Indonesi
an
Journa
l
of El
ect
ri
cal Engineer
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
9
, No
.
3
,
Ma
rch
201
8
,
p
p.
650
~
654
IS
S
N:
25
02
-
4752
,
DOI: 10
.11
591/
ijeecs
.
v9.i
3
.
pp
650
-
654
650
Journ
al h
om
e
page
:
http:
//
ia
es
core.c
om/j
ourn
als/i
ndex.
ph
p/ij
eecs
Tra
in
Ob
stac
l
e Dete
ction S
ystem Using Av
r Mic
rocontr
oll
er
and SR0
4 U
l
tras
onic S
ensor
A.A.
Az
iz
, W.
R.
W
.
Ah
m
ad
Facul
t
y
of
El
e
ct
r
ic
a
l
Eng
ineeri
ng
,
Univer
si
ti Te
kn
ologi
MA
RA (UiT
M
),
4045
0
Shah
Alam,
Sela
ng
or,
Mal
a
y
s
ia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Sep
2
9
, 201
7
Re
vised
Dec
2
7
, 2
01
7
Accepte
d
Ja
n
1
7
, 2
01
8
Malay
s
ia,
ra
i
lwa
y
is
conside
r
e
d
as
the
bac
kb
one
of
tra
nsport
,
connect
in
g
peopl
e
from
all
ac
ross
the
coun
t
r
y
.
W
it
h
the
cur
re
nt
state
of
e
co
nom
y
,
m
ore
peopl
e
pr
efe
r
to choose
train
as
m
ai
n
tra
nsporta
tion e
spec
i
al
l
y
in
big
ci
t
y
ar
e
a
such
as
Kuala
Lumpur.
W
it
h
l
ower
cost
and
r
el
a
ti
ve
l
y
th
e
saf
est
form
of
tra
nsports
compare
d
to
th
e
othe
r
tra
nsports,
li
k
e
ca
rs,
m
otorcy
c
les
or
buss
es,
it
is
a
wise
choic
e
to
use
train
as
dai
l
y
comm
ute
t
ra
nsport.
Now
ad
a
y
s
,
the
r
ai
l
tra
ffi
c
ne
twork
i
n
Malay
sia
are
get
ti
ng
busier
w
it
h
tr
ai
ns
tr
ave
l
i
ng
at
high
er
spee
ds
and
ca
rr
y
ing
m
ore
pass
enge
rs
with
he
a
vie
r
axl
e
lo
ads
tha
n
b
efo
re
.
W
it
h
the
in
cre
as
e
of
p
assenge
r,
t
he
risk
invol
ved
in
da
ily
tra
in
op
era
t
ion
wil
l
signifi
c
ant
l
y
inc
r
ea
se.
An
improv
ed
safe
t
y
s
y
s
te
m
is
re
qu
ire
d
to
k
e
ep
up
wi
th
the
ev
er
growin
g
tra
in
loa
ds.
Th
e
proposed
safe
t
y
s
y
s
te
m
is
applied
to
alert
the
tr
ai
n
oper
at
o
rs.
The
whole
s
y
stem
is
compri
sed
of
an
ult
r
asonic
sensor
conne
c
te
d
to
a
dat
ab
ase
and
an
Atm
ega
328P
mi
cro
con
trol
l
er
m
ounte
d
on
a
custom
PC
B
bo
ard
.
It
is
found
tha
t
th
e
tr
ai
n
in
thi
s
count
r
y
re
quire
s
a
dista
nc
e
of
77
m
et
er
in
ord
er
t
o
complet
e
l
y
sto
p
the
tr
ai
n
wi
th
re
gar
ds
to
a
fe
w a
ss
um
pti
ons on
th
e ave
r
age
m
ass a
nd
the spe
ed
of
the t
r
ai
n
.
Ke
yw
or
d
s
:
Ardu
i
no
D
at
abase
U
lt
rasonic se
nsor
M
ic
ro
co
ntr
oller
V
isual
basic
Copyright
©
201
8
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
:
A.A.
Aziz
,
Faculty
of Elec
tric
al
Engineer
ing
,
Un
i
ver
sit
i Te
knol
og
i M
ARA
(U
iTM
)
,
40450 S
hah A
l
a
m
, S
el
ango
r,
Ma
la
ysi
a
.
