I
AE
S
I
n
t
e
r
n
at
ion
al
Jou
r
n
al
of
Ar
t
if
icial
I
n
t
e
ll
ig
e
n
c
e
(
I
J
-
AI
)
Vol.
14
,
No.
2
,
Apr
il
2025
,
pp.
91
7
~
92
4
I
S
S
N:
2252
-
8938
,
DO
I
:
10
.
11591/i
jai
.
v
14
.i
2
.
pp
91
7
-
92
4
917
Jou
r
n
al
h
omepage
:
ht
tp:
//
ij
ai
.
iaes
c
or
e
.
c
om
In
t
e
r
ac
t
iv
e
c
o
m
m
u
n
ic
a
t
io
n
h
u
m
an
-
r
o
b
ot
i
n
t
e
r
f
a
c
e
f
o
r
r
e
d
u
c
e
d
m
ob
il
ity p
e
op
le
assi
st
a
n
c
e
Robi
n
s
on
Jim
e
n
e
z
M
or
e
n
o
1
,
Ann
y
As
t
r
id
E
s
p
it
i
a
Cub
il
los
2
,
E
s
p
e
r
an
z
a
Rodrígu
e
z
Car
m
on
a
2
1
M
e
c
h
a
tr
oni
c
E
ngi
ne
e
r
in
g P
r
ogr
a
m
, F
a
c
ul
ty
of
E
ngi
ne
e
r
in
g
, U
ni
ve
r
s
it
y M
il
it
a
r
N
ue
va
G
r
a
na
da
, B
ogot
á
, C
ol
ombi
a
2
I
ndus
tr
ia
l
E
ngi
ne
e
r
in
g P
r
ogr
a
m
, F
a
c
ul
ty
of
E
ngi
ne
e
r
in
g
, U
ni
v
e
r
s
it
y M
il
it
a
r
N
ue
va
G
r
a
na
da
, B
ogot
á
, C
ol
ombi
a
Ar
t
icle
I
n
f
o
AB
S
T
RA
CT
A
r
ti
c
le
h
is
tor
y
:
R
e
c
e
ived
M
a
r
12,
2024
R
e
vis
e
d
Oc
t
22,
2024
Ac
c
e
pted
Nov
14,
2024
Co
mmu
n
i
ca
t
i
o
n
b
e
t
w
ee
n
a
ro
b
o
t
a
n
d
i
t
s
u
s
er
i
s
es
s
e
n
t
i
al
fo
r
t
h
e
ex
ec
u
t
i
o
n
o
f
t
as
k
s
,
ev
en
mo
re
s
o
i
n
a
s
cen
ari
o
w
h
ere
t
h
e
ro
b
o
t
i
s
d
es
i
g
n
e
d
t
o
as
s
i
s
t
p
eo
p
l
e
w
i
t
h
red
u
ced
m
o
b
i
l
i
t
y
.
T
h
i
s
d
o
c
u
men
t
p
re
s
en
t
s
t
h
e
ev
al
u
at
i
o
n
o
f
a
co
n
v
ers
a
t
i
o
n
s
cr
i
p
t
b
et
w
e
en
a
h
u
ma
n
u
s
er
an
d
a
ro
b
o
t
fo
r
as
s
i
s
t
an
ce
u
s
i
n
g
p
re
-
reco
r
d
ed
res
p
o
n
s
e
s
,
fo
r
t
h
i
s
a
met
h
o
d
o
l
o
g
y
w
i
t
h
t
h
ree
p
h
as
e
s
w
a
s
p
ro
p
o
s
ed
a
n
d
a
p
p
l
i
e
d
:
e
s
t
a
b
l
i
s
h
men
t
o
f
t
h
e
t
r
ai
n
i
n
g
s
c
h
eme
o
f
a
co
n
v
o
l
u
t
i
o
n
a
l
n
et
w
o
rk
t
h
at
al
l
o
w
s
reco
g
n
i
ze
u
s
er's
w
o
rd
s
fo
r
ex
ecu
t
i
o
n
o
f
t
as
k
s
b
y
t
h
e
r
o
b
o
t
,
g
en
era
t
i
o
n
o
f
d
i
a
l
o
g
u
e
b
et
w
een
t
h
e
u
s
er
a
n
d
p
o
s
s
i
b
l
e
i
n
t
eract
i
o
n
s
w
i
t
h
t
h
e
as
s
i
s
t
i
v
e
r
o
b
o
t
a
n
d
fi
n
al
l
y
,
t
h
e
meas
u
rem
en
t
o
f
p
ercep
t
i
o
n
o
f
i
n
t
erface
u
s
er
s
.
Res
u
l
t
s
s
h
o
w
a
h
i
g
h
l
ev
el
o
f
acc
u
racy
w
i
t
h
w
o
r
d
s
s
e
l
ect
e
d
t
o
co
mma
n
d
t
h
e
ro
b
o
t
,
u
s
i
n
g
a
co
n
v
o
l
u
t
i
o
n
a
l
n
eu
ra
l
n
et
w
o
r
k
,
w
i
t
h
an
au
d
i
o
i
n
p
u
t
d
i
s
cri
m
i
n
a
t
ed
i
n
i
t
s
co
mp
o
n
en
t
s
mel
freq
u
en
c
y
cep
s
t
ra
l
co
effi
c
i
en
t
s
(MFCCs
)
an
d
co
mma
n
d
s
e
t
s
o
f
mal
e
an
d
f
emal
e
v
o
i
ce
s
.
It
w
as
p
o
s
s
i
b
l
e
t
o
e
s
t
a
b
l
i
s
h
a
d
i
al
o
g
u
e
mo
d
el
w
i
t
h
t
h
ree
s
ce
n
es
t
o
reco
g
n
i
ze
t
h
e
res
i
d
en
t
i
a
l
en
v
i
ro
n
men
t
,
ren
ame
s
p
ace
s
an
d
e
x
ecu
t
e
act
i
o
n
co
mma
n
d
s
t
o
mo
v
e
el
emen
t
s
.
It
i
s
co
n
cl
u
d
e
d
t
h
e
d
es
i
g
n
ed
i
n
s
t
ru
me
n
t
i
s
rel
i
ab
l
e
an
d
t
h
e
p
ercep
t
i
o
n
o
f
p
ro
p
o
s
ed
i
n
t
eract
i
v
e
co
mm
u
n
i
cat
i
o
n
i
n
t
er
f
ace
i
s
g
o
o
d
i
n
t
e
rms
o
f
u
s
a
b
i
l
i
t
y
(effec
t
i
v
en
e
s
s
,
effi
c
i
en
c
y
,
an
d
u
s
er
s
a
t
i
s
fact
i
o
n
).
K
e
y
w
o
r
d
s
:
C
onvolut
ional
ne
twor
ks
I
nter
a
c
ti
ve
c
omm
unica
ti
on
I
nter
f
a
c
e
M
a
n
r
obot
c
omm
unica
ti
on
Us
a
bil
it
y
Th
i
s
i
s
a
n
o
p
en
a
c
ces
s
a
r
t
i
c
l
e
u
n
d
e
r
t
h
e
CC
B
Y
-
SA
l
i
ce
n
s
e.
C
or
r
e
s
pon
din
g
A
u
th
or
:
R
obins
on
J
im
e
ne
z
M
or
e
no
M
e
c
ha
tr
onic
E
nginee
r
ing
P
r
ogr
a
m
,
F
a
c
ult
y
o
f
E
ng
inee
r
ing
,
Unive
r
s
idad
M
il
it
a
r
Nue
va
Gr
a
na
da
C
a
r
r
e
r
a
11
#101
-
80,
B
ogotá
,
C
olom
bia
E
mail:
r
obins
on.
ji
mene
z
@unim
il
it
a
r
.
e
du.
c
o
1.
I
NT
RODU
C
T
I
ON
B
oth
a
r
ti
f
icia
l
int
e
ll
igenc
e
a
nd
r
oboti
c
s
ha
ve
unde
r
gone
e
xpone
nti
a
l
de
ve
lopm
e
nt
in
r
e
c
e
nt
ye
a
r
s
,
with
va
r
ious
a
ppli
c
a
ti
ons
in
s
e
ve
r
a
l
f
ields
o
f
kno
wle
dge
.
I
n
c
ons
ider
a
ti
on
of
[
1]
,
the
us
a
ge
o
f
r
obot
s
is
quit
e
c
omm
on
in
the
p
r
od
uc
ti
ve
s
e
c
tor
,
but
pr
ogr
e
s
s
ha
s
be
e
n
made
r
e
c
e
ntl
y
in
the
s
e
r
vice
s
e
c
tor
,
withi
n
[
2]
,
[
3
]
s
hows
the
im
pa
c
t
pr
oduc
e
d
f
or
s
oc
iety
a
nd
c
ult
ur
e
.
I
ns
ide
[
3]
the
c
ha
nge
s
in
the
pr
oc
e
s
s
e
s
of
c
omm
unica
ti
ve
media
ti
on
be
twe
e
n
man
a
nd
mac
hines
a
r
e
pe
r
c
e
ived,
c
ons
ider
ing
hybr
id
opti
ons
a
nd
the
n
e
e
d
f
or
tr
a
ns
dis
c
ipl
inar
y
wor
k
to
f
ind
mor
e
f
r
iendly
int
e
r
f
a
c
e
s
f
or
man
-
mac
hine
c
omm
unica
ti
on.
Now
mul
ti
ple
a
pps
a
r
e
a
va
il
a
ble
f
or
s
e
r
vice
s
ba
s
e
d
in
r
obots
.
Almeida
[
1]
c
onf
i
r
m
that
r
obo
ts
c
a
n
be
ne
f
it
f
or
the
tour
is
m.
Ac
tually
digi
tal
a
ge
nts
of
tr
a
ve
ls
f
r
om
o
r
ga
niza
ti
ons
s
uc
h
a
s
E
xpe
dia
a
nd
B
ooking,
a
r
e
c
a
pa
ble
of
us
ing
a
ps
e
udo
na
tu
r
a
l
langua
ge
in
it
s
c
omm
unica
ti
on
.
Ga
r
c
ía
[
2]
s
hown
a
ppli
c
a
ti
ons
in
jour
na
li
s
m,
f
o
r
e
xa
mpl
e
,
ins
ide
nor
th
Ame
r
ica
n
media
a
s
T
he
W
a
s
hingt
on
P
os
t,
T
he
W
a
ll
S
tr
e
e
t
J
our
na
l,
F
or
be
s
a
nd
c
onvolut
ional
ne
u
r
a
l
ne
twor
k
,
a
nd
a
uthor
s
pr
opos
e
c
oll
a
bor
a
ti
ve
wo
r
k
man
-
mac
hine
s
im
il
a
r
to
the
one
pr
opos
e
d
withi
n
the
p
r
e
s
e
nt
doc
ument.
M
is
s
e
s
e
t
al.
[
4]
e
xpos
e
s
the
c
ons
tr
uc
ti
on
to
a
r
obot
que
qua
li
f
ies
s
ome
pa
r
a
mete
r
s
pr
e
s
e
nt
in
the
huma
n
body
f
o
r
p
r
oc
e
s
s
of
tea
c
hing
-
lea
r
ning
f
or
a
nonna
ti
ve
langua
ge
.
