I
n
t
e
r
n
at
io
n
al
Jou
r
n
al
of
A
d
van
c
e
s
i
n
A
p
p
li
e
d
S
c
ie
n
c
e
s
(
I
JA
A
S
)
V
ol
.
14
, N
o.
3
,
S
e
pt
e
m
be
r
20
25
, pp.
724
~
739
I
S
S
N
:
2252
-
8814
,
D
O
I
:
10.11591/
ij
a
a
s
.
v14.
i
3
.
pp724
-
739
724
Jou
r
n
al
h
om
e
page
:
ht
tp
:
//
ij
aas
.i
ae
s
c
or
e
.c
om
A
u
t
on
om
ou
s
n
avi
gat
i
on
sy
st
e
m
f
or
a r
ove
r
w
i
t
h
r
ob
ot
i
c
ar
m
u
si
n
g c
on
vol
u
t
i
on
al
n
e
u
r
al
n
e
t
w
or
k
s
A
z
i
z
E
l
M
r
ab
e
t
1
, H
ic
h
am
H
ih
i
1
, M
oh
am
m
e
d
K
h
al
il
L
agh
r
ai
b
1
, M
b
ar
e
k
C
h
ah
b
ou
n
1
,
A
ym
an
e
A
m
al
aou
i
2
1
L
a
bor
a
t
or
y of
E
ngi
ne
e
r
i
ng, S
ys
t
e
m
s
a
nd A
ppl
i
c
a
t
i
ons
, N
a
t
i
ona
l
S
c
hool
of
A
ppl
i
e
d S
c
i
e
nc
e
s
, S
i
di
M
oha
m
e
d
B
e
n A
bde
l
l
a
h U
ni
ve
r
s
i
t
y,
F
e
z
, M
or
oc
c
o
2
L
a
bor
a
t
or
y of
I
nnova
t
i
ve
T
e
c
hnol
ogi
e
s
, N
a
t
i
ona
l
S
c
hool
of
A
ppl
i
e
d S
c
i
e
nc
e
s
,
S
i
di
M
oha
m
e
d B
e
n A
bde
l
l
a
h U
ni
ve
r
s
i
t
y, F
e
z
,
M
or
oc
c
o
A
r
t
ic
le
I
n
f
o
A
B
S
T
R
A
C
T
A
r
ti
c
le
h
is
to
r
y
:
R
e
c
e
iv
e
d
O
c
t
22, 2024
R
e
vi
s
e
d
M
a
y 21, 2025
A
c
c
e
pt
e
d
J
un 8, 2025
The
aim
of
this
project
is
to
design
and
develop
an
autonomous
rover
equipped
with
a
KUKA
robotic
arm.
This
mobile
vehicle
will
be
able
to
move autonomo
usly thanks to th
e
use of ma
chine
learning te
chniques.
It will
also
be
able
to
detect
and
retrieve
objects
using
the
KUKA
arm.
The
rover
will
feature
Mecanum
wheels
for
improved
maneuverability
and
will
be
controll
ed
by
a
Raspberry
Pi
3
board,
with
machine
learning
algo
rithms
implemented
using
TensorFlow
and
Python.
The
developmen
t
proce
ss
will
follow
the
V
-
methodology.
The
use
of
such
an
autonomous
rover
and
its
manipulative
capabilitie
s
opens
the
way
to
many
practical
applic
ations,
including
sampling
in
dangerou
s
or
difficult
-
to
-
access
environments,
search
and
rescue
operations
in
the
event
of
natural
disasters
or
industrial
acc
idents,
and
inspection
and
maintenance
of
industrial
or
construction
sites.
Th
e
rover
could
also
be
used
for educational
purposes,
enabling
students
to
expl
ore
the
concepts o
f roboti
cs and art
ificial
intell
igence.
K
e
y
w
o
r
d
s
:
C
onvolut
io
na
l
ne
ur
a
l
ne
twor
ks
M
A
T
L
A
B
P
yt
hon
R
a
s
pbe
r
r
y P
i
3
R
obot
a
r
m
V
-
s
ha
pe
d
This is an
open
acce
ss artic
le unde
r the
CC BY
-
SA
license.
C
or
r
e
s
pon
di
n
g A
u
th
or
:
A
z
iz
E
l
M
r
a
be
t
L
a
bor
a
to
r
y of
E
ng
in
e
e
r
in
g, S
ys
te
m
s
a
nd A
ppl
ic
a
ti
ons
, N
a
ti
ona
l
S
c
hool
of
A
ppl
ie
d S
c
ie
nc
e
s
S
id
i
M
oha
m
e
d B
e
n A
bde
ll
a
h U
ni
v
e
r
s
it
y
F
e
z
, M
or
oc
c
o
E
m
a
il
:
a
z
iz
.e
lm
r
a
be
t@us
m
ba
.a
c
.m
a
1.
I
N
T
R
O
D
U
C
T
I
O
N
T
he
r
a
pi
d
a
dva
nc
e
m
e
nt
of
r
obot
ic
s
ha
s
le
d
to
th
e
de
v
e
lo
pm
e
n
t
of
a
ut
onomous
s
ys
te
m
s
c
a
pa
bl
e
of
pe
r
f
or
m
in
g
in
tr
ic
a
te
ta
s
ks
w
it
h
m
in
im
a
l
hum
a
n
in
te
r
ve
nt
io
n.
A
s
ig
ni
f
ic
a
nt
br
e
a
kt
hr
ough
in
th
is
f
ie
ld
is
th
e
in
te
gr
a
ti
on
of
r
obot
ic
a
r
m
s
w
it
h
m
obi
le
pl
a
tf
or
m
s
,
w
hi
c
h
e
nha
nc
e
s
th
e
ir
v
e
r
s
a
ti
li
ty
a
nd
e
na
bl
e
s
ope
r
a
ti
on
in
di
ve
r
s
e
e
nvi
r
onm
e
nt
s
.
S
uc
h
s
ys
t
e
m
s
ha
ve
f
ound
a
ppl
ic
a
ti
on
s
in
s
e
a
r
c
h
a
nd
r
e
s
c
ue
,
in
dus
tr
ia
l
a
ut
om
a
ti
on,
a
nd
a
gr
ic
ul
tu
r
a
l
r
obot
ic
s
,
a
m
ong
ot
he
r
s
.
H
ow
e
v
e
r
,
de
s
pi
te
not
a
bl
e
pr
ogr
e
s
s
in
th
e
de
ve
lo
pm
e
nt
of
m
obi
le
r
obot
s
a
nd
r
obot
ic
a
r
m
s
,
m
os
t
e
xi
s
ti
ng
s
ol
ut
io
ns
te
nd
to
f
oc
us
on
e
it
he
r
m
obi
li
ty
or
m
a
ni
pul
a
ti
on,
li
m
it
in
g
th
e
ir
a
bi
li
ty
t
o pe
r
f
or
m
a
b
r
oa
d r
a
nge
of
t
a
s
ks
a
ut
onomous
ly
. F
or
i
ns
ta
nc
e
, s
ys
t
e
m
s
de
s
ig
n
e
d f
or
a
ut
onomous
s
pa
c
e
r
e
nde
z
vous
of
te
n
pr
io
r
it
iz
e
na
vi
ga
ti
on
u
s
in
g
a
c
ti
ve
s
e
ns
or
s
li
ke
li
ght
de
te
c
ti
on
a
nd
r
a
ngi
ng
(
L
iDA
R
)
,
r
e
le
ga
ti
ng
vi
s
ua
l
s
e
ns
or
s
s
u
c
h
a
s
c
a
m
e
r
a
s
to
s
e
c
onda
r
y
r
ol
e
s
[
1]
,
[
2]
.
T
he
s
e
s
y
s
te
m
s
a
r
e
ty
pi
c
a
ll
y
hi
ghl
y
s
pe
c
ia
li
z
e
d
a
nd t
a
il
or
e
d t
o s
pe
c
if
ic
e
nvi
r
onm
e
nt
s
, w
hi
c
h r
e
s
tr
ic
ts
t
he
ir
br
oa
de
r
a
ppl
ic
a
bi
li
ty
.
R
e
c
e
nt
a
dv
a
nc
e
m
e
nt
s
in
m
a
c
hi
n
e
le
a
r
ni
ng
a
nd
s
e
n
s
or
te
c
hnol
ogi
e
s
ha
ve
e
na
bl
e
d
s
ig
ni
f
ic
a
nt
im
pr
ove
m
e
nt
s
in
a
ut
onomous
na
vi
ga
ti
on
a
nd
obj
e
c
t
m
a
ni
pul
a
t
io
n
[
3]
.
N
e
ve
r
th
e
le
s
s
,
m
a
ny
of
th
e
s
e
s
ys
te
m
s
r
e
ly
on
s
pe
c
ia
li
z
e
d
h
a
r
dw
a
r
e
or
la
c
k
th
e
f
le
xi
bi
li
ty
r
e
qui
r
e
d
f
or
di
ve
r
s
e
r
e
a
l
-
w
or
ld
a
ppl
ic
a
ti
ons
.
W
hi
le
s
tu
di
e
s
ha
ve
de
m
ons
tr
a
te
d
th
e
e
f
f
e
c
ti
ve
ne
s
s
of
m
a
c
hi
ne
l
e
a
r
ni
ng
a
lg
or
it
hm
s
in
r
obot
ic
s
ys
te
m
s
f
or
ta
s
k
s
s
u
c
h
a
s
s
or
ti
ng
a
nd
obj
e
c
t
c
la
s
s
if
ic
a
ti
on
[
4]
,
th
e
in
te
gr
a
ti
on
of
th
e
s
e
a
lg
or
it
hm
s
w
it
h
m
obi
le
pl
a
tf
or
m
s
a
nd
r
obot
ic
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
dv A
ppl
S
c
i
I
S
S
N
:
2252
-
8814
A
ut
onomous
nav
ig
at
io
n s
y
s
te
m
f
o
r
a r
ov
e
r
w
it
h r
obot
ic
a
r
m
us
i
ng c
onv
ol
ut
io
nal
ne
ur
al
…
(
A
z
iz
E
l
M
r
abe
t)
725
a
r
m
s
r
e
m
a
in
s
unde
r
e
xpl
or
e
d.