Em
a
il
:
anees@sala
m
.u
itm
.ed
u.m
y
1.
INTROD
U
CTION
The
prob
le
m
a
rises
w
he
n,
the
re
is
an
unwa
nt
ed
obsta
cl
e
present
in
the
tr
a
in
track
.
This
will
po
se
a
gr
eat
da
ng
e
r
t
o
the
trai
n
co
nduct
or
an
d
it
s
passe
ng
e
r.
I
n
a
norm
al
trai
n
op
e
rati
on,
t
rain
c
onduct
or
w
il
l
scan
the
track
ahea
d
to
ens
ur
e
th
e
track
is
fr
ee
fr
om
any
ob
s
ta
cl
e
[1
]
.
How
ever,
there
is
a
dr
aw
bac
k
wi
th
this
syst
e
m
,
wh
ic
h
is
the
con
ce
ntr
at
ion
of
trai
n
c
onduct
or
can
be
deterio
rated
ov
e
r
tim
e
thu
s
al
lowing
a
roo
m
fo
r
m
ist
akes.
The
safety
syst
e
m
is
a
design
to
gi
ve
warn
to
the
trai
n
c
onduct
or
an
d
c
on
t
ro
l
cente
r
wh
e
n
an
ob
sta
cl
e
is
det
ect
ed
on
t
he
tr
ack.
Furthe
rm
or
e
,
the
syst
e
m
al
so
c
on
t
ro
l
t
he
trai
n
br
a
kes
i
f
the
t
rain
c
on
du
ct
or
fail
s
to
res
pond
from
the
warnin
g.
P
re
vi
ou
s
st
ud
y
ha
s
been
c
ondu
ct
ed
[
2]
wh
ic
h
cha
ract
erizi
ng
MR
T
op
e
rati
on
in
Ma
nila
to
i
de
ntify
any
pro
bl
e
m
m
igh
t
occ
ur.
I
n
I
ndia
,
a
stu
dy
has
bee
n
done
to
re
duce
t
he
nu
m
ber
o
f
rail
way
incide
nts w
hic
h
is
a
bout r
ai
lway
anti
-
c
ol
li
sion
syst
em
us
in
g
dig
it
al
si
ng
le
le
ns
r
efle
x
(d
sl
r)
and
ultraso
nic
sens
or
to
det
ec
t
the
pr
esence
of
obsta
cl
e
on
the
track
[
3].
In
Ind
on
e
sia
,
ul
traso
nic
sens
or
and
Atem
ega 3
28
Ardu
i
no were
us
e
d
in
a stu
dy
to
m
easur
e t
he
p
lo
ughi
ng d
e
pt
h
el
evati
on
of
dr
ai
nag
e
ch
a
nn
el
[
4].
In
t
his
pro
j
ect
,
an
ultras
onic
s
ens
or
wh
ic
h
is
placed
onboar
d
of
t
he
trai
n
it
sel
f
is
us
e
d.
T
he
se
ns
or
is
us
e
d
to
detect
any
obsta
cl
e
present
on
the
tr
ack.
With
a
pr
esence
of
any
ob
sta
cl
e,
it
will
sen
d
the
dista
nce
of
detect
ion
t
o
th
e
co
ntro
l
ce
nte
r
an
d
if
the
ob
sta
cl
e
is
ver
y
c
lose
to
t
he
trai
n,
t
he
syst
em
will
cut
off
t
he
trai
n’
s
m
oto
r
po
we
r
s
upply.B
esi
de
us
in
g
ultras
on
i
c
sensor,
com
pu
te
r
ai
ded
sci
ence
su
c
h
as
im
age
ench
a
nc
e
m
en
t
with
m
achine
visio
n
[5
]
an
d
real
tim
e
histo
gr
am
or
ie
nted
gr
a
dients
(
H
O
G)
[
6]
are
us
e
d
fo
r
detect
ing
distance
of
obj
e
ct
.
T
he
pro
j
ect
is
ai
m
e
d
to
be
l
ow
co
st,
affo
rd
a
ble
a
nd
reli
able
s
ol
ution
to
trai
n
operat
or
i
n
Ma
la
ysi
a,
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Tra
in
Obst
acle Detect
ion S
yst
em Usin
g
Av
r
Mi
crocontr
oller a
nd SR0
4 Ult
ra
s
on
ic
Se
nsor
(
A.A.
