
ntage.
If
im
plem
ented
with
pro
per
plan
ning
a
nd
r
eso
ur
ces
,
this
f
ram
ewo
rk
c
ou
l
d
se
rv
e
to
be
a
great
he
lp
t
o
t
he
s
ociet
y.
H
e
nce,
there
is
ne
ed
of
su
c
h
syst
em
s t
hat could
sa
ve t
he
li
ves
i
n
vol
ved w
it
h ac
ci
de
nts.
REFERE
NCE
S
[1]
Andrea
Z
and
L
ore
nzo
V
.,
“
Internet
of
Thi
ngs
f
or
Sm
art
Cit
ie
s
,”
IEEE
Inte
rne
t
o
f
Things
Journal
,
v
ol
/
issue:
1
(
1
)
,
Feb
2014
.
[2]
Isna
K
.
and
S
.
D
.
Sawant
,
“
Int
egr
ation
of
Cloud
Com
puti
ng
a
nd
Inte
rne
t
of
T
hings
,
”
Inte
rnati
onal
Journal
of
Adv
anc
ed Re
sea
rch
in
Comput
er
and
Comm
unication Engi
ne
erin
g
,
v
ol
/i
ss
ue:
5
(
4
)
,
Apr
2016
.
[3]
Sonali
D
.
T
.,
“
Cloud
Com
puti
ng
and
Softw
are
-
Based
Int
ern
et
of
Thi
ngs
,”
Int
ernati
onal
Jour
nal
of
Adv
an
ced
Re
search in Co
mputer
Scienc
e and Sof
tw
are
En
gine
ering
,
v
ol
/i
s
sue:
6
(
4
)
,
Apr 2
014
.
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.
5
,
Oct
ober
201
9
:
4
0
2
0
-
4
0
2
6
4026
[4]
Jonatha
n
K
.,
“
Us
ing
Acti
ve
Queue
Mana
gement
to
As
sist
I
OT
A
ppli
c
at
ion
Flows
in
Hom
eBroa
db
and
Networks
,
”
2017
IEEE
In
te
r
net
o
f
Things
Jo
urnal
,
v
ol
/i
ss
ue:
4
(
5
)
,
Oc
t
2017
.
[5]
Pengfie
Z
.
,
et
al.,
“
Secur
e
Locat
ion
of
Thi
ngs(SL
OT)
:
Miti
gati
ng
Loc
al
Spoofi
ng
Atta
cks
in
Inte
rne
t
of
Thi
ngs
,
”
IEE
E
Inte
rn
et
o
f
Things
Journal
,
v
ol. 4, Dec 201
7
.
[6]
Akriti
S
.
,
et
al.,
“
Inte
lligent
Ac
ci
den
t
Mana
g
ement
S
y
s
te
m
usin
g
IoT
and
C
lou
d
Com
puti
ng
,”
2nd
Inte
rnat
ion
al
Co
nfe
renc
e
on
N
ex
t
Gene
ratio
n
Computing
Tech
nologi
es
,
Oct
20
16
.
[7]
C.
Chat
rap
at
hi
a
nd
N
.
R
.
Venka
te
sakum
ar
,
“
VANE
T
base
d
Inte
g
rat
ed
Fram
ework
for
S
m
art
Acc
ide
nt
Mana
geme
nt
S
y
stem
,”
Int
ernati
onal
Con
fe
ren
ce
on
Sof
t
-
Computing
and
Net
wo
rk
Sec
uri
ty
,
Feb
2015
.
[8]
Pri
y
a
l
R
.
and
V
a
ntha
na
S
.,
“
Car
Acc
ide
n
t
Notif
icati
on
S
y
stem
b
a
sed
on
Internet
of
Thi
ngs
,”
Int
e
rnational
Journ
al
of
Computer
Ap
pli
cation
s
,
v
ol
/i
s
sue:
107
(
17
)
,
De
c
2014
.
[9]
H
.
