Int
ern
at
i
onal
Journ
al of Inf
orm
at
ic
s
and
Co
m
munic
at
i
on
Tec
hn
olog
y (IJ
-
I
CT)
Vo
l.
6
,
No.
3
,
D
ece
m
ber
201
7
, pp.
139
~
145
IS
S
N:
22
52
-
8776
,
DOI: 10
.11
591/ijict
.
v6
i
3
.
pp
139
-
145
139
Journ
al h
om
e
page
:
http:
//
ia
esj
ou
r
nal.co
m/
on
li
ne/in
dex
.php
/
IJ
ICT
Driver
Dro
wsiness D
etection S
yste
m Using S
ensors
Kusum
a Kum
ari BM
*
1
,
Sampad
a Set
hi
2
,
R
amakant
h
Kum
ar
P
3
,
Ni
sha
n
t Kum
ar
4
,
A
tu
li
t Sh
an
k
ar
5
1
Depa
rte
m
ent of
Com
pute
r
Sci
en
ce
,
Univer
si
t
y
C
oll
eg
e
of
Scie
n
c
e,
Tumkur Univ
ersity
,
Tumakur
u,
Karn
at
ak
a
2,3,4,5
Depa
rtment
of
Inform
ation
Scie
nc
e and
Eng
ine
er
ing, R.
V.
C
oll
eg
e
of
Engi
n
e
eri
ng,
Bang
al
ore
,
Karn
at
ak
a
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
A
ug
1
4
th
,
20
1
7
Re
vised
Oct
19
th
, 201
7
Accepte
d
Nov
3
th
, 201
7
A
low
-
cost
and
sim
ple
distri
buted
sensors
m
odel
tha
t
is
par
ti
cu
larl
y
sui
ta
b
l
e
for
m
ea
suring
e
y
e
b
li
nk
of
the
drive
r,
acci
den
t
and
hand
pos
it
ion
on
a
stee
ring
whe
el
.
The
se
sensors
ca
n
be
used
i
n
aut
om
oti
ve
a
ct
iv
e
sa
f
e
t
y
s
y
stems
tha
t
aim
at
detec
t
ing
d
rive
r’
s
fa
t
igue,
a
m
aj
or
issue
to
pre
vent
ro
ad
ac
c
ide
nts.
Th
e
k
e
y
point
of
thi
s
appr
oac
h
is
to
d
esign
a
pro
toty
p
e
of
sensor
unit
s,
so t
ha
t
it
c
an
serve
as
pl
at
f
orm
for
int
egr
ati
ng
diffe
re
n
t
kind
s of
sensors
int
o
the
st
ee
ring
whee
l.
Sin
ce
th
e
sensors
are
att
ac
hed
to
the
st
e
eri
ng
whee
l,
the
re
for
e
th
e
y
c
an’
t
b
e
de
tache
d
b
y
th
e
driv
er.
I
t
will
a
lso
detec
t
dang
ero
us
st
y
l
ish
drivi
ng
which
m
a
y
le
ad
to
fa
ta
l
acci
den
ts.
The
m
aj
or
dra
w
bac
k
is
that
the
e
y
e
bli
nk
se
nsors
fra
m
e
worn
b
y
th
e
dr
ive
r
ca
n
b
e
re
m
oved
ca
using
the
se
nsor
non
-
oper
at
ion
al
.
The
ou
tcom
e
is
tha
t
the
v
ibra
tor
at
t
ac
h
ed
t
o
e
y
e
b
li
n
k
sensor’s
fra
m
e
vibra
t
es
if
the
drive
r
shuts
his
e
y
es
for
appr
o
ximate
l
y
3
sec
onds
and
al
so
the
LCD
display
s
th
e
re
spec
t
ive
warni
ng
m
essage
.
The
whee
l
is
slowed
or
stopped
d
epending
on
the
con
dit
ion.