I
ns
ide
[
5
]
the
us
e
of
r
obots
in
the
he
a
lt
h
s
e
c
tor
,
or
iente
d
to
r
e
pe
ti
ti
ve
o
r
high
-
r
is
k
a
c
ti
vit
ie
s
,
whic
h
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
2
,
Apr
il
2025
:
91
7
-
92
4
918
incr
e
a
s
e
d
a
f
ter
the
pa
nde
mi
c
c
a
us
e
d
by
C
OV
I
D
-
19,
whe
r
e
a
ppli
c
a
ti
ons
c
a
n
be
f
ound
in:
tele
medic
ine,
he
a
lt
hc
a
r
e
,
tele
he
a
lt
h,
onli
ne
c
ouns
e
li
ng,
e
-
He
a
lt
h,
home
c
a
r
e
,
logi
s
ti
c
s
,
p
lanning,
c
oo
r
dinat
ion
a
nd
c
omm
unica
ti
on,
vir
us
p
r
otec
ti
on
a
nd
s
a
nit
a
ti
on,
ongoing
im
muno
logi
c
a
l
a
s
s
e
s
s
ment,
job
r
e
de
s
ign
a
nd
r
e
or
ga
niza
ti
on,
ps
yc
hologi
c
a
l
s
uppor
t
,
a
nd
c
onti
nu
ing
medic
a
l
e
duc
a
ti
on
.
I
n
the
s
a
me
s
e
ns
e
,
S
a
ntos
e
t
al.
[
6
]
indi
c
a
tes
that
r
oboti
c
s
in
th
is
s
e
c
tor
c
a
n
be
us
e
f
ul
f
o
r
both
e
mpl
oye
e
s
a
nd
c
us
tom
e
r
s
,
the
f
i
r
s
t
one
s
,
f
or
e
xa
mpl
e
,
to
a
void
the
e
xe
c
uti
on
of
r
e
s
pe
c
ti
ve
tas
ks
,
whe
n
c
a
r
r
ying
out
s
ur
ge
r
i
e
s
pr
e
c
is
e
ly
(
with
or
without
s
pe
c
ialis
t
s
upe
r
vis
ion)
,
a
s
we
ll
a
s
f
or
c
ontr
oll
ing
medic
a
ti
on
s
upply.
T
he
latter
,
f
or
e
xa
mpl
e
,
with
t
he
c
ontr
ol
of
the
a
dmi
nis
tr
a
ti
on
of
medic
a
ti
ons
a
nd
by
p
r
ovid
ing
e
mot
ional
s
uppor
t
in
ps
yc
hologi
c
a
l
t
r
e
a
tm
e
nts
.
R
e
s
e
a
r
c
h
by
C
a
s
ti
ll
o
e
t
al.
[
5]
,
a
s
e
r
vice
r
obot
is
de
f
ined
a
s
a
n
e
nti
ty
wi
th
ba
s
ic
be
ha
vior
s
us
e
d
f
o
r
c
ompl
e
x
tas
ks
in
r
e
a
l
e
nvir
onments
c
a
pa
ble
of
int
e
r
a
c
ti
ng
thr
ough
int
e
r
f
a
c
e
s
that
c
a
n
be
gr
a
phica
l
o
r
na
tur
a
l
(
langua
ge
,
ge
s
ti
c
ulation,
e
ye
tr
a
c
king
,
f
a
c
ial
e
xpr
e
s
s
ion,
e
mot
ional
be
ha
vior
,
a
nd
langua
ge
body)
.
Ac
c
or
ding
to
P
ineda
-
de
-
Alc
á
z
a
r
[
3]
,
human
-
mac
hine
int
e
r
a
c
ti
ons
a
r
e
a
na
lyze
d
a
t
a
c
omm
un
ica
ti
ve
le
ve
l.
T
he
ludi
bot
c
r
e
a
ted
by
[
4]
int
e
gr
a
tes
ve
r
ba
l
a
nd
no
n
-
ve
r
ba
l
c
omm
unica
ti
on
in
human
-
r
obot
int
e
r
a
c
t
ion.
T
he
a
uthor
s
a
f
f
i
r
m
that
the
im
pleme
ntation
of
r
obots
mus
t
be
s
uppor
ted
by
the
c
ons
ider
a
ti
ons
of
the
d
e
s
ign
of
biom
e
dica
l
pr
oduc
ts
a
nd
ge
ne
r
a
l
c
r
it
e
r
ia,
s
uc
h
a
s
us
e
r
s
,
us
a
bil
it
y,
m
a
nuf
a
c
tur
ing,
c
yc
le
pr
oduc
t
li
f
e
,
a
nd
s
a
f
e
ty
guidelines
f
or
r
obots
f
or
pe
r
s
ona
l
us
e
.
T
he
W
or
ld
He
a
lt
h
Or
ga
niza
ti
on
(
W
HO
)
[
7]
r
e
por
ts
1.
3
b
il
li
on
pe
ople
ha
ve
s
ome
majo
r
dis
a
bil
it
y
,
that
c
or
r
e
s
ponds
to
16
%
o
f
the
population.
E
mer
g
ing
oppor
tun
it
ies
f
o
r
the
a
ppli
c
a
ti
on
o
f
r
oboti
c
s
to
s
uppor
t
he
a
lt
hc
a
r
e
may
be
r
e
f
lec
ted
in
r
e
duc
ing
medic
a
l
c
os
ts
a
nd
incr
e
a
s
ing
c
onve
nienc
e
f
or
pa
ti
e
nt
us
e
r
s
.
M
a
ny
of
them
ne
e
d
a
s
s
is
tanc
e
,
a
c
c
or
ding
to
[
6]
,
whe
r
e
they
r
e
por
t
t
ha
t
the
mos
t
f
r
e
que
nt
us
e
of
r
oboti
c
s
in
hu
man
c
a
r
e
is
ther
a
py
a
nd
e
nter
tainment,
a
nd
the
mos
t
us
e
d
mea
ns
of
int
e
r
a
c
ti
on
be
twe
e
n
humans
a
nd
r
obots
is
ve
r
ba
l
c
omm
unica
ti
on,
whic
h
is
us
e
f
ul
in
he
lpi
ng
the
e
l
de
r
ly,
c
hil
d
r
e
n
a
nd
pe
ople
with
a
menta
l
he
a
lt
h
dis
or
de
r
.
Huma
n
-
r
obot
int
e
r
a
c
ti
on
is
a
n
im
po
r
tant
a
s
pe
c
t
in
t
he
int
r
oduc
ti
on
o
f
r
obots
in
da
il
y
human
a
c
ti
vit
ies
.
Ac
c
or
ding
to
S
c
hne
ider
e
t
al.
[
8]
,
the
ne
e
d
to
br
e
a
k
the
ba
r
r
ie
r
o
f
the
us
e
of
r
obots
f
r
om
a
c
a
de
mi
a
to
the
c
ompany
is
e
s
tablis
he
d
unde
r
the
us
e
of
s
oc
ial
r
oboti
c
s
.
T
his
f
ield
pr
e
s
e
nts
de
ve
lopm
e
nts
s
uc
h
a
s
tea
c
hing
r
obots
f
or
s
c
ienc
e
e
xplana
ti
ons
in
mus
e
ums
[
9]
,
r
obots
a
s
r
e
c
e
pti
onis
ts
[
10]
,
a
s
a
s
s
is
tants
f
or
manipul
a
ti
ng
a
n
e
leva
tor
[
11]
,
or
f
or
pur
c
ha
s
in
g
it
e
ms
[
12]
.
I
n
a
ll
thes
e
c
a
s
e
s
,
the
us
e
of
human
-
r
obot
int
e
r
a
c
ti
on
in
a
pr
a
c
ti
c
a
l
wa
y
is
e
vident
,
a
s
in
[
13
]
.
T
he
a
s
s
is
tanc
e
of
r
obots
to
pe
ople
in
a
s
uppor
t
s
i
tuation,
s
uc
h
a
s
that
r
e
s
ult
ing
f
r
om
the
C
OV
I
D
19
pa
nde
mi
c
,
whic
h
r
e
quir
e
s
them
to
be
is
olate
d,
is
vit
a
l
if
ther
e
a
r
e
mobi
li
ty
li
mi
tations
or
the
inabili
ty
to
pick
up
a
n
objec
t.
De
r
ived
f
r
om
the
a
bove
,
th
is
wor
k
pr
e
s
e
nts
the
de
ve
lopm
e
nt
of
a
dialogue
model
f
or
a
c
a
r
e
r
obot
in
a
r
e
s
idential
e
nvir
onment
that
int
e
r
a
c
ts
s
oc
iall
y
with
a
us
e
r
.
S
o,
the
a
im
is
to
a
s
s
e
s
s
the
us
a
bil
it
y
of
the
i
nter
a
c
ti
ve
man
-
r
obot
c
omm
unica
ti
on
int
e
r
f
a
c
e
de
s
igned
to
a
s
s
is
t
p
e
ople
wit
h
r
e
duc
e
d
mobi
li
ty.
C
onvolut
ional
ne
ur
a
l
ne
twor
ks
ha
ve
r
e
c
e
ntl
y
be
e
n
us
e
d
to
im
pr
ove
the
inf
e
r
e
nc
e
of
int
e
r
r
oga
ti
ve
s
e
ntenc
e
s
in
dialogue
s
with
r
obots
[
14]
,
to
maintain
or
f
oll
ow
a
dialogue
[
15]
,
[
16]
a
nd
in
ge
ne
r
a
l
to
a
ll
ow
human
-
r
obot
int
e
r
a
c
ti
on
b
y
voice
c
omm
a
nds
[
17]
or
c
onve
r
s
a
ti
ons
[
18]
.
T
hi
s
is
why
the
dialogue
model
pr
e
s
e
nted
in
thi
s
wor
k
us
e
s
c
o
nvolut
ional
ne
tw
or
ks
to
r
e
c
ognize
wor
ds
o
f
a
f
f
i
r
m
a
ti
on
or
de
nial
a
nd
a
c
ti
on
by
the
us
e
r
,
in
a
c
ti
ons
of
c
a
r
r
yin
g
a
nd/or
b
r
ingi
ng
s
pe
c
if
ic
objec
ts
.
T
he
s
tr
uc
tur
e
of
thi
s
doc
ument
c
ons
is
ts
of
a
s
tate
of
the
a
r
t
in
the
pr
e
s
e
nt
int
r
oduc
tor
y
s
e
c
ti
on,
f
oc
us
ing
on
the
us
e
of
r
obo
ts
a
nd
human
-
r
obot
i
nter
a
c
ti
on,
in
s
e
c
ti
on
2
the
methodology
us
e
d
is
pr
e
s
e
nted
thr
ough
a
f
low
c
ha
r
t
.
S
e
c
ti
on
3
p
r
e
s
e
nts
the
r
e
s
ult
s
obtaine
d
in
the
de
s
ign
of
the
c
onvolut
ional
ne
twor
ks
f
or
the
r
e
c
ognit
ion
of
us
e
r
o
r
de
r
s
a
nd
the
d
ialogue
m
ode
l
f
or
the
r
obot
ic
c
omm
a
nd,
whic
h
s
ubs
e
que
ntl
y
a
ll
ows
us
to
e
va
luate
a
nd
pr
e
s
e
nt
the
us
a
bil
it
y
model
o
f
th
e
dialogue
.