F
ur
th
e
r
m
or
e
,
th
e
de
v
e
lo
pm
e
nt
of
a
ut
onomous
r
ove
r
s
e
qui
pp
e
d
w
it
h
M
e
c
a
num
w
he
e
ls
e
n
a
bl
in
g
om
ni
di
r
e
c
ti
ona
l
m
ove
m
e
nt
h
a
s
pr
ovi
de
d
s
up
e
r
io
r
m
a
ne
uve
r
a
bi
li
ty
c
om
pa
r
e
d
to
tr
a
di
ti
ona
l
w
he
e
l
c
onf
ig
ur
a
ti
ons
[
5]
.
T
hi
s
c
a
pa
bi
li
ty
is
c
r
it
ic
a
l
f
or
n
a
vi
ga
ti
ng
c
om
pl
e
x
e
nvi
r
onm
e
nt
s
,
tr
a
ve
r
s
in
g
obs
ta
c
le
s
,
a
nd r
e
a
c
hi
ng t
a
r
ge
t
lo
c
a
ti
ons
w
it
h pr
e
c
is
io
n.
T
he
in
te
gr
a
ti
on
of
a
K
U
K
A
r
obot
ic
a
r
m
f
ur
th
e
r
e
nha
nc
e
s
th
e
r
ove
r
'
s
obj
e
c
t
d
e
te
c
ti
on
a
nd
m
a
ni
pul
a
ti
on
c
a
pa
bi
li
ti
e
s
,
e
na
bl
in
g
it
to
r
e
tr
ie
ve
obj
e
c
ts
,
p
e
r
f
or
m
in
s
pe
c
ti
ons
,
a
nd
e
xe
c
ut
e
hi
ghl
y
pr
e
c
is
e
ta
s
ks
.
T
he
c
om
bi
na
ti
on
of
r
obot
ic
a
r
m
s
w
it
h
m
obi
le
pl
a
tf
or
m
s
ha
s
be
e
n
e
xpl
or
e
d
in
num
e
r
ous
s
tu
di
e
s
,
de
m
ons
tr
a
ti
ng
th
e
ir
pot
e
nt
ia
l
a
c
r
os
s
in
dus
tr
ie
s
s
uc
h
a
s
a
gr
ic
ul
tu
r
e
a
nd
m
a
nuf
a
c
tu
r
in
g
[
6]
–
[
9]
.
A
dva
nc
e
d
m
a
c
hi
ne
le
a
r
ni
ng
te
c
hni
que
s
,
in
c
lu
di
ng
de
e
p
le
a
r
ni
ng
a
nd
m
ul
t
it
a
s
k
c
onvolut
io
na
l
ne
ur
a
l
ne
twor
ks
(M
C
N
N
)
,
ha
ve
f
ur
th
e
r
im
pr
ove
d
th
e
a
da
pt
a
bi
li
ty
a
nd
a
c
c
ur
a
c
y
of
th
e
s
e
s
ys
te
m
s
.
F
or
e
xa
m
pl
e
,
th
e
Y
O
L
O
-
M
C
N
N
m
ode
l
ha
s
pr
ove
n
e
f
f
e
c
ti
ve
in
c
om
pl
e
ti
ng
m
ul
ti
pl
e
ta
s
ks
,
s
uc
h
a
s
ta
r
ge
t
de
te
c
ti
on,
pos
e
e
s
ti
m
a
ti
on,
a
nd
obs
ta
c
le
s
e
gm
e
nt
a
ti
on,
w
hi
c
h
a
r
e
e
s
s
e
nt
ia
l
f
or
a
ut
onomous
ope
r
a
ti
ons
[
7]
,
[
10
]
.
T
he
s
e
a
dva
nc
e
m
e
nt
s
r
e
duc
e
th
e
ne
e
d f
or
m
a
nua
l
in
te
r
ve
nt
io
n a
nd e
nha
nc
e
ope
r
a
ti
ona
l
e
f
f
ic
ie
nc
y.
T
hi
s
pa
pe
r
pr
e
s
e
nt
s
a
s
tr
uc
tu
r
e
d
a
nd
va
li
da
te
d
a
ppr
oa
c
h
to
th
e
de
ve
lo
pm
e
nt
of
a
n
a
ut
onomous
r
ove
r
s
ys
te
m
,
e
m
pl
oyi
ng
th
e
V
-
m
e
th
odol
ogy
a
pr
ove
n
pr
oc
e
s
s
f
or
m
is
s
io
n
-
c
r
it
ic
a
l
pr
oj
e
c
ts
r
e
qui
r
in
g
hi
gh
r
e
li
a
bi
li
ty
a
nd
pe
r
f
or
m
a
nc
e
.
T
he
r
ove
r
'
s
de
s
ig
n
in
c
or
por
a
te
s
a
R
a
s
pbe
r
r
y
P
i
3
pl
a
tf
or
m
,
w
hi
c
h
pr
oc
e
s
s
e
s
r
e
a
l
-
ti
m
e
da
ta
f
r
om
s
e
ns
or
s
a
nd
c
a
m
e
r
a
s
to
gui
de
th
e
ve
hi
c
le
th
r
ough
it
s
e
nvi
r
onm
e
nt
.
A
ddi
ti
ona
ll
y,
th
e
s
tu
dy
e
xpl
or
e
s
th
e
in
te
gr
a
ti
on
of
m
a
c
hi
ne
le
a
r
ni
ng
a
lg
or
it
hm
s
,
pa
r
ti
c
ul
a
r
ly
c
onvolut
io
na
l
ne
ur
a
l
ne
twor
ks
(
C
N
N
s
)
,
w
hi
c
h
a
r
e
c
e
nt
r
a
l
to
th
e
r
ove
r
'
s
c
a
pa
bi
li
ti
e
s
.
T
he
m
odul
a
r
de
s
ig
n
of
th
e
s
y
s
te
m
,
de
ve
lo
p
e
d
us
in
g
C
a
ti
a
V
5,
is
di
s
c
us
s
e
d
a
lo
ng
s
id
e
th
e
a
ppl
ic
a
ti
on
of
a
dva
n
c
e
d
a
lg
or
it
hm
s
a
n
d
th
e
in
te
gr
a
ti
on
of
a
K
U
K
A
r
obot
ic
a
r
m
w
it
h
th
e
m
obi
le
pl
a
tf
or
m
.
T
he
pa
pe
r
a
ls
o
hi
ghl
ig
ht
s
th
e
im
por
ta
nc
e
of
s
e
c
ur
e
c
om
m
uni
c
a
ti
on
pr
ot
oc
ol
s
,
s
uc
h
a
s
ope
n pl
a
tf
or
m
c
om
m
uni
c
a
ti
ons
uni
f
ie
d a
r
c
hi
te
c
tu
r
e
(
O
P
C
U
A
)
,
a
nd a
dva
nc
e
d da
ta
m
a
na
g
e
m
e
nt
t
e
c
hni
que
s
i
n
e
ns
ur
in
g
th
e
r
e
li
a
bi
li
ty
a
nd
s
a
f
e
ty
of
th
e
a
ut
onomous
s
ys
te
m
.
B
y
pr
ovi
di
ng
a
c
om
p
r
e
he
ns
iv
e
ove
r
vi
e
w
of
th
e
de
ve
lo
pm
e
nt
pr
oc
e
s
s
a
nd
de
s
ig
n
c
ons
id
e
r
a
ti
ons
,
th
is
p
a
pe
r
a
im
s
to
c
ont
r
ib
ut
e
to
th
e
br
oa
de
r
f
ie
ld
of
a
ut
onomous
r
obot
ic
s
.
T
he
f
ol
lo
w
in
g
s
e
c
ti
ons
de
lv
e
in
to
th
e
s
ys
te
m
'
s
de
s
ig
n,
th
e
m
e
th
odol
ogy
e
m
pl
oye
d,
a
nd
th
e
e
va
lu
a
ti
on of
i
ts
pe
r
f
or
m
a
nc
e
, w
hi
le
a
ls
o di
s
c
us
s
in
g i
m
pl
ic
a
ti
ons
f
or
f
ut
ur
e
r
e
s
e
a
r
c
h.
2.
M
E
T
H
O
D
T
he
de
v
e
lo
pm
e
nt
of
th
is
r
obot
f
ol
lo
w
e
d
a
s
tr
uc
tu
r
e
d
a
nd
s
ys
te
m
a
ti
c
a
ppr
oa
c
h
ba
s
e
d
on
th
e
V
-
m
e
th
odol
ogy. T
he
V
-
s
ha
pe
d de
v
e
lo
pm
e
nt
m
ode
l
is
a
n i
te
r
a
ti
ve
, i
nc
r
e
m
e
nt
a
l
a
ppr
oa
c
h f
or
s
of
twa
r
e
pr
oj
e
c
ts
th
a
t
ha
ve
w
e
ll
-
de
f
in
e
d
r
e
qui
r
e
m
e
nt
s
but
ne
e
d
f
le
xi
bi
li
ty
.
T
hi
s
de
ve
lo
pm
e
nt
c
yc
le
f
ol
lo
w
s
a
V
-
s
ha
pe
w
it
h
s
e
que
nt
ia
l
a
nd
p
a
r
a
ll
e
l
s
te
ps
,
in
c
lu
di
ng
r
e
qui
r
e
m
e
nt
s
a
na
ly
s
is
,
de
s
ig
n,
im
pl
e
m
e
nt
a
ti
on,
te
s
ti
ng
,
a
nd
va
li
da
ti
on.
T
hi
s
m
od
e
l
is
p
a
r
ti
c
ul
a
r
ly
w
e
ll
-
s
ui
te
d
to
m
is
s
io
n
-
c
r
it
ic
a
l
pr
oj
e
c
ts
;
e
a
c
h
ph
a
s
e
w
a
s
m
e
ti
c
ul
ous
ly
e
xe
c
ut
e
d t
o e
ns
ur
e
t
he
s
ys
te
m
'
s
r
e
li
a
bi
li
ty
, f
unc
ti
ona
li
ty
, a
nd r
e
pr
oduc
ib
il
it
y.
2.1. Re
q
u
ir
e
m
e
n
t
s
an
al
ys
is
A
c
om
pr
e
he
ns
iv
e
a
na
ly
s
i
s
of
bot
h
f
unc
ti
ona
l
a
nd
non
-
f
unc
ti
ona
l
r
e
qui
r
e
m
e
nt
s
w
a
s
c
onduc
te
d
to
de
f
in
e
th
e
s
c
ope
a
nd
obj
e
c
ti
ve
s
of
th
e
pr
oj
e
c
t.