Th
is
is
a
c
companie
d
b
y
the
v
ehi
c
le
’s owner
bei
ng
notified
through
the
GS
M
m
odule
,
so
the
owner
ca
n
r
et
ri
eve
the
drive
r’s
loc
a
ti
o
n,
photogr
aph
a
nd
a
l
ist
of
n
earb
y
po
lic
e
stat
ions
through
an
andr
o
id
m
obil
e
appl
i
cation.
The
r
efo
re
,
driv
er
c
an
b
e
al
er
te
d
du
ring
dr
ows
ine
ss
and
th
e
owner ca
n
be
n
oti
fie
d
sim
ult
an
e
ousl
y
.
Ke
yw
or
d:
A
ccel
er
om
et
er
A
ndr
oi
d
D
r
owsiness
E
ye
b
li
nk se
nsor
E
ye
shut
S
ens
or
Copyright
©
201
7
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
:
Ku
s
um
a K
um
ari B.M
,
Dep
a
rtem
ent o
f
Com
pu
te
r
Sci
ence,
Un
i
ver
sit
y C
ol
le
ge
of
Scie
nce
, T
um
ku
r Un
i
ve
rsity
B.H
R
oad, T
um
aku
r
u, Ka
rn
a
ta
ka
57
2103, I
nd
ia
.
Em
a
il
: ku
su
m
a.kuku
@g
m
ai
l.
com
1.
INTROD
U
CTION
Dr
i
ver
fati
gue
is
a
sig
nificant
facto
r
in
an
a
m
ple
of
ve
hicle
acci
den
ts
.
Re
cent
sta
ti
sti
cs
est
i
m
at
e
that
annuall
y
1,2
00
deaths
an
d
76,
000
i
nju
ries
a
r
e
a
case
of f
at
igu
e
relat
ed
c
ra
sh
es
al
l
ove
r
th
e
w
or
ld
.
T
her
e
fore,
a
te
chnolo
gy
to
pr
e
ve
nt
an
d
de
te
ct
dr
ive
r
dro
wsin
ess
is
a
m
ajor
c
halle
nge
in
the
fiel
d
of
acci
den
t
a
vo
i
da
nce
syst
e
m
s.
Since
the
drow
si
nes
s
presents
a
ha
zard
on
the
r
oad,
the
refo
re
counteract
ive
m
et
ho
ds
nee
d
to
be
dev
el
op
e
d
t
o
de
al
with
it
s
ef
f
ect
s.
The
ai
m
of
t
his
proj
e
ct
is
to
de
velo
p
a
prototy
pe
f
or
drowsi
ness
det
ect
ion
syst
e
m
wh
ic
h
helps
preve
nt
ro
a
d
acci
den
ts
on
la
r
ge
scal
e.
T
he
f
oc
us
i
s
on
desig
ning
a
syst
em
that
will
accuratel
y
m
o
nitor
t
he
du
rati
on
of
ope
n
or
cl
os
ed
sta
te
of
the
dri
ve
r’
s
e
ye
s
in
real
-
ti
m
e.
By
m
on
it
or
i
ng
t
he
ey
es,
the
dri
ve
r
ca
n
be
al
e
rted
well
in
ad
va
nce
t
o
a
vo
i
d
ac
ci
den
ts
a
nd
t
his
is
po
s
sible
if
longer
ey
e
blin
ks
a
re
detect
ed.
T
he
analy
sis
of
fac
e
i
m
ages
is
a
r
esearch
a
rea
w
it
h
fo
ll
owin
g
f
eat
ur
es
a
nd
a
ppli
cat
ion
s
as
vi
rtual
too
ls
fa
ce
re
cogniti
on,
a
nd
hum
an
ide
ntific
at
ion
sec
ur
it
y
syst
e
m
s
bu
t
t
his
pr
oj
ect
fo
c
us
es
on
IR
transm
ittanceand
rece
ption
by
ey
e
blink
se
ns
or
a
nd
t
he
s
yst
e
m
is
design
ed
t
o
dete
rm
i
ne
w
hethe
r
the
ey
es
of
the drive
r
a
re
opene
d o
r
cl
os
e
d du
e
to fat
ig
ue
foll
ow
e
d by
acci
den
t
detect
ion
.