A
ziz
)
651
wh
ic
h
w
ou
l
d
al
low
m
or
e
us
age
of
this
te
chnolo
gy
in
rai
lway
trai
n
infrast
ru
ct
ure.
T
he
dev
el
opm
ent
of
th
e
syst
e
m
will
a
lso
ope
n
up
the
po
s
sibil
it
y
of
im
ple
m
enting
the
in
ven
te
d
sa
fet
y
syst
e
m
on
oth
e
r
area
whic
h
is
no
t l
im
it
ed
on
t
rain
sa
fety
app
li
cat
ion
alo
ne.
2.
RESEA
R
CH MET
HO
D
2.1
.
Br
ea
king
D
ist
an
ce
Iden
tifica
tion
Fo
r
a
trai
n
to
s
top
be
fore
hitt
i
ng
the
obsta
cl
es,
it
re
qu
i
res
a
certai
n
am
ou
nt
of
dista
nce
for
the
ine
rtia
of
t
he
trai
n
t
o
be
re
duced
a
nd
el
im
inate
d.
The
trai
n
sto
ppin
g
dista
nce
i
s
aff
ect
e
d
by
s
ever
al
factors
su
c
h
as
trai
n
weig
ht,
t
rain
sp
ee
d,
an
d
th
e
trac
k
ge
ogra
ph
y.
I
n
t
hi
s
pro
j
ect
,
we
search
f
or
the
m
ini
m
u
m
det
ect
ion
distance
before
the
trai
n
hit
the
ob
sta
cl
e.
B
el
ow
is
t
he
f
orm
ula
us
ed
to
c
al
culat
e
the
br
akin
g
distance
of
t
he
trai
n
with
dif
f
eren
t
detect
io
n
distances
.
A
ssu
m
ing
con
st
ant
gr
a
dient
tr
ack,
the
braki
ng
d
ist
ance
can
be
cal
culat
ed usin
g
,
0
)
(
)
(
2
1
2
1
2
h
h
g
m
U
m
S
a
m
(1)
Fr
om
the
eq
uat
ion
,
“m
”
ref
ers
to
the
trai
n
m
a
ss
in
kg,
w
hile
“a”
ind
ic
at
es
t
he
acce
le
rati
on
rate
w
hile
“
-
a”
an
no
ta
te
s
for
decele
rati
on.
“S”
is
the
de
sired
c
om
po
ne
nt
in
the
eq
ua
ti
on
w
hich
is
t
he
m
ini
m
u
m
br
aki
ng
distance.
For
com
po
ne
nt
“U
”,
it
sy
m
bo
li
zes
the
diff
e
rence
of
the
s
pee
d
at
w
hich
de
cel
erat
ion
be
gun.
“
g”
ref
e
rs
to
the
acce
le
rati
on
pro
vid
e
d
by
gr
avity
.
The
la
st
co
m
po
ne
nt
in
the
eq
uatio
n
is
“h1
-
h2”
wh
ic
h
sy
m
bo
li
zes the
d
if
fer
e
nce i
n h
ei
gh
t
(whe
n goi
ng
on sl
op
e
).
Ther
e
a
re
seve
ral
factor
s
that
aff
ect
br
a
king
distance
of
a
trai
n
[7
]
.
T
he
factor
a
re
the
trai
n
sp
ee
d
wh
e
n
the
bra
ke
is
app
li
ed
,
decele
rati
on
rate
wh
e
n
a
br
a
ke
is
eng
a
ged
wh
ic
h
va
ries
accor
ding
to
the
coeffic
ie
nt
of
f
rict
ion
betwee
n
wh
eel
an
d
ra
il
,
the
ti
m
e
delay
wh
e
n
the
br
akes
a
re
iss
ue
d
by
t
he
trai
n
dri
ver
,
and
the
co
ndit
ion
of
br
a
k
e
pa
ds
us
e
d
al
s
o
aff
ect
t
he
br
a
ki
ng
pe
rfor
m
ance
w
hich
co
nse
qu
e
ntly
aff
ec
t
the
br
a
king
dista
nc
e
besides
the
la
st
factor
w
hi
ch
is
the
ge
ography
of
t
he
tr
ack.