T
he
f
unc
ti
ona
l
r
e
qui
r
e
m
e
nt
s
f
oc
us
on
th
e
r
ove
r
'
s
c
or
e
c
a
pa
bi
li
ti
e
s
,
w
hi
c
h
in
c
lu
de
a
ut
onomous
na
vi
ga
ti
on
us
in
g
m
a
c
hi
ne
le
a
r
ni
ng
te
c
hni
que
s
,
om
ni
di
r
e
c
ti
ona
l
m
ove
m
e
nt
e
na
bl
e
d
by
M
e
c
a
num
w
he
e
l
s
,
obs
ta
c
le
de
te
c
ti
on
a
nd
a
voi
da
nc
e
u
s
in
g
ul
tr
a
s
oni
c
s
e
n
s
or
s
,
a
nd
obj
e
c
t
de
te
c
ti
on
a
nd
r
e
tr
ie
va
l
f
a
c
il
it
a
te
d
by
a
KUKA
a
r
m
a
nd
a
c
a
m
e
r
a
-
ba
s
e
d
vi
s
io
n
s
ys
t
e
m
.
T
h
e
s
e
c
a
pa
bi
li
ti
e
s
e
n
s
ur
e
th
e
r
ove
r
c
a
n
op
e
r
a
te
e
f
f
e
c
ti
ve
ly
in
dyna
m
ic
e
nvi
r
onm
e
nt
s
,
lo
c
a
t
e
ta
r
ge
t
obj
e
c
ts
,
a
nd
in
te
r
a
c
t
w
it
h i
ts
s
ur
r
oundings
.
I
n
a
ddi
ti
on
to
f
unc
ti
ona
l
r
e
qui
r
e
m
e
nt
s
,
th
e
s
ys
te
m
w
a
s
de
s
i
gne
d
to
m
e
e
t
s
e
ve
r
a
l
non
-
f
unc
ti
ona
l
r
e
qui
r
e
m
e
nt
s
,
s
uc
h
a
s
pe
r
f
or
m
a
nc
e
,
r
e
li
a
bi
li
ty
,
a
nd
us
a
bi
li
ty
.
T
he
r
ove
r
m
us
t
ope
r
a
te
w
it
hout
f
a
il
ur
e
f
or
e
xt
e
nde
d
pe
r
io
ds
,
r
e
c
ove
r
f
r
om
e
r
r
or
s
a
ut
onomous
ly
,
a
nd
pr
ov
i
de
c
le
a
r
s
ta
tu
s
upda
te
s
to
non
-
te
c
hni
c
a
l
us
e
r
s
.
H
a
r
dw
a
r
e
s
pe
c
if
ic
a
ti
ons
in
c
lu
de
th
e
u
s
e
of
a
R
a
s
pbe
r
r
y
P
i
3
a
s
th
e
c
e
nt
r
a
l
pr
oc
e
s
s
in
g
uni
t,
M
e
c
a
num
w
h
e
e
ls
f
or
e
nha
nc
e
d
m
a
ne
uve
r
a
bi
li
ty
,
a
KUKA
a
r
m
f
or
obj
e
c
t
r
e
tr
ie
va
l,
a
nd
a
c
a
m
e
r
a
f
or
obj
e
c
t
de
t
e
c
ti
on
a
nd
r
e
c
ogni
ti
on.
O
n
th
e
s
of
twa
r
e
s
id
e
,
th
e
s
ys
te
m
le
v
e
r
a
ge
s
T
e
n
s
or
F
lo
w
a
nd
P
yt
hon
f
or
m
a
c
hi
ne
le
a
r
ni
ng
a
nd
c
ont
r
ol
a
lg
or
it
hm
s
.
F
in
a
ll
y,
th
e
de
s
ig
n
a
dhe
r
e
s
to
s
pe
c
if
ic
budge
t,
s
iz
e
,
a
nd
w
e
ig
ht
c
ons
tr
a
in
ts
to
e
ns
ur
e
pr
a
c
ti
c
a
li
ty
a
nd f
e
a
s
ib
il
it
y.
2.2. S
ys
t
e
m
ar
c
h
it
e
c
t
u
r
e
T
he
s
ys
te
m
a
r
c
hi
te
c
tu
r
e
is
m
odul
a
r
,
w
it
h
e
a
c
h
m
odul
e
pe
r
f
or
m
in
g
a
s
pe
c
if
ic
f
unc
ti
on
a
nd
c
om
m
uni
c
a
ti
ng
a
s
ync
hr
onou
s
ly
vi
a
de
di
c
a
te
d
to
pi
c
s
.
T
hi
s
de
s
i
gn
e
ns
ur
e
s
f
le
xi
bi
li
ty
,
s
c
a
la
bi
li
ty
,
a
nd
e
a
s
e
of
m
a
in
te
na
nc
e
.
T
he
c
a
m
e
r
a
m
odul
e
c
a
pt
ur
e
s
im
a
g
e
s
u
s
in
g
th
e
r
ove
r
'
s
onboa
r
d
c
a
m
e
r
a
a
nd
publ
is
he
s
th
e
m
to
th
e
"
im
a
ge
_t
opi
c
"
f
o
r
us
e
by
ot
he
r
m
odul
e
s
.
T
he
obj
e
c
t
de
te
c
ti
on
m
odul
e
s
ubs
c
r
ib
e
s
to
th
is
to
pi
c
a
nd
a
na
ly
z
e
s
th
e
pi
c
tu
r
e
s
w
it
h
a
C
N
N
ba
s
e
d
on
th
e
r
e
s
id
ua
l
ne
t
w
or
ks
(
R
e
s
N
e
ts
)
-
50
a
r
c
hi
te
c
tu
r
e
.
T
hi
s
m
odul
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8814
I
nt
J
A
dv A
ppl
S
c
i
,
V
ol
.
14
, N
o.
3
,
S
e
pt
e
m
be
r
20
25
:
724
-
739
726
de
te
c
ts
a
nd
id
e
nt
if
ie
s
obj
e
c
ts
w
it
hi
n
th
e
im
a
ge
s
,
publ
is
hi
ng
th
e
r
e
s
ul
ts
(
e
.g.,
obj
e
c
t
c
oor
di
na
te
s
a
nd
ty
pe
)
to
th
e
"
obj
e
c
t_
de
te
c
ti
on_t
opi
c
"
.
T
he
KUKA
a
r
m
c
ont
r
ol
m
odul
e
s
ubs
c
r
ib
e
s
to
th
e
"
obj
e
c
t_
de
te
c
ti
on_t
opi
c
"
to
r
e
c
e
iv
e
obj
e
c
t
c
oor
di
na
te
s
a
nd
c
a
lc
ul
a
te
s
th
e
a
r
m
m
ove
m
e
nt
s
r
e
qui
r
e
d
to
gr
a
s
p
th
e
de
te
c
te
d
obj
e
c
ts
u
s
in
g
in
ve
r
s
e
ki
ne
m
a
ti
c
s
.
I
t
th
e
n
publ
i
s
he
s
a
r
m
m
ove
m
e
nt
c
om
m
a
nd
s
to
th
e
"
a
r
m
_c
ont
r
ol
_t
opi
c
"
.
T
he
w
he
e
l
c
ont
r
ol
m
odul
e
s
ubs
c
r
ib
e
s
to
th
is
to
pi
c
a
nd
a
dj
u
s
ts
th
e
r
ove
r
'
s
M
e
c
a
num
w
he
e
ls
to
po
s
it
io
n
th
e
r
ove
r
f
or
obj
e
c
t
r
e
tr
ie
va
l,
publ
is
hi
ng
w
he
e
l
m
ove
m
e
nt
c
om
m
a
nds
to
th
e
"
w
he
e
ls
_c
ont
r
ol
_t
opi
c
"
.
F
in
a
ll
y,
th
e
ul
tr
a
s
oni
c
s
e
ns
or
m
odul
e
m
e
a
s
ur
e
s
di
s
ta
nc
e
s
to
s
ur
r
ounding
obj
e
c
ts
a
nd
publ
is
he
s
th
is
da
ta
to
th
e
s
ys
te
m
,
e
na
bl
in
g
obs
ta
c
le
a
voi
da
n
c
e
a
nd na
vi
ga
ti
on.
2.3. I
m
p
le
m
e
n
t
at
io
n
d
e
t
ai
ls
T
he
im
pl
e
m
e
nt
a
ti
on
pha
s
e
in
vol
ve
d
in
te
gr
a
ti
ng
ha
r
dw
a
r
e
c
om
pone
nt
s
a
nd
de
ve
lo
pi
ng
s
of
twa
r
e
a
lg
or
it
hm
s
to
br
in
g
th
e
s
ys
te
m
to
li
f
e
.
T
he
R
a
s
pbe
r
r
y
P
i
3
w
a
s
s
e
t
up
to
c
onne
c
t
w
it
h
th
e
M
e
c
a
num
w
he
e
ls
,
KUKA
a
r
m
,
c
a
m
e
r
a
,
a
nd
ul
tr
a
s
oni
c
s
e
ns
or
s
,
f
unc
ti
oni
ng
a
s
th
e
c
e
nt
r
a
l
hub
f
or
da
ta
pr
oc
e
s
s
in
g
a
nd
c
ont
r
ol
.
T
he
c
ont
r
ol
a
lg
or
it
hm
s
w
e
r
e
e
xe
c
ut
e
d
in
P
yt
hon,
ut
il
iz
in
g
T
e
ns
or
F
lo
w
f
or
C
N
N
-
ba
s
e
d
obj
e
c
t
de
te
c
ti
on.
T
he
R
e
s
N
e
t
-
50
m
ode
l
w
a
s
c
ho
s
e
n
f
or
it
s
ba
la
nc
e
of
a
c
c
ur
a
c
y
a
nd
c
om
put
a
ti
ona
l
e
f
f
ic
ie
nc
y,
a
nd
it
w
a
s
f
in
e
-
tu
ne
d
f
or
our
s
pe
c
if
ic
obj
e
c
t
de
te
c
ti
on t
a
s
k us
in
g a
d
a
ta
s
e
t
of
l
a
be
le
d i
m
a
ge
s
.