M.
Ali
and
Z
.
S.
Alwan
,
“
Car
Acc
ide
n
t
Det
ec
t
ion
and
Notifi
cat
ion
S
y
stem
Us
ing
S
m
art
p
hone
,”
Int
ernati
on
al
Journal
of
Computer
Sc
ie
nc
e
and
Mobile
Computi
ng
,
v
ol
/i
ss
ue:
4
(
4
)
, p
p
.
620
-
635
,
Apr 2015
.
[10]
Aishw
ar
y
a
S.
R.
,
et
al.,
“
An
IoT
Based
Acc
ide
n
t
Preve
nti
on
and
Tra
ck
ing
S
y
stem
for
Night
Drive
rs
,”
Inte
rnationa
l
Journal
of
Innov
ati
v
e
R
ese
arch
i
n
Computer
and
Comm
unic
ati
on
Engi
ne
ering
,
v
ol
/i
ss
ue:
3
(
4
)
,
Apr
2015
.
[11]
Koneti
S
.
and
Ponnam
R
,
“
No
vel
drunke
n
dri
ving
det
e
ct
ion
a
nd
pre
vent
ion
m
ode
ls
using
In
te
rne
t
of
thi
ngs
,”
Inte
rnational
Co
nfe
renc
e
on
Rec
ent
Tr
ends
in
E
l
ec
tri
cal
,
E
le
c
tronic
s and
Comput
i
ng
Techno
logi
e
s
,
2017
.
[12]
V.
K.
Pra
ti
ksha
and
R
aj
esh
G
.,
“
Propos
ed
Model
for
the
Sm
art
Acc
ide
nt
Det
ec
t
ion
S
y
st
em
f
or
Sm
art
Vehicl
es
using
Arduino
boar
d,
Sm
art
Sensors
,
GP
S
and
GS
M
,”
Inte
rnati
onal
Journal
of
Eme
rging
Tr
ends
and
Technol
ogy
in
Computer
S
cienc
e
,
2015
.
[13]
Kishw
er
K
.
,
et
a
l.
,
“
Protot
y
pe
of
Autom
at
ic
Acc
i
dent
Detect
ion
a
nd
Mana
gement
in
Vehic
ul
ar
En
vironment
Us
ing
VA
NET
and
IoT
,”
11th
Int
ernati
onal
Confe
r
enc
e
on
Soft
wa
re,
Knowle
dge
,
Information
Manage
ment
a
nd
Appl
ic
a
ti
ons
,
20
17
.
[
1
4
]
N
.
H
.
Sane,
et
al
.
,
“
Rea
l
T
ime
Vehic
le
Ac
ci
d
e
nt
Dete
c
ti
on
an
d
Tra
cki
ng
Us
in
g
GP
S
and
GSM
,”
Inte
rnation
al
Journal
on
Rece
nt
and
Inno
vat
io
n
Tr
ends
in
Com
puti
ng
and
Comm
unic
ati
on
.
[15]
Upendra
Y
.
an
d
Kam
al
kanna
n
,
“
Sm
art
Vehicle
Moni
tori
ng
S
y
stem
using
IOT
,”
Int
ernati
o
nal
Journal
for
Dev
e
lopmen
t
o
f Com
pute
r Sc
i
en
ce
and
Techno
lo
gy
,
v
ol
.
5,
I
-
3
,
SW
-
31
.
[16]
Sarishm
a
R
.
and
K
.
Ragh
av
,
“
Fire
Acc
ide
n
t
Det
ec
t
ion
and
Prev
ent
ion
m
onit
orin
g
S
y
stem
using
W
ire
le
ss
Sensor
Network
en
abl
e
d
Android
Appli
ca
t
ion
,”
Indian
J
ournal
of
S
ci
en
c
e
and
Te
chnol
og
y
,
v
ol
/i
ss
ue:
9(1
7),
Ma
y
2016
.
[17]
Kav
y
a
K.
and
Gee
tha
C
.