I
n
this
pr
oj
ect
,
scal
e
m
od
el
is
us
e
d
to
dem
onstrat
e
the
detect
ion
from
the
ultraso
nic
sen
so
r
.
The
sc
al
e
is
based
on
1:1
60
scal
es.
T
he
trai
n
m
od
el
us
ed
is
160
ti
m
es
s
m
aller
than
t
he
rea
l
trai
n.
The
det
ect
ion
ra
nge
is
al
so
be
scal
e
d
up
t
o
com
par
e
with
the
cal
culat
ion
produce
d
f
rom
the
fo
rm
ula
and
t
o
m
a
tc
h
the
trai
n
siz
e.
T
he
detect
io
n
ra
ng
e
scal
e
-
up
f
or
m
ul
a
is give
n
as
equ
at
ion
(2),
y
x
160
1
(2)
Wh
e
re
“
x”
is
t
he
scal
ed
up
ultraso
nic
val
ue
(x1
60)
a
nd
“y
”
is
the
m
easur
ed
va
lue
directl
y
fr
om
the
ultr
aso
nic
sens
or
.
Figure
1. SR0
4 Ti
m
ing
Diag
r
a
m
. Repr
inte
d wit
h perm
issi
on
from
I
ndowa
re
data sh
ee
Ultraso
nic
Se
nsor
SR0
4
i
nvol
ved
i
n
this
st
ud
y
is
us
i
ng
wav
e
pr
op
a
gat
ion
c
oncept
to
m
easur
e
the
distance
bet
we
en
the
obsta
cl
e
and
the
trai
n.
Figure
1
i
nd
ic
at
es
the
tim
ing
diagr
am
wh
e
n
the
sens
or
is
unde
r
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vol
.
9
,
No.
3
,
Ma
rc
h
201
8
:
650
–
654
652
op
e
rati
onal
.
From
the
tim
ing
diagr
am
,
a
short
10µs
pu
lse
is
su
ppli
ed
to
trigg
e
r
in
pu
t
t
o
sta
rt
the
ra
nge
of
detect
ion,
the
m
od
ule
will
em
it
an
8
cy
cl
e
burst
of
40
kH
z
ultraso
und
a
nd
raise
it
s
ech
o.
T
he
dista
nc
e
of
the
obj
ect
is
m
eas
ur
e
d
f
ro
m
the
echo
t
hat
is
propo
rtion
to
wi
dth
a
nd
range.
The
ra
nge
can
be
cal
culat
ed
by
us
ing
the tim
e interv
al
b
et
wee
n
se
ndin
g
tri
gg
e
r
si
gnal
and
recei
vi
ng ech
o si
gnal
,
as stat
ed
i
n
e
quat
ion
3.
cm
s
58
(3)
Or
an
oth
e
r
e
quivale
nt for
m
ula;
2
340
s
m
v
e
l
o
c
i
t
y
t
i
m
e
l
e
v
e
l
h
i
g
h
r
a
n
g
e
(4)
In this case
, me
asur
em
ent of
a cy
cl
e o
f
ove
r
60 m
s is r
eco
m
m
end
ed
t
o pre
ven
t t
rig
ger sign
al
t
o
ec
ho si
gn
al
.
2.2.
I
mple
men
tation
The
syst
em
is
sta
rted
w
he
n
an
ob
j
ect
is
detect
ed
ahea
d
on
the
track,
the
n
the
ultraso
nic
sens
or
will
ping
the
pa
rtic
ular
directi
on,
as
ind
ic
at
e
d
in
Figure
2(
a
).
If
the
sen
sor
detect
s
obj
ect
with
in
ra
ng
e
bel
ow
of
77
m
,
the
m
ic
ro
con
t
ro
ll
er
will
cut
off
the
e
ngine
po
wer
supp
ly
i
m
m
ediately
to
stop
the
trai
n.
If
there
is
no
ob
sta
cl
e
detect
ed,
the
trai
n
w
il
l
con
ti
nu
e
it
s
j
our
ney
as
norm
al
.