F
inally
,
the
c
onc
lus
ions
a
r
e
pr
e
s
e
nted.
2.
M
E
T
HO
D
T
o
a
c
hieve
the
p
r
ojec
t
objec
ti
ve
of
e
va
luating
a
p
r
opos
e
d
model
of
c
onve
r
s
a
ti
on
s
c
r
ipt
be
twe
e
n
man
a
nd
r
o
bot
unde
r
pr
e
r
e
c
or
de
d
r
e
s
pons
e
s
,
a
thr
e
e
-
pha
s
e
methodology
wa
s
u
s
e
d.
T
he
f
ir
s
t
pha
s
e
c
or
r
e
s
ponds
to
the
tr
a
ini
ng
s
c
he
me
o
f
it
s
own
c
onvolut
ional
ne
tw
or
k
that
a
ll
ows
the
r
e
c
ognit
ion
of
the
us
e
r
's
ba
s
e
wor
ds
f
or
the
e
xe
c
uti
on
of
tas
ks
by
the
r
obot
.
F
o
r
f
ir
s
t
ph
a
s
e
,
the
a
im
is
to
e
s
tablis
h
the
tr
a
ini
ng
s
c
he
me
f
or
two
c
onvolut
ional
ne
ur
a
l
ne
twor
ks
,
f
or
whic
h
the
s
tar
ti
ng
point
is
to
de
ter
mi
ne
the
ba
s
ic
a
c
ti
ons
that
the
r
obot
c
a
n
e
xe
c
ute,
a
nd
in
thi
s
c
a
s
e
,
it
will
be
to
a
s
s
is
t
a
p
e
r
s
on
in
c
ondit
ions
of
r
e
duc
e
d
mo
bil
it
y
or
qua
r
a
nt
ine
-
type
is
olation
in
a
r
e
s
idential
e
nvi
r
onment.
T
he
r
obot's
a
c
ti
ons
will
be
to
c
a
r
r
y
a
nd
br
ing
ba
s
ic
objec
ts
s
uc
h
a
s
a
towe
l,
a
glas
s
or
a
medic
ine,
be
ing
a
c
ti
va
ted
by
the
wor
d
“
R
obot”
a
nd
c
onf
ir
mi
ng
the
c
omm
a
nds
de
s
ir
e
d
by
the
us
e
r
a
nd
i
nte
r
pr
e
ted
by
ba
s
ic
a
f
f
i
r
mation
“
Ye
s
”
or
de
nial
“
No”
.
C
ons
ider
ing
wha
t
wa
s
point
e
d
o
ut
by
[
2]
,
a
r
ti
f
icia
l
int
e
ll
igenc
e
invol
ve
s
the
int
e
r
a
c
ti
on
of
a
us
e
r
with
a
r
obot
that,
s
ince
it
is
not
c
ompl
e
tely
human,
r
e
quir
e
s
media
ti
on
s
of
twa
r
e
,
whic
h
t
r
a
ns
f
or
ms
it
i
nto
a
mor
e
r
a
ti
ona
l
than
e
mot
ional
c
omm
un
ica
ti
ve
pr
oc
e
s
s
,
with
li
mi
ted
opti
ons
.
F
or
s
e
c
ond
pha
s
e
,
a
s
c
r
ipt
m
ode
l
is
e
s
tablis
he
d
to
guide
how
a
us
e
r
will
a
s
k
the
r
obot
to
c
a
r
r
y
or
br
ing
a
n
objec
t.
F
or
thi
s
pur
pos
e
,
a
mi
ni
mum
int
e
r
a
c
ti
on
is
s
ought
in
the
f
o
r
m
of
a
s
hor
t
dialogue
guided
by
the
r
obot
.
F
o
r
thi
r
d
pha
s
e
,
human
-
r
o
bot
int
e
r
a
c
ti
on
is
e
va
luate
d
thr
ough
the
dialogu
e
model
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
I
nter
ac
ti
v
e
c
omm
unication
human
-
r
obot
int
e
r
face
for
r
e
duc
e
d
mobili
ty
…
(
R
obins
on
J
ime
ne
z
-
M
or
e
no
)
919
a
ppli
e
d
by
a
gr
oup
o
f
us
e
r
s
.
T
his
c
or
r
e
s
ponds
to
us
a
bil
it
y
pha
s
e
without
the
r
e
a
l
r
oboti
c
a
c
ti
on,
only
c
onve
r
s
a
ti
ona
l
one
.
W
hich
a
l
lows
us
e
r
s
to
a
ns
we
r
a
s
ur
ve
y
a
bout
int
e
r
a
c
ti
on
whe
r
e
they
indi
c
a
te
their
pe
r
c
e
pti
on
on
a
L
iker
t
s
c
a
le.
L
iker
t
s
c
a
les
a
r
e
a
s
e
t
of
opti
ons
in
whic
h
indi
viduals
mus
t
c
hoos
e
their
leve
l
of
a
gr
e
e
ment
with
a
s
e
r
ies
of
s
tate
ments
.
T
he
y
a
r
e
metr
ics
c
omm
only
us
e
d
in
human
-
r
obot
int
e
r
a
c
ti
on
to
mea
s
ur
e
pe
r
c
e
pti
ons
a
nd
a
tt
it
ude
s
[
19]
.
T
he
n,
the
C
r
onba
c
h's
Alpha
va
lue
i
s
c
a
lcula
ted
to
v
e
r
if
y
the
r
e
li
a
bil
it
y
o
f
the
s
ur
ve
y,
whic
h
a
c
c
or
ding
to
[
19
]
i
t
mea
s
ur
e
s
the
int
e
r
na
l
c
ons
is
tenc
y
of
a
s
c
a
le;
if
it
s
va
lue
is
gr
e
a
ter
than
0.
7
,
it
is
us
ua
ll
y
c
ons
ider
e
d
a
c
c
e
ptable
.
F
inally,
a
n
a
na
lys
is
of
the
a
ns
we
r
s
obtaine
d
in
e
a
c
h
que
s
ti
on
is
c
a
r
r
ied
out
us
ing
de
s
c
r
ipt
ive
s
tatis
ti
c
s
.
3.
RE
S
UL
T
S
AN
D
DI
S
CU
S
S
I
ON
3.
1.
F
irs
t
p
h
as
e
:
E
s
t
ab
li
s
h
m
e
n
t
of
t
h
e
t
r
ain
in
g
s
c
h
e
m
e
o
f
it
s
own
c
on
vol
u
t
ion
al
n
e
t
wor
k
t
h
at
all
ows
r
e
c
ogn
izin
g
t
h
e
u
s
e
r
's
b
as
e
wor
d
s
f
or
t
h
e
e
xe
c
u
t
ion
of
t
as
k
s
b
y
t
h
e
r
ob
o
t
T
o
r
e
c
ognize
the
gr
oup
of
s
e
lec
ted
wor
ds
that
a
ll
ow
the
r
obot
to
be
c
omm
a
nde
d,
whic
h
a
r
e
tabu
late
d
in
T
a
ble
1,
the
de
s
ign
of
a
pe
r
s
ona
li
z
e
d
a
r
c
hit
e
c
tur
e
of
a
c
onvolut
ional
ne
ur
a
l
ne
twor
k
pr
e
s
e
nted
i
n
[
20]
is
us
e
d,
whic
h
ha
s
s
ix
c
onvolut
ion
laye
r
s
.
F
or
thi
s
n
e
twor
k,
a
n
a
udio
input
dis
c
r
im
inate
d
in
it
s
mel
f
r
e
que
nc
y
c
e
ps
t
r
a
l
c
oe
f
f
icie
nts
(
M
F
C
C
s
)
c
omponents
is
us
e
d
a
t
a
f
r
e
que
nc
y
of
16000
Hz
.
S
e
ts
o
f
male
a
nd
f
e
ma
le
voice
c
omm
a
nds
a
r
e
us
e
d
in
e
qua
l
p
r
opor
ti
on;
F
igur
e
1
il
lus
tr
a
tes
s
pe
c
tr
ogr
a
ms
of
the
“
r
obot
c
a
r
r
y
m
e
dicine
”
c
omm
a
nd,
in
F
igur
e
1
(
a
)
the
male
voice
a
nd
in
F
ig
ur
e
1(
b)
the
f
e
male
voice
.
T
a
ble
1.
Gr
oup
o
f
wo
r
ds
s
e
lec
ted
to
c
omm
a
nd
the
r
obot
A
c
ti
on
O
bj
e
c
t
R
obot
--
S
to
p
--
T
r
a
e
r
P
a
pe
l/
T
oa
ll
a
/
V
a
s
o /
M
e
di
c
in
a
L
le
va
r
P
a
pe
l/
T
oa
ll
a
/
V
a
s
o /
M
e
di
c
in
a
(
a
)
(
b)
F
igur
e
1.
C
omm
a
nd
r
obo
t
c
a
r
r
y
medic
ine
f
or
(
a
)
f
e
male
voice
a
nd
(
b
)
male
vo
ice
[
20]
F
or
the
r
e
c
ognit
ion
of
the
ye
s
a
nd
no
r
e
s
pons
e
c
omm
a
nds
,
the
s
a
me
c
onvolut
ional
a
r
c
hit
e
c
tur
e
of
s
i
x
c
onvolut
ion
laye
r
s
is
us
e
d,
but
with
a
da
taba
s
e
ba
s
e
d
on
thes
e
c
omm
a
nds
,
whe
r
e
the
r
e
s
pon
s
e
is
va
li
da
ted
only
if
the
c
onf
idenc
e
leve
l
obtaine
d
a
t
the
output
of
the
ne
twor
k
is
gr
e
a
ter
than
90%
.
T
he
us
e
of
thi
s
s
e
pa
r
a
te
ne
twor
k
is
due
to
the
r
e
duc
e
d
ti
me
a
nd
pa
r
a
m
e
ter
va
r
iation
p
r
e
s
e
nted
by
ye
s
/no
r
e
s
pons
e
s
,
whic
h
a
r
e
c
onf
us
e
d
with
phone
mes
of
l
onge
r
wor
ds
a
nd/or
phr
a
s
e
s
whe
n
us
ing
the
s
a
me
ne
twor
k.
F
igur
e
2
il
lus
tr
a
tes
the
c
onf
us
ion
matr
ice
s
of
both
ne
twor
ks
,
in
F
igu
r
e
2(
a
)
the
r
e
s
ult
s
a
r
e
pr
e
s
e
nted
f
or
the
c
omm
a
nd
s
a
nd
in
F
igur
e
2(
b
)
f
or
the
r
e
s
pons
e
s
,
in
both
c
a
s
e
s
the
high
leve
l
of
pr
e
c
is
ion
a
c
hieve
d
c
a
n
be
obs
e
r
ve
d.
3.
2.
S
e
c
on
d
p
h
as
e
:
Gener
at
io
n
of
t
h
e
d
ial
ogu
e
m
od
e
l
t
h
a
t
wil
l
b
e
c
ar
r
ied
ou
t
b
e
t
we
e
n
t
h
e
u
s
e
r
an
d
t
h
e
p
os
s
ib
le
in
t
e
r
ac
t
ion
s
wit
h
t
h
e
as
s
is
t
an
c
e
r
ob
ot
T
he
s
e
c
ond
pha
s
e
o
f
the
methodology
c
or
r
e
s
ponds
to
the
dialogue
model
to
be
c
a
r
r
ied
out
be
twe
e
n
the
us
e
r
a
nd
the
r
obot.