A
publ
is
h
-
s
ubs
c
r
ib
e
m
ode
l
f
a
c
il
it
a
te
d
c
om
m
uni
c
a
ti
on
be
t
w
e
e
n
m
odul
e
s
.
T
hi
s
m
od
e
l
a
ll
ow
s
a
s
ync
hr
onous
a
nd
in
de
p
e
nde
nt
op
e
r
a
ti
on
w
hi
le
e
ns
ur
in
g s
e
a
m
l
e
s
s
da
ta
e
xc
ha
nge
.
T
hi
s
a
ppr
oa
c
h e
na
bl
e
s
e
a
c
h
m
odul
e
t
o f
unc
ti
on a
ut
onomous
ly
w
hi
le
r
e
m
a
in
in
g s
ync
hr
oni
z
e
d w
it
h t
he
ove
r
a
ll
s
ys
te
m
.
2.4. Ve
r
if
ic
at
io
n
an
d
val
id
at
io
n
E
a
c
h
m
odul
e
w
a
s
r
ig
or
ous
ly
te
s
te
d
in
i
s
ol
a
ti
on
a
nd
a
s
pa
r
t
of
t
he
in
te
gr
a
te
d
s
y
s
te
m
to
e
n
s
ur
e
it
m
e
t
th
e
s
pe
c
if
ie
d
r
e
qui
r
e
m
e
nt
s
.
T
he
obj
e
c
t
de
te
c
ti
on
a
lg
or
it
hm
w
a
s
va
li
da
te
d
us
in
g
a
da
ta
s
e
t
of
la
be
le
d
im
a
ge
s
.
T
he
r
ove
r
'
s
n
a
vi
ga
ti
on
a
nd
obj
e
c
t
r
e
tr
ie
va
l
c
a
pa
bi
li
ti
e
s
w
e
r
e
te
s
te
d
in
a
c
ont
r
ol
le
d
e
nvi
r
onm
e
nt
,
c
onf
ir
m
in
g
th
a
t
a
ll
f
unc
ti
ona
l
a
nd
non
-
f
unc
ti
ona
l
r
e
qui
r
e
m
e
nt
s
w
e
r
e
m
e
t.
T
he
s
e
t
e
s
ts
de
m
ons
tr
a
te
d
th
e
s
ys
te
m
'
s
a
bi
li
ty
to
a
ut
onomous
ly
na
vi
ga
te
, de
te
c
t
a
nd a
voi
d obs
ta
c
le
s
, l
oc
a
t
e
t
a
r
ge
t
obj
e
c
ts
, a
nd r
e
tr
ie
ve
t
he
m
us
in
g
KUKA
a
r
m
.
3.
O
B
JE
C
T
D
E
T
E
C
T
I
O
N
A
L
G
O
R
I
T
H
M
T
hi
s
s
tu
dy
f
oc
us
e
s
on
th
e
im
pl
e
m
e
nt
a
ti
on
of
m
a
c
hi
n
e
le
a
r
ni
ng
a
nd
de
e
p
le
a
r
ni
ng
m
e
th
odol
ogi
e
s
to
de
s
ig
n
a
n
obj
e
c
t
de
te
c
ti
on
f
r
a
m
e
w
or
k.
A
s
a
br
a
nc
h
o
f
a
r
ti
f
ic
ia
l
in
te
ll
ig
e
nc
e
,
m
a
c
hi
ne
le
a
r
ni
ng
a
ll
ow
s
c
om
put
a
ti
ona
l
s
ys
te
m
s
to
a
ut
onomous
ly
im
pr
ove
pe
r
f
or
m
a
nc
e
th
r
ough
da
ta
-
dr
iv
e
n
le
a
r
ni
ng
a
ut
onomous
ly
,
bypa
s
s
in
g t
he
ne
e
d f
or
di
r
e
c
t
pr
ogr
a
m
m
in
g. S
uc
h t
e
c
hni
que
s
a
r
e
a
ppl
ie
d e
xt
e
ns
iv
e
ly
i
n doma
in
s
r
a
ngi
ng f
r
om
voi
c
e
a
nd
f
a
c
ia
l
r
e
c
ogni
ti
on
to
t
a
r
ge
t
id
e
nt
if
ic
a
ti
on
s
y
s
te
m
s
.
M
a
c
hi
ne
le
a
r
ni
ng
s
tr
a
te
gi
e
s
a
r
e
ty
pi
c
a
ll
y
c
a
te
gor
iz
e
d
in
to
th
r
e
e
pr
im
a
r
y
ty
pe
s
:
s
upe
r
vi
s
e
d
le
a
r
ni
ng
f
or
la
be
le
d
da
ta
a
na
ly
s
i
s
,
un
s
upe
r
vi
s
e
d
le
a
r
ni
ng
f
or
pa
tt
e
r
n di
s
c
ove
r
y i
n unla
be
le
d da
ta
s
e
ts
, a
nd r
e
in
f
or
c
e
m
e
nt
l
e
a
r
ni
ng f
or
de
c
is
io
n
-
m
a
ki
ng opti
m
iz
a
ti
on t
h
r
ough
it
e
r
a
ti
ve
f
e
e
dba
c
k.
D
e
e
p
le
a
r
ni
ng,
on
th
e
ot
he
r
ha
nd,
is
a
br
a
nc
h
of
m
a
c
hi
ne
le
a
r
ni
ng
th
a
t
e
xpl
oi
ts
a
r
ti
f
ic
ia
l
ne
ur
a
l
ne
twor
ks
to
s
ol
ve
c
om
pl
e
x
pr
obl
e
m
s
.
A
s
il
lu
s
tr
a
te
d
in
F
ig
u
r
e
1,
de
e
p
ne
ur
a
l
ne
twor
ks
,
s
uc
h
a
s
C
N
N
s
,
a
r
e
pa
r
ti
c
ul
a
r
ly
w
e
ll
-
s
ui
te
d
to
c
om
put
e
r
vi
s
io
n
a
nd
im
a
ge
c
la
s
s
if
ic
a
ti
on
ta
s
ks
.
I
n
our
ob
je
c
t
de
te
c
ti
on
s
ys
te
m
,
w
e
ha
ve
us
e
d a
pr
e
-
e
nt
r
a
in
e
d
C
N
N
.
F
ig
ur
e
1. D
e
e
p ne
ur
a
l
ne
twor
k a
r
c
hi
te
c
tu
r
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
dv A
ppl
S
c
i
I
S
S
N
:
2252
-
8814
A
ut
onomous
nav
ig
at
io
n s
y
s
te
m
f
o
r
a r
ov
e
r
w
it
h r
obot
ic
a
r
m
us
i
ng c
onv
ol
ut
io
nal
ne
ur
al
…
(
A
z
iz
E
l
M
r
abe
t)
727
3.1.
P
r
e
s
e
n
t
at
io
n
of
t
h
e
n
e
u
r
al
n
e
t
w
or
k
c
on
vol
u
t
io
n
al
gor
it
h
m
an
d
t
h
e
R
e
s
N
et
-
50 m
od
e
l
C
N
N
s
r
e
pr
e
s
e
nt
a
c
la
s
s
of
de
e
p
le
a
r
ni
ng
m
ode
ls
c
om
m
onl
y
us
e
d
f
or
im
a
ge
c
la
s
s
if
ic
a
ti
on.
A
s
s
how
n
in
F
ig
ur
e
2,
t
he
y
e
xpl
oi
t
c
onvolut
io
n
ope
r
a
ti
ons
to
e
xt
r
a
c
t
r
e
l
e
va
nt
f
e
a
tu
r
e
s
f
r
om
im
a
ge
s
,
th
e
r
e
by
r
e
duc
in
g
th
e
c
om
pl
e
xi
ty
of
t
he
m
ode
l
in
t
e
r
m
s
of
pa
r
a
m
e
te
r
s
t
o be
l
e
a
r
ne
d. T
he
s
e
c
onvolut
io
na
l
ne
twor
ks
a
r
e
ge
ne
r
a
ll
y
c
om
pos
e
d
of
c
onvolut
io
n,
pool
in
g,
a
nd
f
ul
ly
c
onne
c
te
d
la
ye
r
s
,
e
na
bl
in
g
c
la
s
s
if
ic
a
ti
on
to
be
pe
r
f
or
m
e
d.
T
he
ne
twor
k
is
tr
a
in
e
d
by
it
e
r
a
ti
ve
ly
a
dj
us
ti
ng
th
e
w
e
ig
ht
s
of
th
e
v
a
r
io
us
la
ye
r
s
,
ba
s
e
d
on
th
e
tr
a
in
in
g
d
a
ta
.
C
N
N
ha
s
pr
ove
n
its
w
or
th
i
n m
a
ny c
om
put
e
r
vi
s
io
n a
ppl
ic
a
ti
ons
[
11]
, [
12]
.
F
ig
ur
e
2. E
dge
de
te
c
ti
on pr
oc
e
s
s
T
he
R
e
s
N
e
t
a
r
c
hi
te
c
tu
r
e
w
a
s
s
e
l
e
c
te
d
f
or
th
is
pr
oj
e
c
t.
R
e
s
N
e
t
is
a
de
e
p
c
onvolut
io
na
l
ne
twor
k
th
a
t
us
e
s
r
e
s
id
ua
l
c
onne
c
ti
ons
,
m
a
ki
ng
it
e
a
s
y
to
tr
a
in
ve
r
y
de
e
p
m
ode
ls
.
T
he
ne
twor
k
is
bui
lt
f
r
om
s
ta
c
ke
d
r
e
s
id
ua
l
c
onvolut
io
n
bl
oc
ks
.
I
n
a
ddi
ti
on,
tr
a
ns
f
e
r
le
a
r
ni
ng
i
s
of
te
n
e
m
pl
oye
d
w
it
h
R
e
s
N
e
t,
e
na
bl
in
g
pr
e
-
tr
a
in
e
d c
onvolut
io
n l
a
ye
r
s
t
o be
e
xpl
oi
te
d f
or
i
m
a
ge
f
e
a
tu
r
e
e
xt
r
a
c
ti
on.
E
a
r
ly
a
ppr
oa
c
he
s
to
im
a
ge
pr
oc
e
s
s
in
g
w
e
r
e
ba
s
e
d
on
th
e
us
e
of
f
il
te
r
s
to
e
xt
r
a
c
t
f
e
a
tu
r
e
s
s
uc
h
a
s
obj
e
c
t
c
ont
our
s
.