[1]
To
acc
omm
od
at
e
each
a
nd
e
ver
y
obj
ect
w
hi
ch
exists
i
n
th
is
wo
rld
or
li
ke
ly
to
exist
in
the
com
ing
fu
t
ur
e,
the
I
nt
ern
et
of
T
hing
s
is
the
ex
pa
nsi
on
of
t
he
c
urren
t
In
te
rn
et
s
erv
ic
es
w
hich
const
it
utes
this
.
T
his
repor
t
f
oc
us
es
on
the
per
s
pec
ti
ves,
chall
en
ge
s
and
opportu
niti
es
beh
in
d
a
fu
tu
re
inter
net
that
fu
ll
y
su
pport
s
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2252
-
8776
IJECE
V
ol.
6
,
No.
3
,
Decem
ber
20
1
7
:
139
–
145
140
the
“t
hin
gs”
,
al
ong
with
ho
w
these
things
can
help
in
the
de
sign
of
a
m
or
e
synerg
ist
ic
f
uture
In
te
r
net.
There
are
thin
gs
wh
i
ch
f
or
m
par
t
of
inter
net
of
th
ing
s
ha
ving
id
entit
ie
s
and
vir
tual
per
s
onal
it
ie
s
operati
ng
in
sm
art
sp
aces
us
i
ng
i
ntell
igent
inte
r
fac
es
t
o
c
onne
ct
an
d
c
omm
un
ic
at
e
within
s
ocial
,
e
nv
i
ronm
ental
,
an
d
us
er
con
te
xts. [2]
2.
LIT
ERATUR
E REVIE
W
Ther
e
a
re
var
i
ou
s
ways
to
carry
out
detect
ion
but
a
few
are
m
entioned
unde
r
this
sect
ion
su
c
h
as:
sensing
of
ph
ysi
olo
gical
ch
aracte
risti
cs,
sensing
of
dr
iv
er’
s
ey
es,
ve
hi
cl
e
resp
onse
and
m
on
it
or
i
ng
the
respo
ns
e
of
dri
ver
at
any
m
om
ent.
These
te
chn
iq
ues
ca
n
be
i
m
ple
m
ent
ed
in
two
w
a
ys:
by
accuratel
y
or
pr
act
ic
al
ly
m
e
asur
i
ng
c
hang
es
in
ph
ysi
ologica
l
sign
al
s,
su
c
h
as
ey
e
blink
in
g;
and
r
eal
tim
e
m
eas
ur
i
ng
ph
ysi
cal
c
hanges
s
uc
h
as
le
anin
g
of
the
dri
ver’s
hea
d
an
d
the
ope
n/
cl
os
ed
sta
te
s
of
the
ey
es.
Dr
i
ver’s
op
e
rati
on
a
nd
veh
ic
le
be
hav
i
our
ca
n
be
im
plem
ented
in
f
ollow
i
ng
ways
,
by
m
on
it
ori
ng
the
ste
eri
ng
wh
eel
m
ov
e
m
ents,
acce
le
ro
m
et
er
or
br
a
ke
patt
ern
s
,
ve
hicle
’
s
s
peed,
la
te
r
al
acce
le
rati
on
cause
d
a
nd
la
te
ral
disp
la
cem
ents
her
e
.
The
final
te
chn
iq
ue
is
by
m
on
it
or
in
g
t
he
res
ponse
of
the
dri
ve
r.
T
he
dr
i
ver
requir
es
to
consi
ste
ntly
sen
d t
he res
ponse
but eve
ntua
ll
y t
his b
ec
om
es
ti
reso
m
e and a
nnoyin
g for th
e driver
. [3]
3.
METHO
DOL
OGY
The
se
ns
in
g
of
dro
wsi
ness,
warnin
g
the
dr
ow
sy
dr
i
ver
t
hro
ugh
a
vibrat
or
w
hich
ca
n
be
fit
in
th
e
seat
/st
eering
a
nd
se
nd
i
ng
det
ai
ls
to
the
ow
ne
r
are
the
m
os
t
essenti
al
go
al
s,
howe
ver
the
per
f
orm
ance
of
the
syst
e
m
u
nd
e
r r
eal
tim
e con
str
ai
nts m
us
t
al
so
b
e ac
hiev
ed
. [4]
3.1.