The
ultrason
ic
se
nsor
at
te
nd
s
as
a
n
in
pu
t
t
o
the
m
ic
ro
con
tr
oller
an
d
it
is
us
e
d
to
dete
ct
the
unwa
nt
ed
ob
sta
cl
e
on
the
t
rack
an
d
fee
d
t
he
data
into
ATm
ega3
28P
as
sh
own
in
Fi
gure
2(b
).
AT
m
ega3
28P
act
as
so
ur
ce
co
nt
ro
ll
er
of
open
hard
war
e
usual
ly
us
ed
in
va
rio
us
a
pp
li
cat
ion
s.
It
is
al
so
use
d
with
cam
era
pix
y
CM
Ucam
5
to
detect
the
pr
es
ence
of
obj
ect
s
wit
h
sp
eci
fic c
olor
[
8]
All
t
he
cal
c
ulati
on
s
by
ta
ki
ng
into
acc
ount
va
rio
us
c
on
diti
on
s
are
im
plem
ented
in
ATm
ega3
28P
m
ic
ro
co
ntro
ll
e
r.
T
he
m
ic
ro
c
on
t
ro
ll
er
m
easur
es
the
dista
nce
us
in
g
t
he
relat
ed
f
orm
ul
as,
a
nd
tran
sm
it
s
the
inf
or
m
at
ion
to
the
co
ntr
ol
center
via
W
i
-
Fi
(E
SP82
66).
I
f
the
o
bs
ta
cl
e
is
too
cl
os
e
f
or
safety
,
the
m
ic
ro
co
ntro
ll
e
r
will
iss
ue
a
n
e
m
erg
ency
bra
king
to
sto
p
t
he
trai
n
im
m
edi
at
el
y.
At
the
sa
m
e
tim
e,
an
al
ert
will
be display
ed
in
Co
ntro
l C
e
ntr
e b
y
us
in
g Vis
ual Basi
c.
Figure
2. (a
)
P
r
ocess fl
ow
of obstac
le
d
et
ect
ion
in t
he new
im
pr
ov
e
d
sa
fety
syst
e
m
, (
b) the
blo
c
k
dia
gra
m
o
f
the safety
syst
e
m
, in
te
gr
at
ing
ultraso
nic se
nsor
and m
ic
ro
co
ntr
oller ATM
ega32
8P
.
The
m
ai
n
hard
war
e
c
om
pone
nts
use
d
in
t
his
ex
per
im
ent
are
an
A
Tm
ega3
28
Mi
cr
oc
on
t
ro
ll
er,
a
5V
Vo
lt
age
regula
tor
(L
M
7805)
,
a
3.
3V
V
oltag
e
Re
gu
la
to
r
(L
M1117,
to
po
wer
the
wifi
m
odule)
,
an
ultr
aso
nic
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Tra
in
Obst
acle Detect
ion S
yst
em Usin
g
Av
r
Mi
crocontr
oller a
nd SR0
4 Ult
ra
s
on
ic
Se
nsor
(
A.A.
A
ziz
)
653
sens
or
(
SR0
4),
a
W
i
-
Fi
Mo
dule
(ESP8
266),
a
DC
Motor,
and
a
cust
om
PCB
bo
ar
d.
T
he
PCB
bo
a
rd
us
e
d
in
this
pro
j
ect
is
custom
fab
rica
te
d
us
in
g
si
ng
l
e
sided
PCB
.
The
PCB
r
outi
ng
is
done
by
us
in
g
P
ro
te
us
s
of
t
war
e
with
T3
0
wi
re th
ic
kn
e
ss
in
or
der
to p
re
ve
nt
wire
disc
onnec
ts
du
ri
ng
et
ching
p
r
ocess
.
Fig
ur
e 3
(a
)
in
dicat
es
the
PCB
la
yout
use
d
in
this
pro
je
ct
,
m
eanw
hile
Fig
ur
e
3(
b)
sh
ows
the
real
fa
br
ic
at
ed
PC
B.
F
or
t
he
s
of
tware
i
m
ple
m
en
ta
ti
o
n,
Ardu
i
no
I
D
E
is
us
ed
a
s
a
platfo
rm
to
wr
it
e
an
d
c
om
pi
le
the
codes.