T
o
do
thi
s
,
the
r
obo
t
a
c
ti
v
a
ti
on
a
nd
de
a
c
ti
va
ti
on
c
ondit
ions
a
r
e
e
s
tablis
he
d
us
ing
the
“
r
obot”
a
nd
“
s
top”
c
omm
a
nds
r
e
s
pe
c
ti
ve
ly.
T
he
i
ni
ti
a
l
pa
r
t
c
or
r
e
s
ponds
to
the
r
obot's
a
bil
it
y
to
na
v
igate
the
e
nvir
onment
to
r
e
c
ognize
the
s
pa
c
e
s
in
the
r
e
s
ide
nti
a
l
e
nvir
onment
in
whic
h
it
will
int
e
r
a
c
t,
a
s
de
s
c
r
ibed
in
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
2
,
Apr
il
2025
:
91
7
-
92
4
920
[
21]
.
T
his
s
tage
is
c
a
ll
e
d
“
no
ini
c
ializa
do”
,
that
is
,
th
e
r
obot
mus
t
e
xplor
e
the
r
e
s
idential
e
nvi
r
on
ment
to
identif
y
the
s
pa
c
e
s
whe
r
e
it
will
move
a
nd
a
s
s
ign
e
a
c
h
a
r
e
a
a
labe
l,
f
or
e
xa
mpl
e
,
ba
thr
oom
1
,
be
dr
oo
m
2,
thi
s
int
e
r
a
c
ti
on
c
or
r
e
s
ponds
to
the
dialogue
in
s
c
e
ne
1,
a
s
f
oll
ows
:
S
C
E
N
E
1
E
S
T
A
D
O
N
O
I
N
I
C
I
A
L
I
Z
A
D
O
U
S
U
A
R
I
O
:
R
obot
R
O
B
O
T
:
H
ol
a
, ¿
va
mos
a
r
e
c
onoc
e
r
e
l
a
mbi
e
nt
e
?
C
ont
e
s
ta
S
i
o N
o
U
S
U
A
R
I
O
:
S
i
R
O
B
O
T
:
R
e
c
onoc
im
ie
nt
o f
in
a
li
z
a
do, ¿
g
e
ne
r
o un r
e
por
te
?
C
ont
e
s
ta
S
i
o N
o
U
S
U
A
R
I
O
:
N
o
E
s
ta
do R
e
pos
o
S
C
E
N
E
1.1
U
S
U
A
R
I
O
:
R
obot
R
O
B
O
T
:
R
e
c
onoc
i
mi
e
nt
o f
in
a
li
z
a
do, ¿
g
e
ne
r
o un r
e
por
te
?
C
ont
e
s
ta
S
i
o N
o
U
S
U
A
R
I
O
:
S
i
R
O
B
O
T
:
E
nc
ont
r
é
e
n e
l
a
mbi
e
nt
e
# e
s
p
a
c
io
s
, e
s
pa
c
io
1, …, E
tc
.
R
O
B
O
T
:
¿
D
e
s
e
a
c
a
mbi
a
r
de
nombr
e
a
lg
ún e
s
pa
c
io
?
C
ont
e
s
t
a
S
i
o N
o
U
S
U
A
R
I
O
:
N
o (
go t
o s
c
e
n 3)
(
a
)
(
b)
F
igur
e
2.
C
onf
us
ion
matr
ix
of
(
a
)
c
omm
a
nds
a
nd
(
b)
a
ns
we
r
s
Af
ter
that,
the
r
obot
wil
l
a
ll
ow
you
to
r
e
a
s
s
ign
na
mes
to
the
s
pa
c
e
s
to
f
a
c
il
it
a
te
int
e
r
a
c
ti
on
wi
th
the
us
e
r
in
their
r
e
s
idential
e
nvir
onment.
T
ha
t
is
,
the
us
e
r
will
be
a
ble
to
r
e
a
s
s
ign
labe
ls
to
the
s
pa
c
e
s
f
ound,
f
or
e
xa
mpl
e
,
ins
tea
d
of
ba
thr
oo
m
1
,
main
ba
th
r
oom,
ins
tea
d
of
be
dr
oo
m
1
,
it
would
be
P
e
dr
o's
be
d
r
oo
m.
T
his
int
e
r
a
c
ti
on
c
or
r
e
s
ponds
to
the
dialogue
in
s
c
e
ne
2,
a
s
f
oll
ows
:
S
C
E
N
E
2
U
S
U
A
R
I
O
:
R
obot
R
O
B
O
T
:
E
nc
ont
r
é
e
n e
l
a
mbi
e
nt
e
# e
s
p
a
c
i
os
, e
s
pa
c
io
1, …, E
tc
.
R
O
B
O
T
:
¿
D
e
s
e
a
c
a
mbi
a
r
de
nombr
e
a
lg
ún e
s
pa
c
io
?
C
ont
e
s
t
a
S
i
o N
o
U
S
U
A
R
I
O
:
N
o (
go t
o s
c
e
n 3)
S
C
E
N
E
2.1
U
S
U
A
R
I
O
:
S
i
R
O
B
O
T
:
¿
C
uá
l
e
s
e
l
nombr
e
de
l
e
s
p
a
c
io
que
va
mo
s
a
c
a
mbi
a
r
?
E
s
c
oge
s
ol
o un e
s
p
a
c
io
1…... E
tc
.
U
S
U
A
R
I
O
:
E
s
pa
c
io
#
R
O
B
O
T
:
¿
C
uá
l
e
s
e
l
nue
vo nombr
e
?
U
S
U
A
R
I
O
:
E
s
pa
c
io
N
ue
vo
R
O
B
O
T
:
C
a
mbi
a
r
é
e
l
e
s
pa
c
io
# por
e
l
E
s
pa
c
io
N
u
e
vo, ¿
s
e
gur
o
qui
e
r
e
s
c
a
mbi
a
r
lo
?
C
ont
e
s
ta
S
i
o N
o
U
S
U
A
R
I
O
:
N
o
R
O
B
O
T
:
¿
D
e
s
e
a
c
a
mbi
a
r
ot
r
o nombr
e
de
e
s
p
a
c
io
?
C
ont
e
s
ta
S
i
o N
o
U
S
U
A
R
I
O
:
N
o
E
S
T
A
D
O
N
O
I
N
I
C
I
A
L
I
Z
A
D
O
U
S
U
A
R
I
O
:
S
I
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
I
nter
ac
ti
v
e
c
omm
unication
human
-
r
obot
int
e
r
face
for
r
e
duc
e
d
mobili
ty
…
(
R
obins
on
J
ime
ne
z
-
M
or
e
no
)
921
T
he
f
inal
a
nd
mos
t
r
e
c
ur
r
ing
s
c
e
ne
will
be
the
us
ua
l
int
e
r
a
c
ti
on
to
e
xe
c
ute
c
a
r
r
y
a
nd
br
ing
a
c
ti
on
c
o
mm
a
nds
,
unde
r
the
f
o
ll
owing
s
c
r
ipt
:
S
C
E
N
E
3
(
f
or
a
ny c
omm
a
nd)
E
S
T
A
D
O
I
N
I
C
I
A
L
I
Z
A
D
O
***
U
S
U
A
R
I
O
:
R
obot
t
r
a
e
r
t
oa
ll
a
…
me
di
c
a
me
nt
o, pa
pe
l…
R
O
B
O
T
:
C
on gus
to
, i
ndi
que
ubi
c
a
c
ió
n
**
U
S
U
A
R
I
O
:
B
a
ño
R
O
B
O
T
:
U
bi
c
a
c
ió
n no r
e
c
onoc
id
a
, r
e
pi
ta
por
f
a
vor
**
R
O
B
O
T
:
¿
U
s
te
d de
s
e
a
t
r
a
e
r
t
oa
ll
a
d
e
ba
ño… me
di
c
a
me
nt
o, p
a
pe
l…
?
R
O
B
O
T
:
¿
e
s
e
s
te
c
oma
ndo v
á
li
do?
S
i
o no
U
S
U
A
R
I
O
:
S
i
U
S
U
A
R
I
O
:
N
o
R
O
B
O
T
:
P
or
f
a
vor
r
e
pi
ta
c
oma
ndo
***
3.
3.
T
h
i
r
d
p
h
as
e
:
M
e
as
u
r
e
m
e
n
t
o
f
t
h
e
p
e
r
c
e
p
t
ion
o
f
var
io
u
s
u
s
e
r
s
r
e
ga
r
d
i
n
g
t
h
e
u
s
a
b
i
l
it
y
o
f
t
h
e
in
t
e
r
f
ac
e
T
a
king
int
o
a
c
c
ount
that
a
c
c
or
ding
to
[
4]
withi
n
the
c
ons
ider
a
ti
ons
o
f
the
de
s
ign
of
biom
e
dica
l
pr
oduc
ts
,
us
a
bil
it
y
is
c
ounted
a
s
one
o
f
the
e
va
lua
ti
on
c
r
i
ter
ia
f
or
r
obots
f
o
r
pe
r
s
ona
l
us
e
.
T
his
c
r
i
ter
ion
wi
ll
be
a
ppli
e
d
in
thi
s
pa
r
ti
c
ular
c
a
s
e
,
to
e
va
luate
the
r
obot
in
a
s
s
is
ti
ng
us
e
r
s
in
the
e
s
tablis
he
d
s
c
e
ne
s
.
About
,
F
e
r
r
e
ir
a
e
t
al
.
[
22]
a
f
f
i
r
ms
that
us
a
bil
it
y
is
c
r
uc
ial
a
s
it
is
a
c
ha
r
a
c
ter
is
ti
c
of
pr
oduc
t
qua
li
ty
by
a
c
hi
e
ving
it
s
pur
pos
e
with
e
f
f
e
c
ti
ve
ne
s
s
,
e
f
f
icie
nc
y
a
nd
us
e
r
s
a
ti
s
f
a
c
ti
on;
I
ts
c
onc
e
pt
wa
s
popular
ize
d
than
ks
to
the
s
tanda
r
d
I
S
O
9241
-
11:2018
[
23]
.
Als
o,
Albo
r
noz
e
t
al.
[
24]
point
s
out
that
us
a
bil
it
y
ha
s
be
c
ome
pa
r
a
mount
f
or
s
of
twa
r
e
qua
li
ty
.
T
o
e
va
luate
the
us
a
bil
it
y
o
f
the
us
e
r
's
int
e
r
a
c
ti
on
with
the
r
obo
t,
unde
r
the
e
s
t
a
bli
s
he
d
dialogue
model,
the
model
in
F
igur
e
3
is
p
r
opos
e
d,
whic
h
r
e
late
s
the
thr
e
e
us
a
bil
it
y
c
r
it
e
r
ia
de
f
ine
d
by
the
s
tanda
r
d
I
S
O
9241
-
11:2018
[
23
]
with
the
va
r
iable
s
pr
opos
e
d.
F
igur
e
3.