S
ys
te
m Ar
chitecture
Drowsy
dr
i
ver
detect
ion
syst
e
m
is
deco
m
p
os
e
d
into
s
ub
-
syst
e
m
s
and
these
s
ub
syst
e
m
s
pr
ovide
so
m
e
relat
ed
serv
ic
es
.
The
identific
at
ion
of
the
m
ajo
r
com
po
ne
nts
of
the
syst
em
and
c
omm
un
ic
at
ions
betwee
n
the
se
c
om
po
ne
nts is
consi
der
e
d u
nder th
is.
The
syst
em
ar
chite
ct
ur
e
belo
w
portr
ay
s
the
blo
c
ks
re
qu
i
r
ed
f
or
the
im
ple
m
ented
syst
em
.
Fig
ure
.1
ou
tl
ines
the
sy
stem
arch
it
ectu
re
.
The
se
nso
rs
nee
d
to
det
ect
resp
ect
ive
par
am
et
ers
and
pass
t
he
sig
nals
to
m
ic
ro
co
ntro
ll
e
r.
T
he
outp
uts
from
SST
m
i
cro
c
ontrolle
r
a
re
us
e
d
as
in
puts
by
va
rio
us
ou
t
pu
t
dev
ic
e
s
to
con
t
ro
l t
he
s
pe
ed
a
nd d
is
play
the m
essages.
[
5]
Figure
1.
Syst
em
A
rch
it
ect
ur
e
4.
IMPLEME
N
TATION
AN
D RESULT
The
pr
opos
e
d
ideol
og
y
in
volv
es
set
ti
ng
up
a
syst
e
m
fo
r
pre
ve
ntin
g
on
ro
a
d
ve
hicle
acci
de
nts
cause
d
du
e
t
o
dri
ve
r
dro
wsin
es
s.
Keil
com
plier
is
us
ed
f
or
the
fun
ct
ion
in
g
of
the
syst
e
m
.
An
dr
oid
a
pp
li
cat
io
n
helps
no
ti
fy t
he own
er
wh
e
ne
ver
a
n ac
ci
den
t
o
cc
urs w
it
h t
he dri
ve
r’
s
locati
on a
nd ph
otogra
ph.
[6]
Fi
gure
2
depi
ct
s
the
schem
at
ic
diagr
am
of
the
SST
m
icr
o
-
co
ntr
oller,
c
on
sist
in
g
of
ports,
ti
m
ers
and PCA
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
ICT
IS
S
N:
22
52
-
8776
Driver
Dro
wsi
ness Dete
ct
ion
System
U
si
ng
Sensors
(
Kus
uma
Ku
mari B.
M
)
141
Figure
2
.
P
89V
51
R
D2 S
c
hem
at
ic
Figure
3
.
P
DI
P
40 p
i
n
C
onfig
urat
ion
4
.
1.
T
as
k
Imp
le
mentatio
n
Im
ple
m
entation
pr
ocedur
e
of
this
syst
e
m
is
sp
li
t
into
va
ri
ou
s
ste
ps.
Eac
h
ste
p
is
neces
sary
in
or
der
to
achie
ve
t
he
end
res
ult. Steps
a
re
i
nterd
e
pe
nd
e
nt o
n
eac
h
oth
e
r.
Step
-
wise
ap
proac
h
gi
ves
a
bette
r
cl
arit
y
for
dev
el
op
m
ent a
nd im
ple
m
enta
ti
on
a.
Sens
i
ng
I
nfo
r
m
at
ion
:
Wh
en
the
veh
ic
le
m
o
ves,
the
sen
sor
s
would
ta
ke
the
input
from
t
he
ey
e
blink
rate
of
the
dr
i
ver,
the
sp
ee
d
of
the
veh
ic
le
and
a
ny
shock
/vibr
at
io
n
cau
s
ed
due
to
acci
den
t.