A
fter
the
cod
e
is
su
ccess
fu
ll
y
c
om
piled,
then
it
is
up
loa
de
d
into
the
ATm
ega32
8P
us
in
g
Ardu
i
no
bo
otloade
r.
A
rduin
o
U
no
dev
el
op
m
ent
bo
a
r
d
is
us
ed
as
a
bu
r
ner
to
up
lo
ad
th
e
cod
e.
T
he
A
Tm
ega3
28P
is
inserted
into
th
e
m
ic
ro
co
ntro
ll
e
r
slot
a
nd
up
l
oa
d
process
will
be
be
gun.
I
n
t
he
c
on
t
ro
l
ce
nt
re,
Mi
cr
osoft
Visu
al
Stu
dio
i
s
us
e
d
to
create
a
visu
al
basic
-
base
d
wi
ndows
a
ppli
cat
ion
.
Fro
m
the
Visu
al
Ba
sic
app
li
cat
ion
c
reated
,
te
c
hn
ic
ia
ns
are
able
to
m
on
it
or
t
he
trai
n
i
n
or
der
to
sens
e
any
unwa
nte
d
obsta
cl
es
on
the
trai
n
trac
k
ahead.
A
datab
ase
is
dev
el
op
e
d
wit
hin
the
local
ho
st
and
ind
i
rectl
y
m
ini
m
iz
ing
po
ssible
er
ror.
My
SQ
L
is
pur
po
s
el
y
add
e
d
to
create
the
databa
se
w
hich
will
be
use
d
to
sto
re
the
input
data
obta
ined
f
r
om
the
Ultraso
nic
Se
nsor
(S
R
04).
T
he
data
from
the d
at
ab
ase is t
he
n fed
into
visu
al
basi
c w
in
dow
appl
ic
at
ion
to
the
di
sp
la
y segm
ent.
Figure
3. (a
)
T
he
PCB
c
reate
d usin
g Pr
oteu
s,
(
b) the
etche
d
PCB
3.
R
ES
ULTS
AND A
NA
L
Y
SIS
The
detect
ion
distances
from
the
ob
sta
cl
e
a
nd
it
s
tolera
nc
e
com
par
in
g
be
tween
ultras
onic
val
ue
a
nd
theo
reti
cal
val
ue
is
in
dicat
ed
in
Table
1.
T
he
a
ver
a
ge
tol
eran
ce
bet
wee
n
real
value
a
nd
bot
h
Ard
uin
o
a
nd
Database
a
re
about
4.3
1%
,
m
eanw
hile
the
highest
tole
ra
nce
is
at
2.0
0
c
m
in
real
va
lue
w
hich
is
9.50%
tolerance
a
nd
the
lo
west
toler
ance
is
1.0
0%
m
easur
ed
between
real
valu
e
and
bo
t
h
A
r
du
i
no
a
nd
Dat
abase
exclu
ding
the
first
m
easur
ed
value
of
0.00c
m
fr
om
real
value,
A
r
du
i
no
a
nd
Data
bas
e.
F
or
a
trai
n
to
sto
p
b
ef
or
e
hitt
ing
t
he
obsta
cl
es,
it
re
qu
i
res
a
cert
ai
n
am
ou
nt
of
distance
f
or
th
e
inerti
a
of
the
trai
n
to
be
re
duce
d
and
el
im
inate
d.
A
fe
w
facto
rs
that
are
af
fect
ing
t
he
trai
n
st
opping
distanc
e,
are
trai
n
wei
gh
t,
trai
n
s
pee
d,
a
nd
the
track
ge
ogr
aph
y.
I
n
this
s
t
ud
y, we
se
arc
h
the
m
ini
m
u
m
detect
ion
d
ist
a
nce b
ef
or
e
t
he
trai
n
hit
the
ob
sta
cl
e.
By
assum
ing
c
on
sta
nt
gr
a
die
nt
trac
k,
t
he
br
akin
g
distance,
S
ca
n
be
cal
cu
la
te
d
us
i
ng
e
quat
ion
(1).
I
n
order
to
us
e this f
or
m
ula, a
few
ass
umpti
ons h
a
d
to be
m
ade w
hich
t
he
m
e
tho
d
reli
es o
n
s
om
e si
m
pl
ify
ing
expr
essions
as im
ple
m
ented
in
Bar
ney et
al
.[
7]
Table
1.