R
obot
us
e
r
c
onve
r
s
a
ti
on
us
a
bil
it
y
e
va
luation
model
F
or
thi
s
,
a
pr
e
li
mi
na
r
y
e
va
luation
ins
tr
ument
is
g
e
ne
r
a
ted
that
c
or
r
e
s
ponds
to
a
us
e
r
s
ur
ve
y
that
is
a
ppli
e
d
a
f
ter
us
ing
the
de
s
igned
int
e
r
f
a
c
e
in
a
c
ontr
oll
e
d
e
nvir
on
ment.
T
he
s
ur
ve
y
s
e
e
ks
to
k
now
the
pe
r
c
e
pti
on
of
us
e
r
s
on
a
L
iker
t
s
c
a
le
f
r
om
1
to
7,
r
e
g
a
r
ding
e
f
f
icie
nc
y,
e
f
f
e
c
ti
ve
ne
s
s
a
nd
f
inally
us
e
r
s
a
ti
s
f
a
c
ti
on
r
e
ga
r
ding
the
c
onve
r
s
a
ti
on
with
the
r
obot.
Addit
ionally,
the
voice
pr
e
f
e
r
e
nc
e
f
o
r
the
r
obot
is
inves
ti
ga
ted.
T
a
ble
2
li
s
ts
the
que
s
ti
ons
,
va
r
iable
s
a
nd
c
r
it
e
r
ia
of
the
s
ur
ve
y.
T
his
pr
e
li
mi
na
r
y
ve
r
s
ion
of
the
ins
tr
ument
wa
s
a
ppli
e
d
to
a
n
ini
ti
a
l
s
a
mpl
e
o
f
10
p
e
ople,
obtaining
the
ge
ne
r
a
l
r
e
s
ult
s
li
s
ted
in
T
a
ble
3.
T
he
a
ve
r
a
ge
e
xe
c
uti
on
a
nd
r
e
s
pons
e
ti
me
f
or
th
e
que
s
ti
onna
ir
e
wa
s
10
mi
nutes
pe
r
pe
r
s
on.
T
he
va
lue
of
C
r
onba
c
h's
a
lpha
is
0.
7454,
s
o
i
t
is
c
ons
ider
e
d
that
the
int
e
r
na
l
c
ons
is
tenc
y
of
the
inf
or
mation
c
oll
e
c
ti
on
ins
tr
ument
de
s
igned
is
a
de
qua
te.
T
he
s
tanda
r
d
de
viation
indi
c
a
tes
that
the
da
ta
a
r
e
g
r
ou
pe
d
a
nd
c
los
e
to
the
mea
n
va
lue.
It
is
obs
e
r
ve
d
that
75
%
o
f
the
que
s
ti
ons
tend
to
ha
ve
a
low
c
onc
e
ntr
a
ti
on
o
f
va
lues
,
that
is
,
they
ha
ve
a
platykur
ti
c
dis
tr
ibut
ion
be
c
a
us
e
their
kur
tos
is
is
ne
ga
ti
ve
a
nd
pr
e
s
e
nts
a
typi
c
a
l
da
t
a
,
while
que
s
ti
ons
with
pos
it
ive
kur
tos
is
ha
ve
a
high
c
onc
e
ntr
a
ti
on
of
va
lues
da
ta
s
howing
a
leptokur
ti
c
dis
t
r
ibut
ion.
Ac
c
or
ding
to
the
c
oe
f
f
icie
nt
o
f
va
r
iat
ion
in
the
q
ue
s
ti
ons
,
the
dis
pe
r
s
ion
a
nd
va
r
iabili
ty
is
low
be
c
a
us
e
the
va
lue
is
les
s
than
25%
,
the
que
s
ti
on
that
pr
e
s
e
nts
the
gr
e
a
tes
t
va
r
iabili
ty
is
que
s
ti
on
4
a
nd
the
one
with
the
lea
s
t
va
r
iabili
ty
is
que
s
ti
on
2,
s
o
it
s
de
duc
e
d
that
t
he
r
e
is
a
r
e
lative
homogene
it
y
be
twe
e
n
the
va
lues
obtaine
d
a
nd
the
a
ve
r
a
ge
is
r
e
pr
e
s
e
ntative.
B
a
s
e
d
on
thes
e
r
e
s
ult
s
,
it
wa
s
de
c
ided
to
a
djus
t
t
he
e
va
luation
ins
tr
ument
in
a
c
c
or
da
nc
e
with
the
pr
e
li
mi
na
r
y
da
ta
,
i
mpr
oving
it
s
wor
ding
,
includi
ng
in
the
inf
o
r
med
c
ons
e
nt,
to
be
c
ompl
e
ted
b
e
f
or
e
it
s
a
ppli
c
a
ti
on,
the
pos
s
ibi
li
ty
o
f
withd
r
a
wing
f
r
om
th
e
tes
t
a
t
a
ny
ti
me.
T
he
labor
a
tor
y
a
r
e
a
tes
t
wa
s
c
on
dit
ioned
in
s
uc
h
a
wa
y
that
ther
e
wa
s
no
int
e
r
f
e
r
e
nc
e
in
t
he
s
ound,
given
the
c
ontr
oll
e
d
tr
a
ini
ng
c
o
ndit
ion
s
of
the
c
onvolut
ional
ne
twor
k
.
On
the
other
ha
nd,
the
r
e
s
pons
e
s
f
r
om
the
pil
o
t
tes
t
ha
d
r
a
nge
s
that
did
not
e
xc
e
e
d
3,
s
o
a
s
c
a
le
of
1
to
7
is
not
us
e
f
ul,
whic
h
is
why
the
L
iker
t
s
c
a
le
is
r
e
duc
e
d
f
r
om
1
to
5.
U
s
a
bi
li
ty
E
f
f
e
c
t
i
ve
ne
s
s
A
s
s
e
r
ti
ve
ne
s
s
E
f
f
e
c
t
i
ve
ne
s
s
U
ti
li
ty
E
f
f
ic
ie
nc
y
C
la
r
it
y
E
a
s
e
L
e
a
r
ni
ng
U
s
e
r
s
a
ti
s
f
a
c
ti
on
N
a
tu
r
a
ln
e
s
s
S
a
ti
s
f
a
c
ti
on
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
2
,
Apr
il
2025
:
91
7
-
92
4
922
T
a
ble
2.
L
is
t
of
que
s
ti
ons
,
va
r
iable
s
a
n
d
us
a
bil
it
y
c
r
it
e
r
ia
U
s
a
bi
li
ty
c
r
it
e
r
ia
# a
s
k
V
a
r
ia
bl
e
E
f
f
ic
ie
nc
y
1
C
la
r
it
y
E
f
f
e
c
ti
ve
ne
s
s
2
A
s
s
e
r
ti
ve
ne
s
s
U
s
e
r
s
a
ti
s
f
a
c
ti
on
3
N
a
tu
r
a
ln
e
s
s
E
f
f
e
c
ti
ve
ne
s
s
4
E
f
f
e
c
ti
ve
ne
s
s
U
s
e
r
s
a
ti
s
f
a
c
ti
on
5
S
a
ti
s
f
a
c
ti
on
E
f
f
ic
ie
nc
y
6
E
a
s
e
E
f
f
e
c
ti
ve
ne
s
s
7
U
ti
li
ty
E
f
f
ic
ie
nc
y
8
L
e
a
r
ni
ng
D
oe
s
not
a
ppl
y
9
V
oi
c
e
T
a
ble
3.
De
s
c
r
ipt
ive
s
tatis
ti
c
s
of
the
pil
o
t
tes
t
P1
P2
P3
P4
P5
P6
P7
P8
H
a
lf
6.3
6.7
6.2
5.3
5.9
6
6.5
5.9
F
a
s
hi
on
6
7
6
5
6
7
7
6
V
a
r
ia
nc
e
0.41
0.21
0.36
0.81
0.49
1
0.45
0.49
D
e
vi
a
ti
on
0.6749
0.4830
0.6325
0.9487
0.7379
1.0541
0.7071
0.7379
M
in
im
um
5
6
5
4
5
4
5
5
M
a
xi
mum
7
7
7
7
7
7
7
7
R
a
nge
2
1
2
3
2
3
2
2
K
ur
to
s
is
-
0.2830
-
1.2245
0.1786
-
0.3469
-
0.7336
-
0.4500
0.57143
-
0.7336
C
oe
f
f
ic
ie
nt
of
va
r
ia
ti
on
10.71%
7.21%
10.20%
17.90%
12.51%
17.57%
10.88%
12.51%
Af
ter
a
pplyi
ng
the
va
li
da
ted
e
va
luation
ins
tr
ument
with
the
a
djus
tm
e
nts
de
s
c
r
ibed
a
bove
,
the
a
ve
r
a
ge
tes
t
ti
me
r
e
duc
e
d
to
5.
7742
mi
nu
tes
pe
r
p
e
r
s
on.
T
a
ble
4
pr
e
s
e
nts
the
de
s
c
r
ipt
ive
s
tatis
ti
c
s
of
the
f
ir
s
t
e
ight
que
s
ti
ons
of
the
f
inal
s
ur
ve
y.
T
he
ove
r
a
ll
a
v
e
r
a
ge
s
c
or
e
obtaine
d
wa
s
4
.
226
out
of
5
,
s
o
the
d
e
gr
e
e
of
ge
ne
r
a
l
s
a
ti
s
f
a
c
ti
on
with
r
e
s
pe
c
t
to
the
us
a
bil
it
y
of
the
pr
opos
e
d
c
onve
r
s
a
ti
on
s
c
r
ipt
be
twe
e
n
a
us
e
r
a
nd
a
r
obot
f
or
a
s
s
is
tanc
e
,
us
ing
pr
e
-
r
e
c
or
de
d
r
e
s
p
ons
e
s
,
wa
s
84.
52%
.
R
e
ga
r
ding
the
us
a
bil
it
y
c
r
it
e
r
ia,
it
is
obs
e
r
ve
d
that
the
de
gr
e
e
of
s
a
ti
s
f
a
c
ti
on
wa
s
82
.
5
8%
,
86.
02%
,
a
nd
84
.
09%
in
e
f
f
e
c
ti
ve
ne
s
s
,
e
f
f
icie
nc
y
a
nd
us
e
r
s
a
ti
s
f
a
c
ti
on,
r
e
s
pe
c
ti
ve
ly.
Only
one
of
the
e
ig
ht
que
s
ti
ons
obtaine
d
a
s
a
ti
s
f
a
c
ti
on
lev
e
l
of
les
s
th
a
n
80%
,
whic
h
wa
s
que
s
ti
on
4,
whic
h
c
or
r
e
s
ponds
to
the
e
f
f
e
c
ti
ve
ne
s
s
va
r
iable
.
T
his
va
r
iable
obtaine
d
the
lowe
s
t
s
c
or
e
be
c
a
us
e
the
tes
t
doe
s
not
s
how
the
r
obot
e
xe
c
uti
ng
the
tas
ks
.
I
t
is
obs
e
r
ve
d
that
the
a
ve
r
a
ge
va
l
ue
s
of
a
ll
the
que
s
ti
ons
a
r
e
c
los
e
to
e
a
c
h
other
,
e
videnc
in
g
that
the
da
ta
a
r
e
c
los
e
to
s
a
id
mea
s
ur
e
given
the
low
de
viation
a
nd
ther
e
f
o
r
e
a
r
e
r
e
latively
c
onc
e
ntr
a
ted.