F
r
om
a
m
a
th
e
m
a
ti
c
a
l
poi
nt
of
vi
e
w
,
th
is
in
vol
ve
s
th
e
a
ppl
ic
a
ti
on
of
c
onvolut
io
n
ope
r
a
ti
ons
,
w
hi
c
h c
ons
is
t
of
s
um
s
of
e
le
m
e
nt
a
r
y pr
oduc
ts
ove
r
i
m
a
ge
bl
oc
ks
. C
onvolut
io
n c
a
n be
de
f
in
e
d on a
2D
m
a
tr
ix
a
nd
c
a
n
b
e
e
xt
e
nd
e
d
to
vol
um
e
s
[
13]
.
T
he
im
a
ge
i
s
th
e
n
r
e
pr
e
s
e
nt
e
d
a
s
a
te
n
s
or
,
w
it
h
di
m
e
ns
io
n
s
f
or
he
ig
ht
,
w
id
th
,
a
nd numbe
r
of
c
ha
nne
ls
[
14]
, [
15]
.
3.
1
.1.
P
ad
d
in
g a
n
d
s
t
r
id
e
i
n
c
on
vol
u
t
io
n
s
P
a
ddi
ng
a
nd
s
tr
id
e
a
r
e
ke
y
ope
r
a
ti
ons
f
or
de
a
li
ng
w
it
h
th
e
lo
s
s
of
in
f
or
m
a
ti
on
a
t
th
e
e
dge
s
of
th
e
im
a
ge
w
he
n
a
ppl
yi
ng
c
onvolut
io
ns
.
P
a
ddi
ng
c
ons
is
t
s
of
a
ddi
n
g
z
e
r
os
a
r
ound
th
e
im
a
ge
to
ta
ke
in
to
a
c
c
ount
pi
xe
ls
lo
c
a
te
d
on
th
e
e
dge
s
,
w
hi
le
s
tr
id
e
c
ont
r
ol
s
th
e
s
iz
e
of
th
e
out
put
by
a
dj
us
ti
ng
th
e
di
s
ta
nc
e
tr
a
ve
le
d
by
th
e
f
il
te
r
dur
in
g c
onvolut
io
n. T
he
f
ol
lo
w
in
g s
e
c
ti
on e
xpl
or
e
s
t
he
s
e
c
onc
e
pt
s
i
n de
ta
il
[
16]
, [
17]
.
P
a
ddi
ng:
w
he
n
a
c
onvolut
io
n
is
a
ppl
ie
d
w
it
h
a
ve
r
ti
c
a
l
e
dge
f
i
lt
e
r
,
pi
xe
ls
in
th
e
im
a
ge
'
s
c
or
ne
r
s
a
r
e
us
e
d
le
s
s
th
a
n
th
os
e
in
th
e
c
e
nt
e
r
,
le
a
di
ng
to
a
lo
s
s
of
e
dge
in
f
or
m
a
ti
on.
T
o
s
ol
ve
th
is
pr
obl
e
m
,
i
t
is
c
om
m
on
to
a
dd
a
f
r
a
m
e
a
r
ound
th
e
im
a
ge
.
A
s
il
lu
s
tr
a
te
d
in
F
ig
ur
e
3,
th
is
pa
ddi
ng
us
ua
ll
y
in
vol
ve
s
a
ddi
ng
z
e
r
os
a
r
ound
th
e
or
ig
in
a
l
im
a
ge
s
o
th
a
t
pi
xe
ls
a
t
th
e
e
dge
s
c
a
n
be
ta
ke
n
in
to
a
c
c
ount
dur
in
g
c
onvolut
io
n.
T
he
dpi
pa
r
a
m
e
te
r
c
or
r
e
s
ponds
t
o t
he
numbe
r
of
e
le
m
e
nt
s
a
dd
e
d t
o e
a
c
h
s
id
e
of
t
he
i
m
a
ge
.
F
ig
ur
e
3. P
a
ddi
ng i
n
C
N
N
s
S
tr
id
e
:
t
he
s
tr
id
e
c
or
r
e
s
ponds
to
th
e
s
pe
e
d
a
t
w
hi
c
h
th
e
f
il
te
r
m
ove
s
ove
r
th
e
im
a
ge
dur
in
g
c
onvolut
io
n.
T
he
s
iz
e
of
th
e
out
put
de
c
r
e
a
s
e
s
w
it
h
a
la
r
ge
r
s
tr
i
de
,
w
hi
le
a
s
m
a
ll
e
r
s
tr
id
e
ke
e
p
s
it
la
r
ge
r
.
T
hi
s
di
s
ta
nc
e
is
r
e
pr
e
s
e
nt
e
d
by
th
e
s
p
a
r
a
m
e
te
r
.
F
or
e
xa
m
pl
e
,
a
s
tr
id
e
of
1
in
di
c
a
te
s
th
a
t
th
e
f
il
te
r
m
ove
s
one
pi
xe
l
a
t
a
ti
m
e
,
w
hi
le
a
s
tr
id
e
of
2
in
di
c
a
te
s
th
a
t
it
m
ove
s
two
pi
x
e
ls
a
t
e
a
c
h
s
te
p.
A
s
s
ho
w
n
in
F
ig
ur
e
4,
th
e
s
e
c
onc
e
pt
s
a
r
e
il
lu
s
tr
a
te
d
in
th
e
f
ol
lo
w
in
g
im
a
ge
s
,
s
how
in
g
a
n
e
xa
m
pl
e
of
pa
ddi
ng
w
it
h
=
1
a
nd
a
c
onvolut
io
na
l
pr
oduc
t
w
it
h
=
1
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8814
I
nt
J
A
dv A
ppl
S
c
i
,
V
ol
.
14
, N
o.
3
,
S
e
pt
e
m
be
r
20
25
:
724
-
739
728
F
ig
ur
e
4. C
onvolut
io
n
o
pe
r
a
ti
on
A
r
ig
or
ous
de
f
in
it
io
n
of
th
e
c
onvolut
io
n
ope
r
a
ti
on
r
e
qui
r
e
s
c
la
r
it
y
on
two
c
r
it
ic
a
l
c
om
pone
nt
s
:
pa
ddi
ng
a
nd
s
tr
id
e
.
P
a
ddi
ng,
pa
r
a
m
e
te
r
iz
e
d
by
p,
pr
e
s
e
r
ve
s
s
pa
ti
a
l
e
dge
in
f
or
m
a
ti
on
by
a
ppe
ndi
ng
z
e
r
os
a
r
ound
th
e
in
put
m
a
tr
ix
.
M
e
a
nw
hi
le
,
s
tr
id
e
(
s
)
gove
r
ns
th
e
di
s
pl
a
c
e
m
e
nt
in
te
r
va
l
of
th
e
f
il
te
r
dur
in
g
c
onvolut
io
n,
di
r
e
c
tl
y
in
f
lu
e
nc
in
g
out
put
d
im
e
ns
io
na
li
ty
.
T
he
c
onvolut
io
na
l
out
put
is
c
om
put
e
d
a
s
a
2D
m
a
tr
ix
,
w
he
r
e
e
a
c
h
e
nt
r
y
c
or
r
e
s
ponds
to
th
e
s
um
m
a
ti
on
of
e
l
e
m
e
nt
-
w
is
e
pr
oduc
ts
be
twe
e
n
th
e
f
il
te
r
’
s
3D
te
ns
or
a
nd
a
n
ove
r
la
ppi
ng
s
ub
-
c
ub
e
of
th
e
in
put
te
ns
or
.
F
or
a
n
im
a
ge
w
it
h
di
m
e
ns
io
ns
[
ℎ
,
,
]
(
ℎ
:
t
he
s
iz
e
of
t
he
he
ig
ht
,
:
t
he
s
iz
e
of
t
he
wid
h
,
a
nd
:
t
he
n
um
b
e
r
of
c
ha
n
n
e
l
s
)
.
I
n
th
e
c
a
s
e
of
a
n
R
G
B
im
a
ge
,
f
or
e
xa
m
pl
e
,
=
3
w
e
ha
v
e
r
e
d,
gr
e
e
n
,
a
nd
bl
ue
.
B
y
c
onve
nt
io
n,
w
e
c
ons
id
e
r
th
a
t
th
e
K
f
il
te
r
is
gr
id
de
d a
nd ha
s
a
n odd dim
e
ns
io
n note
d f
, w
hi
c
h a
ll
ow
s
e
a
c
h
pi
xe
l
to
be
c
e
nt
e
r
e
d i
n t
he
f
il
te
r
a
nd t
he
r
e
f
or
e
to
ta
ke
in
to
a
c
c
ount
a
ll
th
e
e
le
m
e
nt
s
s
ur
r
ounding
it
,
s
o
th
a
t
w
e
a
ppl
y
a
f
il
te
r
of
di
m
e
ns
io
n
[
,
,
]
.
T
h
e
c
onv
ol
u
ti
o
na
l
pr
od
uc
t
be
tw
e
e
n t
h
e
i
m
a
ge
a
nd
th
e
f
il
t
e
r
i
s
a
2D
m
a
tr
ix
, e
a
c
h e
l
e
m
e
nt
of
w
hi
c
h i
s
t
he
s
um
of
th
e
m
u
lt
i
pl
i
c
a
ti
on
pe
r
e
le
m
e
nt
of
th
e
c
u
be
(
f
il
te
r
)
a
nd
t
h
e
s
ub
-
c
u
b
e
o
f
t
he
gi
ve
n i
m
a
ge
,
a
s
i
ll
u
s
tr
a
te
d i
n F
ig
ur
e
5.
F
ig
ur
e
5. T
he
c
ube
(
f
il
te
r
)
a
nd t
he
s
ub
-
c
ube
of
t
he
gi
ve
n i
m
a
ge
M
a
th
e
m
a
ti
c
a
ll
y,
th
e
di
m
e
n
s
io
na
li
ty
of
th
e
c
onvolut
io
n
ope
r
a
ti
o
n
be
twe
e
n
a
n
in
put
im
a
ge
I
a
nd
f
il
te
r
K
i
s
de
f
in
e
d a
s
in
(
1)
.
d
im
(
(
,
)
)
=
{
(
[
+
2
−
+
1
]
,
[
+
2
−
+
1
]
)
;
>
0
(
ℎ
+
2
−
,
+
2
−
)
;
=
0
(
1)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
dv A
ppl
S
c
i
I
S
S
N
:
2252
-
8814
A
ut
onomous
nav
ig
at
io
n s
y
s
te
m
f
o
r
a r
ov
e
r
w
it
h r
obot
ic
a
r
m
us
i
ng c
onv
ol
ut
io
nal
ne
ur
al
…
(
A
z
iz
E
l
M
r
abe
t)
729
W
he
r
e
⌊
⌋
is
t
he
f
lo
or
f
unc
ti
on of
.