T
he
to
le
ran
ce e
rro
r b
et
ween
real
value
a
nd v
al
ues ob
ta
ine
d from
A
r
du
i
no
Real Valu
e
(c
m
)
Ardu
in
o
(
c
m
)
Databas
e (
c
m
)
Tolerance
(%)
0
0
.00
0
.00
0
.00
0
.00
0
2
.00
2
.19
3
.19
9
.50
0
4
.00
4
.17
4
.17
4
.25
0
6
.00
6
.33
6
.33
5
.50
0
8
.00
8
.74
8
.74
9
.25
1
0
.00
1
0
.34
1
0
.34
3
.40
1
2
.00
1
2
.19
1
2
.19
1
.58
1
4
.00
1
4
.14
1
4
.14
1
.00
Fr
om
the
cal
culat
ion
,
it
is
sh
own
t
hat
the
m
i
nim
u
m
br
akin
g
distance
is
a
ppr
ox
im
at
ely
77
m
fr
om
the
ob
sta
cl
e.
In
or
der
for
a
trai
n
stop
e
ff
ic
ie
ntl
y,
the
pre
-
set
stoppin
g
distance
of
a
trai
n
hav
e
t
o
be
m
or
e
tha
n
77
m
.
From
Ta
ble
2,
betwee
n
the
ra
ng
es
0.0
0m
to
70
.
00
m
the
overla
p
sta
tus
is
ye
s,
this
m
eans
the
trai
n
does
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vol
.
9
,
No.
3
,
Ma
rc
h
201
8
:
650
–
654
654
no
t
hav
e
s
uffic
ie
nt
distance
t
o
sto
p,
th
us
hitt
ing
the
ob
sta
c
le
.
W
he
n
the
de
te
ct
ion
ra
ng
e
increase
in
t
he
range
of 80.
00m
to
90.
00
m
, th
e trai
n hav
e
s
uffici
ent sto
ppin
g,
t
hus all
owin
g
t
he
train t
o
a
vo
i
d hit
ti
ng
th
e
ob
st
acl
es.
Table
1
. B
rak
i
ng d
ist
a
nce (S)
calc
ulate
d fr
om
d
et
ect
ion
r
a
ng
e
usi
ng e
qu
a
ti
on
(1)
Detectio
n
Distan
ce
(
Ardu
in
o
,
c
m
)
Detectio
n
Distan
ce
(
x
1
6
0
,
m)
Brak
in
g
Distan
ce (S
),
m
Ov
erlap Statu
s
0
.00
0
.00
-
7
7
.00
Yes
6
.25
1
0
.00
-
6
7
.00
Yes
1
2
.50
2
0
.00
-
5
7
.00
Yes
1
8
.75
3
0
.00
-
4
7
.00
Yes
2
5
.00
4
0
.00
-
3
7
.00
Yes
3
1
.25
5
0
.00
-
2
7
.00
Yes
3
7
.50
6
0
.00
-
1
7
.00
Yes
4
3
.75
7
0
.00
-
7
.00
No
5
0
.00
8
0
.00
3
.00
No
5
6
.25
9
0
.00
1
3
.00
No
4.
CON
CLU
SION
In
c
oncl
us
io
n,
trai
n
colli
sio
ns
c
ou
l
d
be
a
vo
i
ded
w
hen
an
ef
fecti
ve
s
afety
app
li
ed
t
o
the
syst
em
entirel
y.
The
t
rain
ob
sta
cl
e
de
te
ct
ion
by
us
i
ng
A
VR
m
ic
r
ocontr
oller
a
nd
SR
04
Ultras
on
ic
se
nsor
ha
s
bee
n
stud
ie
d
in
this
pro
j
ect
t
o
pro
po
s
e
the
m
os
t
conve
nient
a
nd
a
fforda
ble
s
afety
syst
e
m
into
t
his
i
nfrastru
ct
ur
e
industries.