T
a
ble
4.
De
s
c
r
ipt
ive
s
tatis
ti
c
s
of
the
f
inal
s
ur
ve
y
P1
P2
P3
P4
P5
P6
P7
P8
H
a
lf
4.4839
4.4516
4.4516
3.7419
4.0645
4.3226
4.1935
4.0968
F
a
s
hi
on
5
5
5
4
4
5
5
4
V
a
r
ia
nc
e
0.3143
0.4412
0.3122
0.7721
0.6410
0.6701
0.6077
0.6035
D
e
vi
a
ti
on
0.5699
0.6752
0.5680
0.8932
0.8139
0.8321
0.7924
0.7897
M
in
im
um
3
3
3
2
3
2
3
3
M
a
xi
mum
5
5
5
5
5
5
5
5
R
a
nge
2
2
2
3
2
3
2
2
K
ur
to
s
is
-
0.7180
-
0.3246
-
0.8124
-
0.4375
-
1.4704
0.4808
-
1.2887
-
1.3465
C
oe
f
f
ic
ie
nt
of
va
r
ia
ti
on
12.71%
15.17%
12.76%
23.87%
20.02%
19.25%
18.90%
19.28%
W
it
h
the
a
djus
tm
e
nt
of
the
ins
tr
ument,
a
platykur
ti
c
f
lat
dis
tr
ibut
ion
is
ob
taine
d
be
c
a
us
e
in
87.
5%
o
f
the
que
s
ti
ons
the
kur
tos
is
is
ne
ga
ti
ve
,
ha
ving
va
l
ue
s
that
a
r
e
f
a
r
f
r
om
the
mea
n
a
nd
do
not
a
ll
ow
a
nor
mal
dis
tr
ibut
ion.
How
e
ve
r
,
the
c
oe
f
f
icie
nt
of
va
r
iat
io
n
indi
c
a
tes
that
da
ta
a
r
e
r
e
latively
gr
oupe
d
be
c
a
us
e
the
va
lues
r
e
tur
n
e
d
a
r
e
les
s
than
24%
,
the
que
s
ti
on
1
ha
s
the
gr
e
a
tes
t
gr
ouping
of
the
da
ta
a
nd
the
que
s
ti
on
4
ha
s
the
lea
s
t
gr
ouping
.
T
he
va
lue
of
C
r
onba
c
h's
Alpha
of
the
a
djus
ted
s
ur
ve
y
c
o
r
r
e
s
ponds
to
0
.
8031,
th
is
mea
ns
ther
e
is
good
int
e
r
na
l
c
ons
is
tenc
y
in
the
inf
or
mati
on
c
oll
e
c
ti
on
ins
tr
ument
de
s
igned,
s
o
it
is
r
e
li
a
ble
a
nd
the
c
onc
lus
ions
obtaine
d
f
r
om
i
ts
r
e
s
ult
s
a
r
e
va
li
d
.
A
ddit
ionally,
wi
th
r
e
s
pe
c
t
to
que
s
ti
on
9,
whic
h
r
e
f
e
r
s
to
the
pr
e
f
e
r
e
nc
e
of
the
type
o
f
voice
f
or
the
r
obot
,
it
i
s
obs
e
r
ve
d
that
none
of
the
r
e
s
po
nde
nts
c
hos
e
the
opti
on
male
,
a
doles
c
e
nt
or
c
hil
d,
10%
of
the
pa
r
ti
c
ipants
pr
e
f
e
r
the
voice
of
a
n
a
dult
man
,
29
%
of
a
n
a
dul
t
woma
n
a
nd
61%
of
a
n
a
doles
c
e
nt
woma
n,
s
o
thi
s
las
t
r
e
s
ult
will
be
us
e
d
f
or
the
f
inal
r
e
c
or
dings
of
the
dial
ogue
s
of
the
pr
opos
e
d
s
c
e
ne
s
.
Although
a
c
c
or
ding
to
[
25]
,
ther
e
is
no
unif
ied
model
of
human
c
ognit
ion
a
nd
de
c
is
ion
making,
c
ognit
ive
s
c
ienc
e
c
a
n
be
ve
r
y
us
e
f
ul
f
or
the
de
s
ign
of
a
utonom
ous
s
ys
tems
that
int
e
r
a
c
t
with
hum
a
ns
with
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
I
nter
ac
ti
v
e
c
omm
unication
human
-
r
obot
int
e
r
face
for
r
e
duc
e
d
mobili
ty
…
(
R
obins
on
J
ime
ne
z
-
M
or
e
no
)
923
tool
s
s
uc
h
a
s
:
blac
k
box
methods
a
nd
theor
y
-
ba
s
e
d
methods
,
r
e
s
our
c
e
-
r
a
ti
ona
l
de
c
is
i
on
making,
c
oope
r
a
ti
ve
c
ognit
ion
a
nd
r
a
ti
ona
l
s
pe
e
c
h
a
c
t
(
R
S
A)
theor
y,
whic
h
c
ould
be
e
xplor
e
d
in
the
f
utu
r
e
.
As
f
u
tur
e
r
e
s
e
a
r
c
h,
it
is
s
ugge
s
ted
to
c
a
r
r
y
out
va
li
da
ti
ons
with
other
ne
twor
k
a
r
c
hit
e
c
tur
e
models
,
c
ons
ider
ing
other
f
e
a
tur
e
e
xtr
a
c
ti
on
tec
hniqu
e
s
s
uc
h
a
s
a
F
our
ier
or
W
a
ve
let
s
pe
c
tr
ogr
a
m
that
c
a
n
give
r
e
s
ult
s
in
im
pr
ov
i
ng
voice
identif
ica
ti
on.
I
t
is
a
ls
o
pr
opos
e
d
to
a
pply
the
f
inal
ins
tr
ument
to
a
lar
ge
s
a
mpl
e
that
a
ll
ows
the
a
ppli
c
a
ti
on
of
f
a
c
tor
a
na
lys
is
models
a
nd
the
c
on
s
ider
a
ti
on
of
ne
w
e
va
l
ua
ti
on
va
r
iable
s
a
nd
wor
k
on
human
-
mac
hine
c
omm
unica
ti
on
that
include
s
the
int
e
r
pr
e
tation
of
ge
s
tur
e
s
by
the
r
obot
to
a
s
s
is
t
pe
ople
with
r
e
duc
e
d
mobi
li
ty.
As
we
ll
c
ompl
e
menting
the
us
a
bil
it
y
t
e
s
t
with
a
s
a
ti
s
f
a
c
ti
on
a
na
lys
is
,
whic
h
tur
ns
out
to
be
a
typi
c
a
ll
y
s
ubjec
ti
ve
tas
k.
4.
CONC
L
USI
ON
I
t
wa
s
pos
s
ibl
e
to
e
f
f
icie
ntl
y
r
e
c
ognize
,
unde
r
a
c
ontr
oll
e
d
e
nvir
onment
,
the
g
r
oup
o
f
wor
ds
that
a
ll
ow
the
r
obot
to
be
c
omm
a
nde
d,
us
ing
a
pe
r
s
ona
li
z
e
d
a
r
c
hit
e
c
tur
e
of
a
c
onvolut
ional
ne
ur
a
l
ne
twor
k
with
s
ix
c
onvolut
ion
la
ye
r
s
,
whic
h
gives
the
pos
s
ibi
li
ty
of
be
ing
e
mbedde
d
in
a
lar
ge
number
o
f
mi
c
r
oc
o
ntr
oll
e
d
s
ys
tems
f
or
de
vice
s
por
table
,
thus
pr
ojec
ti
ng
f
utu
r
e
wor
k
in
thi
s
a
s
pe
c
t.
A
dialogue
model
wa
s
e
s
tablis
he
d
be
twe
e
n
the
us
e
r
a
nd
the
r
obot
with
c
ondit
ions
f
or
a
c
ti
va
ti
n
g
a
nd
de
a
c
ti
va
ti
ng
the
r
obot,
ba
s
e
d
on
thr
e
e
s
c
e
ne
s
:
the
f
ir
s
t
s
e
e
ks
f
or
the
r
obot
to
e
xplor
e
a
n
d
r
e
c
ognize
the
r
e
s
idential
e
nvir
onment
,
the
s
e
c
ond
a
ll
ows
the
s
pa
c
e
s
to
be
r
e
na
med,
a
nd
the
thi
r
d
is
int
e
nde
d
to
e
xe
c
uti
ng
a
c
ti
on
c
omm
a
nds
to
move
e
leme
nts
.
I
t
wa
s
e
vident
that
thi
s
dis
tr
ibut
ion
a
ll
ows
the
tot
a
l
int
e
r
a
c
ti
on
of
the
r
obot
to
be
c
ove
r
e
d
,
f
r
om
it
s
ini
ti
a
l
us
e
to
ge
ne
r
a
l
a
s
s
is
tan
c
e
.
T
he
tes
ts
s
how
the
da
ta
c
oll
e
c
ti
on
ins
tr
ument
is
r
e
li
a
ble
a
nd
ther
e
f
o
r
e
the
c
on
c
lus
ions
obtaine
d
f
r
om
the
a
na
lys
is
of
t
he
inf
o
r
mation
c
oll
e
c
ted
a
r
e
a
ls
o
r
e
li
a
ble.
A
ge
ne
r
a
l
s
a
ti
s
f
a
c
ti
on
leve
l
of
84.
52%
wa
s
a
c
hieve
d,
a
nd
g
r
e
a
ter
than
82
%
f
or
the
us
a
bil
it
y
c
r
it
e
r
ia
(
e
f
f
e
c
ti
ve
ne
s
s
,
e
f
f
icie
nc
y
,
a
nd
us
e
r
s
a
ti
s
f
a
c
ti
on)
.
C
onc
ludi
ng
that
the
pr
opos
e
d
us
a
bi
li
ty
mea
s
ur
e
ment
model
c
a
n
be
r
e
pli
c
a
ted
unde
r
s
im
il
a
r
a
ppli
c
a
ti
on
c
ondit
ions
.
B
a
s
e
d
on
the
voice
pr
e
f
e
r
e
nc
e
of
the
pa
r
ti
c
ipants
,
the
us
e
of
a
n
a
doles
c
e
nt
woma
n's
voice
is
de
f
ined
f
o
r
the
f
inal
r
e
c
or
dings
o
f
the
dia
logues
us
e
d
in
the
th
r
e
e
p
r
opos
e
d
s
c
e
ne
s
.
W
hich
s
hows
a
de
gr
e
e
of
f
a
mi
li
a
r
it
y
a
nd
f
r
iendlines
s
of
tr
e
a
tm
e
nt
s
ought
by
us
e
r
s
.
AC
KNOWL
E
DGE
M
E
NT
S
P
r
oduc
t
de
r
ived
f
r
om
the
r
e
s
e
a
r
c
h
p
r
ojec
t
ti
tl
e
d
“
De
s
ign
of
a
human
-
r
obot
int
e
r
a
c
ti
on
model
us
ing
de
e
p
lea
r
ning
a
lgor
it
hms
”
I
NV
-
I
NG
-
3971
f
inanc
e
d
by
the
vice
-
r
e
c
tor
f
or
r
e
s
e
a
r
c
h
of
the
Unive
r
s
idad
M
il
it
a
r
Nue
va
Gr
a
na
da
,
ye
a
r
2024
.