C
om
m
on c
onvolut
io
n va
r
ia
nt
s
in
c
lu
de
:
‒
V
a
li
d c
onvolut
io
n:
=
0
‒
S
a
m
e
c
onvolut
io
n:
out
put
s
iz
e
=
in
put
s
iz
e
→
=
−
1
2
‒
1
×
1 c
onvolut
io
n:
E
m
pl
oys
a
uni
t
-
s
iz
e
d f
il
te
r
(
=
1
)
, of
te
n us
e
d t
o r
e
duc
e
c
ha
nne
l
de
pt
h (
)
w
hi
le
r
e
ta
in
in
g s
pa
ti
a
l
r
e
s
ol
ut
io
n
(
,
)
.
I
n
th
e
il
lu
s
tr
a
ti
ve
e
x
a
m
pl
e
,
F
ig
ur
e
5,
f
il
te
r
va
lu
e
s
a
r
e
m
a
nua
ll
y
in
it
ia
li
z
e
d
f
or
c
la
r
it
y.
H
ow
e
ve
r
,
in
pr
a
c
ti
c
a
l
C
N
N
s
, t
he
×
×
f
il
te
r
pa
r
a
m
e
te
r
s
a
r
e
opt
im
iz
e
d a
ut
om
a
ti
c
a
ll
y vi
a
ba
c
kpr
opa
ga
ti
on dur
in
g t
r
a
in
in
g
.
3.
1
.2.
P
ool
in
g
P
ool
in
g
la
ye
r
s
dow
ns
a
m
pl
e
s
pa
ti
a
l
di
m
e
ns
io
n
s
(
,
)
w
hi
le
pr
e
s
e
r
vi
ng
c
ha
nne
l
de
pt
h
(
)
.
T
hi
s
ope
r
a
ti
on
a
ppl
ie
s
a
f
ix
e
d
-
a
ggr
e
ga
ti
on
f
unc
ti
on
(
non
-
tr
a
in
a
bl
e
)
to
lo
c
a
li
z
e
d
r
e
gi
ons
of
th
e
in
put
te
ns
or
,
tr
a
ve
r
s
e
d by a
f
il
te
r
of
s
iz
e
×
w
it
h s
tr
id
e
. T
he
out
put
di
m
e
ns
io
n
s
a
r
e
gove
r
ne
d by
(
2)
.
d
im
(
(
)
)
=
{
(
[
+
2
−
+
1
]
,
[
+
2
−
+
1
]
,
)
;
>
0
(
ℎ
+
2
−
,
+
2
−
,
)
;
=
0
(
2)
S
ta
nda
r
d
pr
a
c
ti
c
e
e
m
pl
oys
s
qua
r
e
f
il
te
r
s
(
×
)
,
ty
pi
c
a
ll
y
w
it
h
=
2
a
nd
=
2
to
ha
lv
e
th
e
s
pa
ti
a
l
r
e
s
ol
ut
io
n w
hi
le
a
voi
di
ng ove
r
la
p. C
om
m
on pooli
ng f
unc
ti
ons
i
nc
lu
de
:
‒
A
ve
r
a
ge
pooli
ng:
c
om
put
e
s
t
he
m
e
a
n of
va
lu
e
s
w
it
hi
n t
he
f
il
te
r
’
s
r
e
c
e
pt
iv
e
f
ie
ld
.
‒
M
a
x pooli
ng:
e
xt
r
a
c
ts
t
he
m
a
xi
m
um
va
lu
e
f
r
om
t
he
f
il
te
r
’
s
w
in
dow
.
U
nl
ik
e
c
onvolut
io
na
l
la
ye
r
s
,
pool
in
g
ut
il
iz
e
s
pr
e
de
f
in
e
d
op
e
r
a
ti
ons
(
no
le
a
r
na
bl
e
pa
r
a
m
e
te
r
s
)
,
pr
io
r
it
iz
in
g
c
om
put
a
ti
ona
l
e
f
f
ic
ie
nc
y a
nd t
r
a
ns
la
ti
ona
l
in
va
r
ia
nc
e
i
n de
e
p n
e
twor
ks
.
3.
1
.
3
.
B
u
il
d
in
g a
c
on
vol
u
t
io
n
al
n
e
u
r
al
n
e
t
w
or
k
la
ye
r
b
y l
aye
r
A
C
N
N
is
c
ons
tr
uc
te
d
by
s
ta
c
ki
ng
la
ye
r
s
,
e
a
c
h
pe
r
f
or
m
in
g
s
pe
c
if
ic
ope
r
a
ti
ons
li
ke
c
onvolut
io
n,
a
c
ti
va
ti
on,
pool
in
g, a
nd f
ul
ly
c
onne
c
te
d l
a
ye
r
s
. F
or
e
xa
m
pl
e
, i
n
th
e
3
rd
la
ye
r
:
I
nput
:
a
t
−
1
w
it
h s
iz
e
(
(
−
1
)
,
[
−
1
]
,
[
−
1
]
)
,
[
0
]
be
in
g t
he
i
m
a
ge
i
n t
he
i
nput
P
a
ddi
ng:
[
]
S
tr
id
e
:
[
]
N
um
be
r
of
f
il
te
r
s
:
[
]
w
he
r
e
e
a
c
h
ha
s
t
he
di
m
e
ns
io
n:
(
[
]
,
[
]
,
[
−
1
]
)
B
ia
s
of
t
he
ℎ
c
onvolut
io
n
[
]
A
c
ti
va
ti
on f
unc
ti
on:
[
]
]
O
ut
put
:
[
]
w
it
h s
iz
e
(
[
]
,
[
]
,
[
]
)
a
nd:
∀
[
1
,
2
,
…
,
[
]
]
(
[
−
1
]
,
)
,
=
[
]
(
∑
∑
∑
,
,
(
)
+
−
1
,
+
−
1
,
+
[
]
[
−
1
]
=
1
[
−
1
]
=
1
[
−
1
]
=
1
)
(
3)
d
im
(
(
[
−
1
]
,
(
1
)
)
)
=
(
[
]
,
[
]
)
]
]
t
hus
:
[
]
=
[
[
]
(
(
[
−
1
]
,
(
1
)
)
)
,
[
]
(
(
[
−
1
]
,
(
1
)
)
)
,
…
,
[
]
(
(
[
−
1
]
,
(
[
]
)
)
)
]
(
4)
d
im
(
[
]
)
=
(
[
]
,
[
]
,
[
]
)
w
it
h:
[
]
=
{
⌊
[
−
1
]
+
[
]
−
[
]
[
]
+
1
⌋
;
>
0
[
−
1
]
+
2
[
]
−
[
]
;
=
0
(
5)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8814
I
nt
J
A
dv A
ppl
S
c
i
,
V
ol
.
14
, N
o.
3
,
S
e
pt
e
m
be
r
20
25
:
724
-
739
730
w
he
r
e
[
]
=
.
T
he
l
e
a
r
ne
d p
a
r
a
m
e
te
r
s
a
t
th
e
ℎ
la
ye
r
a
r
e
:
‒
F
il
te
r
s
w
it
h
(
[
]
×
[
]
×
[
−
1
]
)
×
[
]
pa
r
a
m
e
te
r
s
.
‒
bi
a
s
w
it
h
(
1
×
1
×
1
)
×
[
]
pa
r
a
m
e
te
r
s
(
br
oa
dc
a
s
ti
ng)
.
P
ool
in
g
la
ye
r
s
,
a
s
s
how
n
in
F
ig
ur
e
6,
hi
e
r
a
r
c
hi
c
a
ll
y
a
bs
tr
a
c
t
s
pa
ti
a
l
f
e
a
tu
r
e
s
by
s
ubs
a
m
pl
in
g
in
put
s
w
hi
le
m
a
in
ta
in
in
g c
ha
nne
l
c
ons
i
s
te
nc
y (
[
]
,
[
−
1
]
,
[
−
1
]
)
. T
he
l
a
ye
r
’
s
ope
r
a
ti
on i
s
f
or
m
a
li
z
e
d a
s
f
ol
lo
w
s
:
I
nput
:
[
]
(
a
c
ti
va
ti
on
f
r
om
la
ye
r
−
1
)
w
it
h
di
m
e
ns
io
ns
(
[
]
,
[
−
1
]
,
[
−
1
]
)
,
a
nd
[
0
]
de
not
e
s
th
e
r
a
w
in
put
im
a
ge
.
P
a
r
a
m
e
te
r
s
:
F
il
te
r
s
iz
e
[
]
, s
tr
id
e
[
]
, a
nd pa
ddi
ng
[
]
(
r
a
r
e
ly
a
ppl
ie
d i
n
pool
in
g)
.
P
ool
in
g f
unc
ti
on:
[
]
(
a
ggr
e
ga
ti
on ope
r
a
ti
on, e
.g., ma
x or
a
ve
r
a
ge
)
.
O
ut
put
:
[
]
w
it
h di
m
e
ns
io
ns
(
[
]
,
[
]
,
[
]
=
[
−
1
]
)
T
he
out
put
a
c
ti
va
ti
on a
t
pos
it
io
n
(
,
,
)
in
l
a
ye
r
ll
is
c
om
put
e
d a
s
:
,
,
[
]
=
(
[
−
1
]
)
,
,
=
[
]
(
(
+
−
1
,
+
−
1
,
)
(
,
)
∈
[
1
,
2
,
…
,
[
]
)
(
6)
d
im
(
[
]
)
=
(
[
]
,
[
]
,
[
]
)
w
it
h:
[
]
=
{
⌊
[
−
1
]
+
[
]
−
[
]
[
]
+
1
⌋
;
>
0
[
−
1
]
+
2
[
]
−
[
]
;
=
0
(
7)
[
]
=
[
−
1
]
F
ig
ur
e
6. T
he
p
ool
in
g l
a
ye
r
s
C
N
N
s
a
r
e
hi
e
r
a
r
c
hi
c
a
ll
y
s
tr
uc
tu
r
e
d
a
r
c
hi
te
c
tu
r
e
s
in
F
ig
ur
e
7
th
a
t
it
e
r
a
ti
ve
ly
a
ppl
y
c
onvolut
io
na
l
la
ye
r
s
,
nonl
in
e
a
r
a
c
ti
va
ti
on
f
unc
ti
ons
,
a
nd
pool
in
g
ope
r
a
ti
o
ns
.