T
hese
inputs are
the
n sent to
SS
T
m
i
cro
c
ontrolle
r
.
b.
SST
m
ic
ro
co
nt
ro
ll
er:
The
m
i
cro
c
ontrolle
r
ta
kes
var
i
ous
in
pu
ts
thr
ough
ports
a
nd
passes
the
si
gn
al
s
to the o
utput
ports s
o
t
hat
nec
essary act
io
ns
cou
l
d be ca
rr
ie
d ou
t.
c.
Ou
t
pu
ts:
The
ou
t
pu
t
is
received
th
rou
gh
t
he
SST
po
rts
and
se
nt
to
LCD,
GS
M
m
od
ule
an
d
the
vibrat
or. T
he f
ollow
i
ng g
i
ves
an
acc
ount
of
the outp
uts:
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2252
-
8776
IJECE
V
ol.
6
,
No.
3
,
Decem
ber
20
1
7
:
139
–
145
142
Step
1:
The
G
SM
m
od
u
le
al
erts
the
owne
r
of
t
he
veh
ic
le
by
gi
ving
op
ti
on
s
as
loc
at
ion,
ph
otogra
ph
view
a
nd
detai
ls
to
be
recei
ve
d
in
owne
r’
s
m
ai
l.
The
neare
st
po
li
ce
sta
ti
on
is
al
erte
d
a
t
the
sam
e
tim
e.
Step
2:
The
drow
sy
dri
ve
r
is
al
erted
thr
ough
a
vib
rato
r
w
he
re
in
the
vibr
a
ti
on
is
sent
to
the
glass
of
the eye
b
li
nk s
ens
or
.
Step
3:
The
D
C
m
oto
r
at
ta
c
hed
to
outp
ut
port
is
slo
we
d
dow
n
wh
e
ne
ver
t
he
value
receive
d
by
acce
le
ro
m
et
er ex
ceed
s the
th
r
esh
old
value
.
d.
Disp
la
y:
Every
war
ni
ng
m
es
sage
ge
ner
at
e
d
du
e
to
any
ca
us
al
it
y
caused
by
the
dr
i
ver
dro
wsin
es
s
is
disp
la
ye
d o
n
t
he
LCD sc
ree
n.
e.
Perfo
rm
ance
Evaluati
on:
The
perform
a
nce
of
t
he
al
gorithm
is
evaluated
by
co
ns
ide
rin
g
th
e
eff
ect
ive
ness
and
c
orrect
nes
s
of
c
om
par
ison
with
the
thres
hold
valu
es.
The
decisi
on
m
akin
g
capab
il
it
y of t
he
al
gorithm
is evaluate
d
f
or a
ccur
acy
a
nd re
li
abili
ty
. [
7]
4.2.
Resul
t
a.
The
Use
r
I
nte
rf
ace:
The
fo
ll
ow
i
ng
a
re
the
resu
lt
s
obta
in
ed
f
or
the
us
e
r
i
nterf
ace
m
od
ule
of
ou
r
syst
e
m
:
1.
The UI
is
on a
n
a
ndr
oid
a
pp
wh
ic
h
ca
n be l
oad
e
d
i
nto
a
ny
andr
oid
phone.
2.
The UI
g
i
ves
a
good l
ook
a
nd
f
eel
to
the
pro
je
ct
.
3.
The use
r
e
nters
the
owne
r’
s
phone
num
ber
in
the
app.
4.
The
m
ai
n
app
li
cat
ion
is
in
owne
r’
s
ph
one
fo
r
pur
po
se
of
ret
rievin
g
phot
ogra
ph,
li
st
of
nearby
po
li
ce s
ta
ti
on
s,
l
ocati
on a
nd trac
king.
b.
The Dr
owsy
Driver
Detect
io
n Sy
ste
m
: Th
e
r
esults o
btaine
d by
run
ning the
m
od
ule are
as
fo
ll
ows:
1.
The o
wn
e
r’
s
phone
receives
an
a
ud
i
o war
ni
ng m
essage.
2.