F
rom
this
pr
oject
,
it
is
c
le
arly
fo
und
that
the
trai
n
in
Ma
la
ysi
a
with
fe
w
assu
m
pt
ion
s
re
gard
ing
t
he
aver
a
ge
m
ass
and
s
pee
d
requires
a
distanc
e
of
77m
to
com
ple
te
ly
stop
the
trai
n,
wh
ic
h
m
ean
s
the
train
ha
s
su
f
fici
ent
dista
nce
to
sto
p
if
t
he
pr
e
-
set
butt
on
is
ap
plied
f
or
t
he
distance
of
detect
io
n
m
or
e
than
77m
.
It
is
exp
ect
e
d
t
hat
the
in
ve
nted
saf
et
y
syst
e
m
in
t
his
re
searc
h
ca
n
be
im
ple
m
ented
in
oth
e
r
tr
a
in
ope
rati
on
se
rv
ic
es
include
s the
o
t
he
r
infrastr
uctu
res
in
Mal
ay
sia
.
REFERE
NCE
S
[1]
Rüder
M,
Möhler
N,
Ahm
ed
F.
An
Obs
ta
cl
e
Detect
ion
S
y
st
em
for
Autom
at
ed
Tra
ins.
IE
EE
In
te
lligent
V
ehicles
Symposium
.
Col
um
bus.
2003;
4:
180
–
185.
[2]
Tomas
U,
Ganir
on
Jr.
Expl
or
i
ng
the
Emergi
ng
Im
pac
t
of
Metro
Rail
Tr
a
nsit
(MRT
-
3)
i
n
Metro
Manila
.
Inte
rnational
Jo
urnal
of Adv
an
c
ed
Sc
ie
n
ce
and
T
ec
hnolog
y
.
2015
;
74(2):
11
-
24.
[3]
Chouhan
S.
Rail
wa
y
Ant
i
-
Coll
ision
S
y
stem
using
DS
LR
Sensor.
Inte
rnational
Jo
urnal
of
Engi
n
e
ering
Scienc
es
&
Re
search te
chno
logy
.
2014;
3(3), 1199
–
1
202.
[4]
Hari
ans
y
ah
M,
Seti
awa
n
R
P
A,
Made
D,
Sape
i
D
S
A.
The
A
ppli
c
at
ion
of
Ul
tra
sonic
Sensor
and
Atemega
32
8
Arduino
to
Mea
sure
the
Plough
ing
Depth
El
ev
at
ion
of
Drai
na
ge
Channel.
Int
ernati
onal
Journal
of
S
ci
en
ce
an
d
Re
search
.
2014;
3(8):
174
–
180.
[5]
Hos
sein
Z,
Mo
hamm
ad
H
D
M.
An
Acc
ura
t
e
Scheme
for
Distance
Mea
sure
m
ent
Us
ing
an
Ordina
r
y
W
ebca
m
.
Inte
rnational
Jo
urnal
of El
e
ct
ri
c
al
and
Comput
er
Engi
n
ee
ring
.
20
17;
7(1), 209
–
21
5.
[6]
Sachi
n
S,
Dharm
esh
S.
Rea
l
-
Ti
m
e
Autom
at
ic
O
bstac
l
e
Dete
c
ti
o
n
and
Aler
t
Sy
st
e
m
for
Drive
r
A
ss
ista
nce
on
Indian
Roads.
Int
ernational
Journal
of
El
e
ct
rica
l
Eng
in
ee
ring a
nd
Computer
Sc
ie
nc
e
.
2
016;
1(3), 635
–
6
46.
[7]
Barney
D,
Haley
D,
Nikandr
os
G.
Cal
culati
ng
Tra
in
Brak
ing
Distanc
e
.
Sixth
Aus
tra
li
an
W
orkshop
on
Indust
ri
a
l
Expe
ri
enc
e
with
Safe
t
y
Cr
it
i
ca
l
S
y
stems
and
Soft
ware
.
Brisbane. 2001;
3:
23
–
30.
[8]
Budi
R,
Hugo
A,
Heru
I,
Ha
m
dani
H.
Dista
nce
Est
imati
on
base
d
on
Color
-
Bloc
k:
A
Sim
ple
Big
-
O
Anal
y
s
is.
Inte
rnational
Jo
urnal
of El
e
ct
ri
c
al
and
Comput
er
Engi
n
ee
ring
.
20
17;
7(4), 2169
–
2
175.
Evaluation Warning : The document was created with Spire.PDF for Python.