RE
F
E
RE
NC
E
S
[
1]
M
.
M
.
A
.
A
lm
e
id
a
,
“
R
obot
s
,
in
te
li
ge
n
c
ia
a
r
ti
f
ic
ia
l
y
r
e
a
li
da
d
vi
r
tu
a
l:
una
a
pr
oxi
ma
c
ió
n
e
n
e
l
s
e
c
to
r
de
l
tu
r
is
mo,”
C
uad
e
r
no
s
de
T
ur
is
m
o
, no.
44, pp. 13
–
26, Nov. 2019, doi:
10.6018/t
ur
is
mo.44.404711.
[
2]
I
. A
. S
.
G
a
r
c
í
a
,
“
L
os
r
obot
s
y
l
a
in
te
li
ge
nc
i
a
a
r
ti
f
ic
i
a
l.
nu
e
vo
s
r
e
t
os
d
e
l
p
e
r
io
di
s
m
o,”
D
o
x
a
C
om
uni
c
ac
i
ón. R
R
e
v
i
s
ta
I
nt
e
r
di
s
c
i
pl
i
na
r
D
e
E
s
tu
di
os
D
e
C
om
un
ic
a
c
ió
n
Y
C
ie
n
c
ia
s
S
oc
i
al
e
s
,
no.
27,
pp.
295
–
3
15,
D
e
c
. 2
018,
doi
:
10.
3192
1/
do
xa
c
om.n
27a
15.
[
3]
M
.
Y
.
P
i
n
e
d
a
-
de
-
A
l
c
á
z
a
r
,
“
I
nt
e
l
i
g
e
n
c
i
a
a
r
ti
f
i
c
i
a
l
y
mo
d
e
lo
s
d
e
c
om
u
n
i
c
a
c
i
ó
n
,
”
R
a
z
ó
n
y
P
al
a
b
r
a
,
v
ol
.
21
,
no
.
4
\
_
9
9,
p
p.
3
32
–
3
46
,
20
1
7
.
[
4]
M
.
A
.
O
.
M
is
s
e
s
,
H
.
S
.
O
c
hoa
,
a
nd
A
.
S
.
L
ópe
z
,
“
L
udi
bot
:
I
nt
e
r
f
a
z
huma
no
-
r
obot
móv
il
pa
r
a
e
l
a
p
r
e
ndi
z
a
je
lú
di
c
o
de
id
io
m
a
s
,”
I
nge
ni
e
r
ía
I
nv
e
s
ti
gac
ió
n y
T
e
c
nol
ogí
a
, vol
. 22, no. 3, pp. 1
–
10,
J
ul
. 2021, doi:
10.22201/f
i.
25940732e
.2021.22.3.021.
[
5]
M
.
E
.
R
.
C
a
s
ti
ll
o,
A
.
J
.
F
.
M
ur
il
lo
,
a
nd
C
.
R
.
C
r
.
M
e
ndo
z
a
,
“
C
ons
id
e
r
a
c
io
ne
s
e
n
e
l
di
s
e
ño
de
r
obot
s
p
a
r
a
la
a
te
nc
ió
n
mé
di
c
a
e
n
e
l
mundo pos
t
C
O
V
I
D
-
19,”
C
ul
tu
r
a C
ie
nt
íf
ic
a y
T
e
c
nol
ógi
c
a
, vol
. 19, no. 1, 2022, doi
:
10.20983/culc
yt
.2022.1.3.3.
[
6]
N
. B
. S
a
nt
os
, R
. S
. B
a
va
r
e
s
c
o, J
. E
. R
. T
a
va
r
e
s
, G
.
D.
O
. R
a
mo
s
, a
nd J
. L
. V
B
a
r
bos
a
, “
A
s
ys
te
ma
ti
c
ma
ppi
ng s
tu
dy of
r
obot
ic
s
i
n
huma
n c
a
r
e
,”
R
obot
ic
s
and A
ut
onomous
Sy
s
t
e
m
s
, vol
. 144, Oc
t.
2021, doi:
10.1016/j
.r
obot
.2021.103833.
[
7]
W
H
O
,
“
D
i
s
a
bi
li
ty
,”
W
or
ld
H
e
al
th
O
r
gani
z
at
io
n
,
2023.
A
c
c
e
s
s
e
d:
S
e
p
.
07,
2023.
[
O
nl
in
e
]
.
A
va
il
a
bl
e
:
ht
tp
s
:/
/ww
w
.
w
ho.i
nt
/n
e
w
s
-
r
oom/
f
a
c
t
-
s
he
e
ts
/d
e
ta
il
/d
is
a
bi
li
ty
-
a
nd
-
he
a
lt
h
[
8]
S
.
S
c
hne
id
e
r
,
W
.
C
la
s
,
a
nd
D
.
B
r
s
c
ic
,
“
H
uma
n
-
r
obot
in
te
r
a
c
ti
on
in
publ
ic
s
pa
c
e
s
,”
in
2022
17t
h
A
C
M
/I
E
E
E
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on Hum
an
-
R
obot
I
nt
e
r
ac
ti
on (
H
R
I
)
, M
a
r
. 2022, pp.
1287
–
1289, doi:
10.1109/HR
I
53351.2022.9889525.
[
9]
J
.
P
a
r
k
e
t
al
.
,
“
U
s
e
r
pe
r
c
e
pt
io
n
on
pe
r
s
ona
li
z
e
d
e
xpl
a
na
ti
on
by
s
c
ie
nc
e
mus
e
um
doc
e
nt
r
obot
,”
in
2022
17t
h
A
C
M
/I
E
E
E
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on Hum
an
-
R
obot
I
nt
e
r
ac
ti
on (
H
R
I
)
,
M
a
r
. 2022, pp. 973
–
975, doi:
10.1109/HR
I
53351.2022.9889654.
[
10]
M
.
M
ouj
a
hi
d,
B
.
W
il
s
on,
H
.
H
a
s
ti
e
,
a
nd
O
.
L
e
mon,
“
D
e
mons
tr
a
ti
on
of
a
r
obot
r
e
c
e
pt
io
ni
s
t
w
it
h
mul
ti
-
pa
r
ty
s
it
ua
te
d
in
te
r
a
c
ti
o
n,”
in
2
022
17t
h
A
C
M
/I
E
E
E
I
n
te
r
nat
io
nal
C
onf
e
r
e
nc
e
on
H
um
an
-
R
obot
I
nt
e
r
ac
ti
on
(
H
R
I
)
,
M
a
r
.
2022,
pp.
1202
–
1203,
doi
:
10.1109/HR
I
53351.2022.9889314.
[
11]
W
. L
a
w
, K
. L
i,
K
. F
a
n, T
. M
o, a
nd C
. P
oon, “
F
r
ie
ndl
y e
le
va
to
r
c
o
-
r
id
e
r
:
A
n H
R
I
a
ppr
oa
c
h f
o
r
r
obot
-
e
le
va
to
r
i
nt
e
r
a
c
ti
on,”
i
n
2
022
17t
h
A
C
M
/I
E
E
E
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
H
um
an
-
R
obot
I
nt
e
r
ac
ti
on
(
H
R
I
)
,
M
a
r
.
2022,
pp.
865
–
869,
doi
:
10.1109/HR
I
53351.2022.9889614.
[
12]
A
.
Y
a
ma
z
a
ki
,
A
.
K
r
umm
he
ue
r
,
a
nd
M
.
I
ma
i,
“
I
nt
e
r
di
s
c
ip
li
na
r
y
e
xpl
or
a
ti
ons
of
pr
oc
e
s
s
e
s
of
mut
ua
l
unde
r
s
ta
ndi
ng
in
in
te
r
a
c
ti
on
w
it
h
a
s
s
is
ti
ve
s
hopping
r
obot
s
,”
in
2022
17t
h
A
C
M
/I
E
E
E
I
nt
e
r
nat
io
nal
C
onf
e
r
e
n
c
e
on
H
um
an
-
R
obot
I
nt
e
r
ac
ti
on
(
H
R
I
)
,
M
a
r
.
2022, pp. 1293
–
1295, doi:
10.1109/HR
I
53351.2022.9889352.
[
13]
L
.
T
a
ka
ya
ma
,
“
P
ut
ti
ng
huma
n
-
r
obot
in
te
r
a
c
ti
on
r
e
s
e
a
r
c
h
in
to
de
s
ig
n
pr
a
c
ti
c
e
,”
in
2022
17t
h
A
C
M
/I
E
E
E
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on H
um
an
-
R
obot
I
nt
e
r
ac
ti
on (
H
R
I
)
, M
a
r
. 2022, p. 1, doi:
10.11
09/
H
R
I
53351.2022.9889659.
[
14]
K
.
K
a
w
a
i,
R
.
R
z
e
pka
,
a
nd T
.
N
e
mot
o,
“
U
s
in
g
c
onvolut
io
na
l
ne
ur
a
l
ne
twor
k
f
or
im
pr
ovi
ng
in
f
e
r
e
nc
e
of
in
te
r
r
oga
ti
ve
s
e
nt
e
nc
e
s
in
a
di
a
lo
gue
s
y
s
te
m,”
in
2022
10t
h
I
nt
e
r
nat
io
nal
C
onf
e
r
e
n
c
e
o
n
A
ff
e
c
ti
v
e
C
om
put
in
g
and
I
nt
e
ll
ig
e
nt
I
nt
e
r
ac
ti
on
W
or
k
s
hops
and
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
2
,
Apr
il
2025
:
91
7
-
92
4
924
D
e
m
os
(
A
C
I
I
W
)
, O
c
t.
2022, pp. 1
–
4, doi:
10.1109/AC
I
I
W
5723
1.2022.10086010.
[
15]
S
. L
iu
,
S
. L
iu
, a
n
d W
. X
u, “
G
a
te
d a
tt
e
nt
iv
e
c
onvolut
io
na
l
ne
tw
or
k di
a
lo
gue
s
ta
te
t
r
a
c
ke
r
,”
i
n
2020 I
E
E
E
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on A
c
ous
ti
c
s
, Spe
e
c
h and Signal
P
r
oc
e
s
s
in
g (
I
C
A
SSP
)
, 2020, p
p. 6174
–
6178, doi:
10.1109/I
C
A
S
S
P
40776.2020.9054225.
[
16]
H
.
S
hi
,
T
.
U
s
hi
o,
M
.
E
ndo,
K
.
Y
a
ma
ga
mi
,
a
nd
N
.
H
or
ii
,
“
A
mul
ti
c
ha
nne
l
c
onvolut
io
na
l
ne
ur
a
l
ne
twor
k
f
or
c
r
os
s
-
la
ngua
ge
di
a
lo
g
s
ta
te
tr
a
c
ki
ng,”
in
2016
I
E
E
E
Spok
e
n
L
anguage
T
e
c
hnol
ogy
W
or
k
s
hop
(
SL
T
)
,
D
e
c
.
2016,
pp.
559
–
564,
doi
:
10.1109/S
L
T
.2016.7846318.
[
17]
H.
-
A
.
R
us
a
n
a
nd
B
.