T
hi
s
s
e
qu
e
nc
e
c
onvolut
io
n
→a
c
ti
va
ti
on→
pool
in
g
is
r
e
pe
a
te
d
a
c
r
o
s
s
s
u
c
c
e
s
s
iv
e
la
y
e
r
s
to
pr
ogr
e
s
s
iv
e
ly
e
xt
r
a
c
t
hi
ghe
r
-
or
de
r
s
pa
ti
a
l
a
nd
s
e
m
a
nt
ic
f
e
a
tu
r
e
s
f
r
om
in
put
im
a
ge
s
.
T
h
e
e
xt
r
a
c
te
d
f
e
a
tu
r
e
m
a
ps
a
r
e
th
e
n
f
la
tt
e
ne
d
a
nd
f
e
d
in
to
f
ul
ly
c
onne
c
te
d (
de
ns
e
)
l
a
y
e
r
s
, a
ugm
e
nt
e
d w
it
h a
ddi
ti
ona
l
a
c
ti
va
ti
ons
, t
o pe
r
f
or
m
c
la
s
s
if
ic
a
ti
on or
r
e
gr
e
s
s
io
n t
a
s
k
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
dv A
ppl
S
c
i
I
S
S
N
:
2252
-
8814
A
ut
onomous
nav
ig
at
io
n s
y
s
te
m
f
o
r
a r
ov
e
r
w
it
h r
obot
ic
a
r
m
us
i
ng c
onv
ol
ut
io
nal
ne
ur
al
…
(
A
z
iz
E
l
M
r
abe
t)
731
F
ig
ur
e
7. T
he
s
e
que
nt
ia
l
s
te
ps
i
nvol
ve
d i
n a
C
N
N
A
ha
ll
m
a
r
k
of
C
N
N
s
is
th
e
ir
a
bi
li
ty
to
s
ys
te
m
a
ti
c
a
ll
y
r
e
duc
e
s
pa
ti
a
l
r
e
s
ol
ut
io
n
w
hi
le
in
c
r
e
a
s
in
g
c
ha
nne
l
de
pt
h
(
C
N
N
)
a
s
th
e
ne
twor
k
de
e
pe
n
s
,
ba
la
nc
in
g
c
o
m
put
a
ti
ona
l
e
f
f
ic
ie
nc
y
w
it
h
r
e
pr
e
s
e
nt
a
ti
ona
l
c
a
pa
c
it
y.
T
hi
s
th
r
e
e
-
di
m
e
ns
io
n
a
l
tr
a
ns
f
or
m
a
ti
on
f
r
om
r
a
w
pi
xe
l
da
ta
to
a
bs
tr
a
c
t
f
e
a
tu
r
e
hi
e
r
a
r
c
hi
e
s
is
vi
s
ua
li
z
e
d i
n F
ig
ur
e
8
. T
hi
s
f
ig
ur
e
il
lu
s
tr
a
t
es
th
e
a
r
c
hi
te
c
tu
r
a
l
e
vol
ut
io
n of
t
e
ns
or
s
a
c
r
os
s
l
a
ye
r
s
.
F
ig
ur
e
8. S
tr
uc
tu
r
e
of
a
C
N
N
3.
1
.
4
.
R
e
s
N
e
t
R
e
s
N
e
t
s
e
nh
a
nc
e
tr
a
di
ti
ona
l
C
N
N
a
r
c
hi
te
c
tu
r
e
s
by
in
te
gr
a
ti
ng
s
ki
p
c
onne
c
ti
ons
t
ha
t
bypa
s
s
in
te
r
m
e
di
a
te
la
ye
r
s
,
a
s
s
how
n
in
F
ig
ur
e
9
.
T
he
s
e
c
onne
c
ti
ons
m
it
ig
a
te
pe
r
f
or
m
a
nc
e
de
gr
a
da
ti
on
in
v
e
r
y
de
e
p
ne
twor
ks
by
e
na
bl
in
g
th
e
pr
opa
ga
ti
on
of
una
lt
e
r
e
d
gr
a
di
e
nt
s
a
nd
f
e
a
tu
r
e
s
.
W
it
hout
s
ki
p
c
onne
c
ti
ons
,
th
e
be
ha
vi
or
of
a
r
e
s
id
ua
l
bl
oc
k i
s
gove
r
ne
d by s
t
a
nda
r
d l
in
e
a
r
a
nd
nonl
in
e
a
r
t
r
a
ns
f
or
m
a
ti
ons
a
s
i
n (
8)
a
nd (
9)
.
[
]
=
∑
,
[
]
[
−
1
]
+
[
]
−
1
=
1
(
8)
⟶
[
]
=
[
]
(
[
]
)
(
9)
B
y
in
it
ia
li
z
in
g
w
e
ig
ht
s
[
]
a
nd
bi
a
s
e
s
[
]
to
z
e
r
o
a
nd
s
e
le
c
ti
ng
a
n
id
e
nt
it
y
a
c
ti
va
ti
on
[
]
,
th
e
r
e
s
id
ua
l
bl
oc
k
s
im
pl
if
ie
s
to
[
]
=
[
−
2
]
,
pr
e
s
e
r
vi
ng
ne
twor
k
pe
r
f
or
m
a
nc
e
e
ve
n
if
a
dd
e
d
la
ye
r
s
c
ont
r
ib
ut
e
m
in
im
a
ll
y.
T
o
e
ns
ur
e
di
m
e
ns
io
na
l
c
om
pa
ti
bi
li
ty
be
twe
e
n
[
−
2
]
,
a
nd
[
]
,
s
a
m
e
c
onvolut
io
ns
(
m
a
tc
hi
ng
s
pa
ti
a
l
di
m
e
ns
io
n
s
)
a
r
e
ty
pi
c
a
ll
y
a
ppl
ie
d.
W
he
n
m
is
m
a
t
c
he
s
o
c
c
ur
,
a
le
a
r
na
bl
e
pr
oj
e
c
ti
on
te
n
s
or
a
dj
us
ts
th
e
s
ki
p c
onne
c
ti
on
:
⟶
[
]
=
[
]
(
[
]
+
[
−
2
]
)
(
10)
w
he
r
e
m
ig
ht
be
a
f
ix
e
d t
e
ns
or
or
a
l
e
a
r
ne
d one
, a
nd
d
im
(
)
=
[
[
]
,
[
−
2
]
]
.
I
nc
e
pt
io
n
ne
twor
ks
in
tr
oduc
e
a
pa
r
a
di
gm
s
hi
f
t
by
de
pl
oyi
ng
m
odul
e
s
th
a
t
pe
r
f
o
r
m
m
u
lt
ip
le
ope
r
a
ti
ons
in
pa
r
a
ll
e
l
w
it
hi
n
a
s
in
gl
e
l
a
ye
r
,
a
s
s
how
n
in
F
ig
ur
e
10.
S
pe
c
if
ic
a
ll
y,
e
a
c
h
I
nc
e
pt
io
n
m
odul
e
a
ppl
ie
s
c
onvolut
io
na
l,
pool
in
g,
a
nd
f
ul
ly
c
onne
c
t
e
d
ope
r
a
ti
o
ns
s
im
ul
ta
ne
ou
s
ly
,
a
ll
ow
in
g
th
e
ne
twor
k
to
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8814
I
nt
J
A
dv A
ppl
S
c
i
,
V
ol
.
14
, N
o.
3
,
S
e
pt
e
m
be
r
20
25
:
724
-
739
732
c
a
pt
ur
e
f
e
a
tu
r
e
s
a
t
va
r
io
us
s
p
a
ti
a
l
s
c
a
le
s
.
T
hi
s
a
r
c
hi
te
c
tu
r
e
e
li
m
in
a
te
s
th
e
ne
e
d
f
or
m
a
nua
l
s
e
le
c
ti
on
of
l
a
ye
r
ty
pe
s
a
nd e
nha
n
c
e
s
m
ode
l
e
f
f
ic
ie
nc
y
a
nd r
e
pr
e
s
e
nt
a
ti
ona
l
r
ic
hn
e
s
s
.
F
ig
ur
e
9. F
e
e
df
or
w
a
r
d ne
ur
a
l
ne
twor
k a
r
c
hi
te
c
tu
r
e
F
ig
ur
e
10. T
he
r
e
s
ul
t
of
a
ll
t
he
ope
r
a
ti
ons
3.2.
R
ob
ot
m
od
u
le
3.2.1.
R
ob
ot
M
e
c
an
u
m
W
he
e
le
d
m
obi
le
r
obot
s
r
e
s
e
m
bl
in
g
m
a
c
hi
ne
s
a
r
e
be
c
om
in
g
i
nc
r
e
a
s
in
gl
y
pr
e
va
le
nt
a
nd
a
r
e
w
id
e
ly
ut
il
iz
e
d
in
in
dus
tr
ia
l
s
e
tt
in
gs
f
o
r
a
ut
om
a
te
d
t
r
a
ns
por
ta
ti
on
or
l
ogi
s
ti
c
a
l
pur
pos
e
s
,
s
uc
h
a
s
th
e
c
onve
ya
n
c
e
of
goods
,
pa
r
ts
,
a
nd
e
v
e
n
pe
opl
e
.
W
he
n
de
a
li
ng
w
it
h
e
xpe
ns
iv
e
a
nd
s
e
n
s
it
iv
e
lo
a
d
s
,
th
e
m
obi
le
r
obot
m
us
t
b
e
r
e
li
a
bl
e
a
nd
s
a
f
e
w
hi
le
pr
ovi
di
ng
e
f
f
ic
ie
nt
m
ove
m
e
nt
.
A
m
a
c
hi
ne
e
qui
ppe
d
w
it
h
w
he
e
ls
c
a
n
m
a
ne
uve
r
a
r
ound,
le
a
di
ng
to
m
or
e
e
f
f
ic
ie
nt
us
a
ge
.