The o
wn
e
r
ca
n retrie
ve
the
lo
cat
ion
of the
dr
iver by c
hoos
i
ng the “l
ocati
on”
op
ti
on.
3.
The
owne
r
ca
n
al
so
g
et
a
li
st of
dri
ve
r
’s
ne
a
rb
y p
olice
sta
ti
on
s
. W
it
h
this,
the p
hoto
gr
a
ph
of
the drive
r
ca
n be se
nt to
ow
ne
r’
s
em
ail ad
dr
ess as s
pecifie
d
in
the clam
ped
phone.
4.
The
wh
eel
/m
oto
r
is
st
oppe
d
as
soo
n
as
a
n
acci
den
t
occur
s,
the
vibrat
or
in
the
ey
e
blink
sens
or
fram
e v
ibrates a
nd d
is
play
s a m
essage on
the
LCD
.
5.
Wh
e
n
t
he
dri
ve
r
fall
s
asl
eep
,
the
vibrat
or
vibrat
es
an
d
t
he
LC
D
dis
pl
ay
s
the
m
essa
ge.
Along wit
h t
hi
s,
the
sp
ee
d o
f t
he
ve
hicle
is
r
edu
ce
d.
6.
Wh
e
n
the
a
cce
le
ro
m
et
er
is
til
te
d
ra
ndom
ly
,
t
hat
act
s
as
the
ste
ering,
a
m
es
sage
is
dis
play
ed
on the
LCD t
he
sp
ee
d of t
he vehicl
e is
re
duced.
7.
W
it
h
al
l
of
t
he
ab
ove
m
e
ntion
e
d,
t
he
a
ndr
oid
a
pp
li
ca
ti
on
s
se
nd
a
nd
recei
ves
det
ai
ls
si
m
ultaneou
sly
.
4.3.
Ti
lted
A
c
cel
erometer
Figure
4
.
Til
te
d Accel
er
om
eter
Fig
ure
4
s
hows
that wh
en
the
ste
ering
is ti
lt
e
d
ra
ndom
ly
,
the v
ibrat
or
v
i
brat
es and
the wheel spee
d
is
slow
e
d d
own.
LCD
disp
la
ys
a m
essage “SL
EEPI
NG”.
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
ICT
IS
S
N:
22
52
-
8776
Driver
Dro
wsi
ness Dete
ct
ion
System
U
si
ng
Sensors
(
Kus
uma
Ku
mari B.
M
)
143
4.4.
E
ye
is
S
h
ut
Figure
5
.
Ey
e is shut
Fig
ure
5
sho
ws
that
wh
e
n
t
he
ey
es
are
cl
os
e
d
the
n
the
vib
rat
or
vi
br
at
es
an
d
the
s
pee
d
of
t
he
wheel
is
decr
ease
d. T
he
LCD dis
play
s a “SLEE
PING
” m
essage.
4.5.
Owner
’s
Pho
n
e Andr
oi
d Appli
ca
tion
Figure
6
.
O
w
ne
r’
s
Pho
ne A
ndr
oid A
ppli
cat
i
on
Fig
ure
6
sho
w
s
the
andro
i
d
app
li
cat
io
n
on
t
he
owne
r’
s
phon
e
.
The
ow
ne
r
can
ch
oose
f
ro
m
a
m
on
g
var
i
ou
s
op
ti
on
s
an
d
retrie
ve
the
re
qu
ire
d
inf
or
m
at
ion
f
r
om
the
cl
am
p
ed
phone.
Be
f
or
e
this,
the
owne
r’
s
phone
play
s a
warnin
g
a
udio
m
essage whe
n no
ti
fie
d by the
GS
M m
od
ul
e
of the syst
em
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2252
-
8776
IJECE
V
ol.
6
,
No.
3
,
Decem
ber
20
1
7
:
139
–
145
144
4.6.
Clam
ped
P
ho
n
e’s
A
c
tiv
ity
S
creen
Figure
7
.
Cl
a
m
ped ph
on
e
’s
ac
ti
vity
scr
een
Fig
ure
7
s
how
s
the
act
ivit
y
screen
of
cl
a
m
ped
phone
wh
e
re
t
he
detai
ls
of
t
he
owner
are
t
o
be
entere
d.