M
oc
a
nu
,
“
H
um
a
n
-
c
omput
e
r
in
te
r
a
c
ti
on
th
r
ough
voi
c
e
c
omm
a
nds
r
e
c
ogni
ti
on,”
in
2022
I
nt
e
r
nat
io
nal
Sy
m
pos
iu
m
on E
le
c
tr
oni
c
s
and T
e
le
c
om
m
uni
c
at
io
n
s
(
I
SE
T
C
)
,
N
ov. 2022, pp. 1
–
4, doi:
10.1109/I
S
E
T
C
56213.2022.10010253.
[
18]
M
.
S
a
lo
umi
e
t
al
.
,
“
S
pe
e
c
h
e
mot
io
n
r
e
c
ogni
ti
on
us
in
g
o
ne
-
di
me
ns
io
na
l
c
onvolut
io
na
l
ne
ur
a
l
ne
twor
ks
,”
in
2023
46t
h
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
T
e
le
c
om
m
uni
c
at
io
n
s
and
Si
gnal
P
r
oc
e
s
s
in
g
(
T
SP
)
,
J
ul
.
2023,
pp.
212
–
216,
doi
:
10.1109/T
S
P
59544.2023.10197766.
[
19]
M
.
L
.
S
c
hr
um,
M
.
J
ohns
on,
M
.
G
huy,
a
nd
M
.
C
.
G
ombol
a
y,
“
F
our
ye
a
r
s
in
r
e
vi
e
w
,”
in
C
om
pani
on
of
th
e
2020
A
C
M
/I
E
E
E
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on Hum
an
-
R
obot
I
nt
e
r
ac
ti
on
, M
a
r
. 2
020, pp. 43
–
52, doi:
10.1145/3371382.3
380739.
[
20]
R
.
J
im
é
ne
z
-
M
or
e
no
a
nd
R
.
A
.
C
a
s
ti
ll
o,
“
D
e
e
p
le
a
r
ni
ng
s
pe
e
c
h
r
e
c
ogni
ti
on
f
or
r
e
s
id
e
nt
ia
l
a
s
s
is
ta
nt
r
obot
,”
I
A
E
S
I
nt
e
r
nat
i
onal
J
our
nal
of
A
r
ti
fi
c
ia
l
I
nt
e
ll
ig
e
nc
e
, vol
. 12, no. 2, pp. 585
–
592, 2023, doi:
10.11591/i
ja
i.
v12.i2.pp585
-
592.
[
21]
J
.
J
.
H
e
r
r
e
r
a
,
R
.
J
im
e
ne
z
-
M
or
e
no,
a
nd
J
.
E
.
M
.
B
a
que
r
o,
“
V
ir
t
ua
l
e
nv
ir
onme
nt
f
or
a
s
s
is
ta
nt
mobi
le
r
obot
,”
I
nt
e
r
nat
io
nal
J
our
nal
of
E
le
c
tr
ic
al
and C
om
put
e
r
E
ngi
ne
e
r
in
g
, vol
. 13, no. 6, pp. 6174
–
6184, De
c
. 2023, doi:
10.11591/i
je
c
e
.v13i6.pp6174
-
6184.
[
22]
J
.
M
.
F
e
r
r
e
ir
a
e
t
al
.
,
“
I
mpa
c
t
of
us
a
bi
li
ty
me
c
ha
ni
s
ms
:
A
n
e
xpe
r
im
e
nt
on
e
f
f
ic
ie
nc
y,
e
f
f
e
c
ti
ve
ne
s
s
a
nd
us
e
r
s
a
ti
s
f
a
c
ti
on,”
I
nf
or
m
at
io
n and Soft
w
ar
e
T
e
c
hnol
ogy
, vol
. 117, J
a
n. 2020, doi:
10.1016/j
.i
nf
s
of
.2019.106195.
[
23]
I
S
O
,
“
E
r
gonomi
c
s
of
huma
n
-
s
ys
te
m
in
te
r
a
c
ti
on
-
P
a
r
t
11:
U
s
a
bi
li
ty
:
D
e
f
in
it
io
ns
a
nd
c
onc
e
pt
s
,”
I
nt
e
r
nat
io
n
al
St
andar
di
z
a
ti
on
O
r
gani
z
at
io
n
,
2018
.
[
O
nl
in
e
]
. A
va
il
a
bl
e
:
ht
tp
s
:/
/ww
w
.i
s
o.or
g/
o
bp/
ui
/#
is
o:
s
td
:i
s
o:
9241:
-
11:
e
d
-
2:
v1:
e
n
[
24]
D
.
A
.
A
lb
or
noz
,
S
.
A
.
M
onc
a
yo,
S
.
R
ua
no
-
H
oyos
,
G
.
E
.
C
ha
n
c
hí
-
G
ol
ond
r
in
o,
a
nd
K
.
M
á
r
c
e
le
s
-
V
il
la
lb
a
,
“
S
is
te
ma
s
of
twa
r
e
pa
r
a
la
e
je
c
uc
ió
n
de
pr
ue
ba
s
de
u
s
a
bi
li
da
d
ba
jo
e
l
e
nf
oque
d
e
mous
e
tr
a
c
ki
ng,”
T
e
c
noL
ógi
c
a
s
,
vol
.
22,
pp.
19
–
31,
D
e
c
.
2019,
doi
:
10.22430/22565337
.1511.
[
25]
M
.
K
.
H
o
a
nd
T
.
L
.
G
r
if
f
it
hs
,
“
C
ogni
ti
ve
s
c
ie
nc
e
a
s
a
s
our
c
e
o
f
f
or
w
a
r
d
a
nd
in
ve
r
s
e
mode
ls
of
huma
n
de
c
is
io
ns
f
o
r
r
obot
ic
s
a
nd
c
ont
r
ol
,”
A
nnual
R
e
v
ie
w
of
C
ont
r
ol
,
R
obot
ic
s
,
and
A
ut
o
nom
ous
Sy
s
te
m
s
,
vol
.
5,
no.
1,
pp.
33
–
53,
M
a
y
2022,
doi
:
10.1146/a
nnur
e
v
-
c
ont
r
ol
-
042920
-
015547.
B
I
OG
RA
P
HI
E
S
OF
AU
T
HO
RS
R
o
bi
n
s
o
n
J
i
m
énez
M
o
ren
o
i
s
an
el
ec
t
ro
n
i
c
en
g
i
n
eer
g
rad
u
at
e
d
fro
m
U
n
i
v
er
s
i
d
ad
D
i
s
t
ri
t
al
Fran
c
i
s
c
o
J
o
s
é
d
e
Cal
d
a
s
i
n
2
0
0
2
.
H
e
recei
v
ed
a
M.
Sc.
i
n
E
n
g
i
n
eer
i
n
g
fro
m
U
n
i
v
er
s
i
d
ad
N
ac
i
o
n
al
d
e
Co
l
o
m
b
i
a
i
n
2
0
1
2
a
n
d
D
o
c
t
o
r
i
n
E
n
g
i
n
eeri
n
g
at
U
n
i
v
ers
i
d
a
d
D
i
s
t
ri
t
a
l
Fran
ci
s
co
J
o
s
é
d
e
Ca
l
d
a
s
i
n
2
0
1
8
.
H
i
s
cu
rre
n
t
w
o
rk
i
n
g
as
A
s
s
o
c
i
at
e
Pr
o
fes
s
o
r
o
f
U
n
i
v
er
s
i
d
ad
Mi
l
i
t
ar
N
u
ev
a
G
ran
ad
a
an
d
res
earch
fo
c
u
s
es
o
n
t
h
e
u
s
e
o
f
co
n
v
o
l
u
t
i
o
n
al
n
e
u
ra
l
n
et
w
o
r
k
s
f
o
r
o
b
j
ec
t
reco
g
n
i
t
i
o
n
a
n
d
i
mag
e
p
ro
ce
s
s
i
n
g
f
o
r
ro
b
o
t
i
c
a
p
p
l
i
ca
t
i
o
n
s
s
u
ch
a
s
h
u
ma
n
-
mach
i
n
e
i
n
t
eract
i
o
n
.
H
e
can
b
e
co
n
t
act
e
d
at
e
mai
l
:
r
o
b
i
n
s
o
n
.
j
i
me
n
ez@
u
n
i
mi
l
i
t
ar.
ed
u
.
co
.
A
nny
A
s
tr
i
d
Es
p
i
ti
a
C
ub
i
l
l
o
s
p
erfo
rmed
h
er
u
n
d
erg
rad
u
a
t
e
s
t
u
d
i
e
s
i
n
i
n
d
u
s
t
r
i
al
en
g
i
n
eer
i
n
g
i
n
t
h
e
U
n
i
v
er
s
i
d
ad
M
i
l
i
t
ar
N
u
e
v
a
G
ran
a
d
a
i
n
2
0
0
2
a
n
d
ma
s
t
er
i
n
i
n
d
u
s
t
ri
a
l
en
g
i
n
eer
i
n
g
fro
m
t
h
e
U
n
i
v
ers
i
d
a
d
d
e
L
o
s
A
n
d
e
s
i
n
2
0
0
6
.
Sh
e
i
s
a
n
A
s
s
o
c
i
at
e
Pro
fe
s
s
o
r
o
n
In
d
u
s
t
ri
a
l
E
n
g
i
n
eeri
n
g
Pr
o
g
ram
at
U
n
i
v
er
s
i
d
ad
Mi
l
i
t
ar
N
u
ev
a
G
ran
a
d
a,
Bo
g
o
t
á,
Co
l
o
mb
i
a.
Sh
e
can
b
e
co
n
t
ac
t
ed
a
t
e
mai
l
:
an
n
y
.
e
s
p
i
t
i
a@
u
n
i
mi
l
i
t
ar.
e
d
u
.
co
.
Es
pera
nza
R
o
dr
í
g
uez
C
a
rm
o
na
w
as
b
o
rn
i
n
Perei
ra
(Co
l
o
m
b
i
a).
Sh
e
p
erf
o
rmed
h
er
u
n
d
erg
ra
d
u
a
t
e
s
t
u
d
i
e
s
i
n
mec
h
an
i
cal
e
n
g
i
n
eer
i
n
g
(1
9
9
7
)
fro
m
t
h
e
U
n
i
v
ers
i
d
a
d
T
ec
n
o
l
ó
g
i
c
a
d
e
Perei
ra
(U
T
P)
a
n
d
m
as
t
er's
d
e
g
ree
i
n
u
n
i
v
er
s
i
t
y
t
eac
h
i
n
g
fro
m
t
h
e
U
n
i
v
er
s
i
d
ad
d
e
L
a
Sal
l
e,
Bo
g
o
t
á,
Co
l
o
mb
i
a.
Sh
e
i
s
an
A
s
s
o
c
i
at
e
Pro
fe
s
s
o
r
o
n
In
d
u
s
t
r
i
a
l
E
n
g
i
n
eeri
n
g
P
ro
g
ram
at
U
n
i
v
er
s
i
d
ad
M
i
l
i
t
ar
N
u
e
v
a
G
ran
ad
a,
B
o
g
o
t
á,
Co
l
o
m
b
i
a.
Sh
e
ca
n
b
e
c
o
n
t
ac
t
ed
a
t
emai
l
:
es
p
era
n
za.
ro
d
r
i
g
u
ez@
u
n
i
m
i
l
i
t
ar.
ed
u
.
co
.
Evaluation Warning : The document was created with Spire.PDF for Python.