A
m
obi
le
r
obot
c
a
pa
bl
e
of
s
e
r
vi
ng
m
ul
ti
pl
e
s
ta
ti
on
s
w
it
hi
n
a
pr
oduc
ti
on l
in
e
c
a
n e
nha
nc
e
pr
oduc
t
c
a
p
a
c
it
y a
nd qua
li
ty
[
18]
, [
19]
.
I
n
th
is
r
e
s
e
a
r
c
h,
a
n
om
ni
di
r
e
c
ti
ona
l
r
obot
is
e
m
pl
oye
d,
w
hi
c
h
ha
s
th
e
a
bi
li
ty
to
m
ove
in
a
ny
di
r
e
c
ti
on.
I
t
is
a
hol
onomi
c
r
obo
t
w
it
h
f
ou
r
s
pe
c
ia
li
z
e
d
w
he
e
ls
,
na
m
e
ly
th
e
M
e
c
a
num
w
he
e
l
s
ys
te
m
,
e
a
c
h
dr
iv
e
n
by
a
s
e
p
a
r
a
te
s
te
pp
e
r
m
ot
or
.
F
or
s
uc
h
a
r
obot
,
th
e
num
b
e
r
of
c
ont
r
ol
le
d
de
gr
e
e
s
of
f
r
e
e
dom
is
e
qua
l
to
th
e
to
ta
l
num
be
r
of
de
gr
e
e
s
of
f
r
e
e
dom
of
th
e
r
obot
[
20]
,
[
21]
.
I
t
c
a
n
m
ove
in
a
ny
di
r
e
c
ti
on
on
a
pl
a
na
r
s
ur
f
a
c
e
due
to
it
s
f
r
e
e
ly
r
ot
a
ti
ng
r
ol
le
r
s
pl
a
c
e
d
on
th
e
w
he
e
l
s
ur
f
a
c
e
a
t
a
45
-
de
gr
e
e
a
ngl
e
.
F
ig
ur
e
11
de
pi
c
ts
a
m
obi
le
r
obot
m
ode
l
w
it
h a
M
e
c
a
num
w
he
e
l,
w
it
h a
c
oor
di
na
te
s
ys
te
m
a
tt
a
c
he
d
a
t
th
e
c
e
nt
e
r
of
t
he
w
he
e
l
hub,
w
he
r
e
th
e
uni
t
a
xi
s
is
d
e
not
e
d.
T
h
e
r
obot
'
s
pos
it
io
n
a
nd
or
ie
nt
a
ti
on
a
r
e
r
e
pr
e
s
e
nt
e
d
a
s
.
T
h
e
r
obot
'
s
li
ne
a
r
ve
lo
c
it
y
is
a
nd
it
s
a
ngul
a
r
ve
lo
c
it
y
is
,
w
hi
le
is
th
e
a
ngul
a
r
ve
lo
c
it
y
of
th
e
i
-
th
w
he
e
l
a
nd
r
e
pr
e
s
e
nt
s
th
e
li
ne
a
r
w
he
e
l
ve
lo
c
it
y. T
he
a
ngl
e
be
twe
e
n t
he
f
r
e
e
-
s
li
di
ng r
ol
le
r
a
xi
s
a
nd t
he
w
he
e
l
hub a
xi
s
c
a
n be
e
it
he
r
pos
it
iv
e
o
r
ne
ga
ti
ve
, de
pe
ndi
ng on whe
th
e
r
t
he
w
h
e
e
l
is
l
e
f
t
or
r
ig
ht
[
22]
–
[
24]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
dv A
ppl
S
c
i
I
S
S
N
:
2252
-
8814
A
ut
onomous
nav
ig
at
io
n s
y
s
te
m
f
o
r
a r
ov
e
r
w
it
h r
obot
ic
a
r
m
us
i
ng c
onv
ol
ut
io
nal
ne
ur
al
…
(
A
z
iz
E
l
M
r
abe
t)
733
F
ig
ur
e
11. T
he
di
r
e
c
ti
on of
m
obi
le
r
obot
m
ot
io
n i
s
ba
s
e
d on
th
e
M
e
c
a
num
's
v
a
r
io
us
ve
lo
c
it
ie
s
of
t
he
w
he
e
ls
T
he
di
r
e
c
ti
on
of
th
e
m
obi
le
r
obot
'
s
m
ot
io
n
ba
s
e
d
on
th
e
M
e
c
a
num
w
he
e
l
m
ot
io
n
is
s
how
n
in
F
ig
ur
e
11.
T
he
f
ol
lo
w
in
g
f
or
w
a
r
d
ki
ne
m
a
ti
c
e
qu
a
ti
on
de
s
c
r
ib
e
s
th
e
r
e
la
ti
on
s
hi
p
be
tw
e
e
n
th
e
m
obi
le
r
obot
'
s
ve
lo
c
it
y a
nd t
he
ve
lo
c
it
y of
e
a
c
h w
he
e
l
a
s
i
n (
12)
.
[
]
=
1
4
[
1
1
1
1
−
1
1
−
1
1
1
1
1
1
−
1
+
1
+
1
+
−
1
+
]
[
1
2
3
4
]
(
12)
W
he
r
e
is
t
he
r
a
di
us
of
th
e
M
e
c
a
num
w
he
e
l
,
a
nd a
a
nd b de
s
c
r
i
be
t
he
l
e
ngt
h a
nd w
id
th
of
t
he
m
obi
le
r
obot
.
3.2.2.
K
U
K
A
ar
m
c
on
t
r
ol
D
e
na
vi
t
-
H
a
r
te
nbe
r
g
(
D
H
)
pa
r
a
m
e
te
r
s
a
r
e
a
c
om
m
onl
y
us
e
d
m
e
th
od
f
or
de
s
c
r
ib
in
g
th
e
ki
ne
m
a
ti
c
s
of
m
a
ni
pul
a
to
r
a
r
m
s
in
r
obot
s
[
25]
,
[
26]
.
T
he
y
pr
ovi
de
a
s
ys
t
e
m
a
ti
c
w
a
y
of
r
e
pr
e
s
e
nt
in
g
th
e
r
e
la
ti
ons
hi
p
s
be
twe
e
n
th
e
va
r
io
us
li
nk
s
of
a
n
a
r
ti
c
ul
a
te
d
m
a
ni
pul
a
to
r
a
r
m
.
T
he
D
H
p
a
r
a
m
e
te
r
m
e
th
od
r
e
li
e
s
on
f
our
m
a
in
pa
r
a
m
e
te
r
s
f
or
e
a
c
h
jo
in
t
of
a
m
a
ni
pul
a
to
r
a
r
m
:
th
e
a
ngl
e
of
r
ot
a
ti
on
a
bout
th
e
c
om
m
on
z
-
a
xi
s
(
)
,
th
e
li
nk
le
ngt
h
(
)
,
th
e
di
s
ta
nc
e
a
lo
ng
th
e
x
-
a
xi
s
be
twe
e
n
th
e
c
om
m
on
z
-
a
xe
s
(
)
,
a
nd
th
e
a
ngl
e
of
in
c
li
na
ti
on
to
th
e
c
om
m
on x
-
a
xi
s
(
)
. T
he
homoge
ne
ous
t
r
a
ns
f
or
m
a
ti
on ma
tr
ix
D
H
pa
r
a
m
e
te
r
s
f
or
t
he
K
U
K
A
a
r
m
a
r
e
obt
a
in
e
d
a
c
c
or
di
ng t
o t
hi
s
a
ppr
oa
c
h by the
f
ol
lo
w
in
g r
e
la
ti
ons
hi
p a
nd T
a
bl
e
1.
F
ig
ur
e
12 s
how
s
t
he
K
U
K
A
a
r
m
r
obot
.
=
(
13)
=
[
0
0
−
0
−
0
1
]
(
14)
T
a
bl
e
1. D
H
pa
r
a
m
e
t
e
r
s
f
or
t
he
K
U
K
A
a
r
m
J
oi
nt
i
θ
(
)
(
)
(
d
e
g
)
1
1
889
0
90
2
2
0
2340
0
3
3
0
0
90
4
4
1440
0
-
90
5
5
40
0
-
90
T
he
r
obot
m
ode
l
w
a
s
im
por
te
d
f
r
om
S
ol
id
W
or
ks
us
in
g
a
n
X
M
L
f
il
e
a
nd
th
e
n
in
te
gr
a
te
d
in
to
M
A
T
L
A
B
/S
im
ul
in
k.
S
pe
c
if
ic
a
ll
y,
w
e
us
e
d
th
e
“
s
m
im
por
t”
c
om
m
a
nd,
w
hi
c
h
i
s
pa
r
t
of
th
e
S
im
s
c
a
pe
M
ul
ti
body
to
ol
box,
to
c
onve
r
t
th
e
X
M
L
f
il
e
in
to
a
S
im
ul
in
k
m
ode
l
c
ont
a
in
in
g
th
e
3D
m
e
c
ha
ni
c
a
l
c
om
pone
nt
s
. A
f
te
r
i
m
por
t,
i
nput
s
w
e
r
e
c
onf
ig
ur
e
d t
o c
ont
r
ol
e
a
c
h j
oi
nt
a
ngl
e
, a
nd output
s
w
e
r
e
s
e
t
to
r
e
a
d t
he
m
e
a
s
ur
e
d
va
lu
e
s
.
T
o
a
c
hi
e
ve
pr
e
c
is
e
m
ot
io
n
c
ont
r
ol
,
w
e
im
pl
e
m
e
nt
e
d
a
f
e
e
dba
c
k
lo
op
us
in
g
a
pr
opor
ti
ona
l
-
in
te
gr
a
l
-
de
r
iv
a
ti
ve
(
P
I
D
)
c
ont
r
ol
le
r
to
m
in
im
iz
e
de
vi
a
ti
ons
be
twe
e
n
th
e
c
om
m
a
nd
e
d
a
nd
m
e
a
s
ur
e
d
jo
in
t
a
ngl
e
s
.
E
a
c
h
jo
in
t
in
th
e
m
ode
l
is
th
e
r
e
f
or
e
a
s
s
oc
i
a
te
d
w
it
h
bot
h
a
c
ont
r
ol
in
put
a
nd
a
c
or
r
e
s
ponding
m
e
a
s
ur
e
m
e
nt
out
put
.
Evaluation Warning : The document was created with Spire.PDF for Python.