T
he
e
m
ai
l add
ress
is entere
d her
e s
o
that t
he p
ho
t
ogra
ph could
be sent t
o
the
own
e
r’
s
em
ai
l
-
id.
4.7.
Ph
otog
r
aph sen
t to ow
ner’s
e
mail
-
id
Figure
8
.
P
ho
t
ogra
ph sen
t t
o owne
r’
s
em
ail
-
id
Figure
8
De
picts
the
photog
raph
recei
ved
in
the
em
ail
of
the
ow
ner
w
hen
t
he
ow
ner
chooses
th
e
photog
raph
opti
on
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
ICT
IS
S
N:
22
52
-
8776
Driver
Dro
wsi
ness Dete
ct
ion
System
U
si
ng
Sensors
(
Kus
uma
Ku
mari B.
M
)
145
5.
CONCL
US
I
O
N
Accor
ding
to
a
stu
dy,
20%
of
the
c
ras
hes
a
r
e
due
t
o
t
he
dri
ver
dro
wsin
e
s
s.
T
her
e
f
or
e,
de
velo
pm
ent
of
syst
em
fo
r
dr
i
ver
dro
wsine
ss
detect
io
n
is
not
only
le
gitim
a
te
bu
t
al
so
it
is
ver
y
nec
essary.
T
he
propo
s
e
d
syst
e
m
deals
to
al
ert
the slee
py
dr
i
ver
w
it
h
the h
el
p
of
the
v
ib
rator
an
d
th
e
LCD w
il
l
dis
play
sle
epin
g. Also
,
2
andr
oid
phone
s
with
di
f
fer
e
nt
an
droid
ap
plica
ti
on
s
a
re
create
d,
one
wh
e
re
the
ow
ner
has
t
o
acc
ess
the
dr
i
ver’s
locat
i
on
a
nd
the
ot
her
is
in
t
he
c
ar
w
hich
is
use
d
to
detect
th
e
locat
ion
of
t
he
car
a
nd
cl
ic
k
the
pictures
of
t
he
dr
i
ver
.
T
he
ne
arb
y
poli
ce
sta
ti
on
s
can
be
ac
cessed
a
nd
th
e
picture
of
t
he
dr
i
ver
ca
n
be
c
li
cked
and
sen
t
to ow
ner’s
em
ai
l
ad
dr
ess
.
I
n
the
ha
rdwar
e p
a
rt
the
fo
ll
owin
g
se
nsors
,
s
uch
as ey
e
blink
sen
sor w
it
h
a
fr
am
e
wh
ic
h
th
e
dri
ve
r
has
t
o
wear
w
hile
dri
ving,
it
detect
s
blin
king
of
th
e
ey
e
an
d
i
ntim
at
es
the
dr
iv
er
with
the
hel
p
of
the
vibrat
or,
a
n
ac
cel
ero
m
et
er
w
hich
is
at
ta
che
d
to
the
ste
erin
g
wh
ic
h
helps
to
detect
if
the
dr
i
ver
is
dr
ivi
ng
ras
hl
y.
The
DC
m
oto
r
slo
ws
dow
n
w
he
n
the
se
ns
ors
detect
th
at
the
dr
i
ver
is
sle
epin
g
an
d
s
top
s
i
f
the
vi
br
at
io
n
s
ens
or
dete
ct
s
a
ny
acci
de
nt
w
hich
has
occ
urred.
Sim
ultaneou
sly
,
t
he
LC
D
in
the
ca
r
disp
la
y
s
appr
opriat
e w
a
rn
i
ng m
essages.
6.
LIMITATI
O
NS
OF THE
DROWS
Y D
RIV
E
R DET
ECTION
S
YST
EM
The follo
wing
are the
lim
it
a
tio
ns o
f
t
he dr
owsy
drive
r det
ect
ion
syst
em
:
a.
The dri
ver has
to w
ea
r
t
he
ey
e
b
li
nk se
nsor
fr
a
m
e w
hile d
riv
ing
.
b.