R
.,
“
Acc
ide
n
t
Detect
ion
and
Am
bula
nce
Rescu
e
usi
ng
Raspbe
rr
y
P
i
,”
Int
ernati
onal
Journal
of
Engi
n
ee
ring a
nd
Tech
nique
s
,
v
ol
/i
ss
ue
:
2
(
3
)
,
2016
.
[18]
P
.
A
.
Ta
rge
and
M
.
P
.
Satone
,
“
VA
NET
base
d
Rea
l
-
Ti
m
e
Inte
l
l
ige
nt
Tr
ansportation
S
y
stem
,”
Int
ernati
onal
Journ
al
of
Computer
Ap
pli
cations
,
v
ol
/i
s
sue:
145
(
4
)
,
Jul
2016
.
[19]
Poorani
K.,
et
a
l.
,
“
IOT
base
d
l
ive
stre
aming
of
Vehic
l
e,
Pos
it
i
on
Acc
id
ent
Pre
vent
ion
and
De
t
ec
t
ion
S
y
st
em
,”
Inte
rnational
Jo
urnal
of Re
ce
nt
Tr
ends
in
Eng
in
ee
ring a
nd
R
ese
arch
,
Mar
2017
.
[20]
Ta
ewung
K
.
and
Hy
un
Y
.
J
.,
“
A
Novel
Algorit
h
m
for
Crash
Det
ec
t
ion
Under
Gene
ral
Road
Sc
en
es
Us
ing
Crash
Probabil
ities
an
d
an
Inte
ra
ct
iv
e
Multi
ple
Model
Parti
cle
Filt
er
,
”
2014
IEE
E
Tr
ansacti
ons
on
Inte
l
li
gent
Tr
anspor
t
Syste
ms
,
v
ol
.
1
5
,
D
ec
2014
.
[21]
Dhanla
kshm
i
an
d
Ez
i
l
S
.
L
.,
“
In
stanc
e
Vehicle
Monitori
ng
and
Tra
ck
ing
with
I
nte
rne
t
of
Thi
ng
s
using
Arduino
,”
Inte
rnational
Jo
urnal
on
Smar
t
S
ensing
and
In
tell
ige
nt
Syste
ms
,
Sep
2017
.
[22]
Pin
W
.
and
Jun
hua
W
.,
“
A
Ve
hic
l
e
Coll
ision
Dete
c
ti
on
Algor
it
hm
at
T
-
shape
d
Inte
rsec
t
ions
base
d
on
Loc
ati
on
Based
Servic
e
,
”
Inte
rnational
Co
nfe
renc
e
on
Re
c
ent
Tr
ends
in
El
ec
tri
cal
,
E
lectronic
s
and
Compu
ti
ng
Technol
ogy
,
Oct
2015
[23]
D
.
S
.
Dim
a
and
D
.
Covaciu
,
“
Soluti
ons
for
acce
l
era
t
ion
m
ea
s
ure
m
ent
in
v
e
hi
cl
e
cra
sh
t
ests
,”
IOP
Confe
renc
e
Serie
s:Materia
ls Sc
i
ence
and
En
gine
ering
,
Mar
2
017
.
[24]
Him
anshu
C
.,
“
Rec
en
t
Tr
ends
of
Mea
surem
ent
and
Deve
lopment
of
Vibr
at
ion
Sensors
,”
IEEE
Sensor
Journal
,
2016
.
[25]
Shu
Y
.,
et
al
.
,
“
A
Press
ure
Sensing
Sy
st
em
for
H
ea
rt
Ra
te
Monito
ring
with
Pol
y
m
er
-
Based
Press
ure
Sensors
and
an
Anti
-
Inte
rf
ere
n
c
e
Pos
t
Proce
ss
in
g
Circ
u
it
,”
I
EEE
Sensor J
ournal
,
2015
.
Evaluation Warning : The document was created with Spire.PDF for Python.