On
ly
on
e
D
C
m
oto
r
has bee
n use
d wh
ic
h re
pr
ese
nts
only
one
wh
eel
of the
v
e
hicle
.
7.
FUTU
RE E
N
HANCE
MEN
TS
The
f
ollow
i
ng
can
be
re
garde
d
as
the
fu
tu
re
enh
a
ncem
ents
to
increase
the
viabili
ty
of
t
he
pro
posed
syst
e
m
:
a.
Four
DC m
otors s
hould
be us
ed wh
ic
h rep
re
sent fo
ur whee
l of the ca
b.
Dr
i
ver
w
on’t
ha
ve
to
w
ea
r
t
he
eye
b
li
nk se
nsor
fram
e d
ur
in
g dr
i
ving
REFERE
NCE
S
[1]
Nikola
os
P,
“Vision
-
based
Det
ec
t
ion
of
Dr
iv
e
r
Fat
igue
,
”
Proc.
IE
EE
Interna
ti
onal
Conf
ere
n
ce
on
I
ntelli
g
en
t
Tra
nsporta
ti
on,
pp.
45
-
60,
2000.
[2]
Zut
ao
Zha
ng
;
Ji
ashu
Zha
ng
,
“A
New
Re
al
-
Tim
e
E
ye
Tr
ack
ing
for
Dr
iv
er
Fatigue
De
tecti
on
,
”
Proc.
2006
6t
h
Inte
rna
ti
ona
l
Co
nfe
re
n
c
e
on
ITS
Te
l
ec
om
m
unic
ations,
pp.
8
-
11,
20
06.
[3]
Grac
e;
R
ic
har
d
,
et
al.
"
A
drows
y
drive
r
detec
t
ion
syste
m
for
heav
y
ve
hi
cl
es.
"
Dig
ita
l
Avionic
s
S
y
st
e
m
s
Confer
enc
e,
1998.
Proce
edi
n
gs,
17th
DA
SC
.
The
AIA
A/IE
EE/SA
E.
Vol.
2.
IE
EE
,
pp.
50
-
70
,
19
98.
[4]
Murata
;
Aps
ua
;
Yasutaka
Hir
amatsu.
"Ev
alua
ti
o
n
of
drows
ine
ss
by
HRV
m
easures
-
basic
study
f
or
drows
y
driver
det
e
ct
ion
.
"
Proc
ee
dings
of
4th
Inte
rna
ti
ona
l
W
orkshop
on
Co
mput
at
ion
al
Int
el
l
i
genc
e
&
Appli
c
at
ions.
Hiroshim
a
:
Hiroshim
a
Univer
sit
y
,
2008,
pp.
38
-
45.
[5]
Ito,
T
ake
hiro
,
e
t
al
.
"
Dr
iv
er
bl
in
k
measur
eme
nt
by
the
motion
p
i
ct
ure
proce
ss
ing
and
it
s
appli
ca
t
ion
to
drows
ine
s
s
de
tecti
on
.
"
In
te
l
l
ige
nt
Tra
nsport
a
ti
on
S
y
s
te
m
s,
P
roc
ee
d
ings.
The
IEEE
5th
Int
er
nat
io
n
al
Conf
er
enc
e
on.
IEEE,
pp.
30
-
35
,
2002
.
[6]
Ji
Qiang;
Zh
iwe
i
Zhu;
Pe
il
in
Lan.
"Rea
l
-
ti
m
e
no
nint
rusive
m
onitoring
and
pre
d
iction
of
driv
er
fa
t
igue
.
"
Vehic
u
la
r
Te
chno
log
y
,
IE
EE
Tr
ansacti
ons
on53.
4
,
1052
-
1
06
8,
pp
.
80
-
90
,
2
004
.
[7]
Vural
,
Esra
,
e
t al
.
"A
utomate
d
dr
ows
ine
ss
det
ecti
on
for
impr
oved
drivi
ng
saf
ety
.
"p
p.
48
-
60
,
2008
.
Evaluation Warning : The document was created with Spire.PDF for Python.