T
E
L
KO
M
NIK
A
, V
ol
.
17
,
No.
6,
Dec
em
be
r
20
1
9,
p
p.2
99
2~
3
00
2
IS
S
N: 1
69
3
-
6
93
0
,
accr
ed
ited
F
irst
Gr
ad
e b
y K
em
en
r
istekdikti,
Decr
ee
No:
21
/E
/
K
P
T
/20
18
DOI:
10.12928/TE
LK
OM
N
IK
A
.v
1
7
i
6
.
11670
◼
29
92
Rec
ei
v
ed
Nov
e
mb
er
2
, 2
01
8
; Rev
i
s
e
d
J
un
e 9
,
20
1
9
;
A
c
c
ep
ted
J
u
l
y
2
, 2
01
9
Embe
dd
e
d
s
y
s
t
em
f
or
up
per
-
li
mb
ex
oskel
eto
n
based
on
electr
om
y
o
gra
ph
y
c
o
nt
rol
T
r
i
w
iy
ant
o
T
r
iw
iy
ant
o
*
1
,
I
P
u
t
u
A
l
it
P
aw
ana
2
, Bamb
ang
G
u
r
u
h
Ir
i
ant
o
3
,
T
r
i Bow
o
Ind
r
ato
4
, I
Dew
a Ged
e Ha
r
i W
is
ana
5
1
,3
,4
,5
Dep
a
rtm
e
n
t
o
f
El
e
c
tr
o
m
e
d
i
c
a
l
En
g
i
n
e
e
ri
n
g
,
Po
l
te
k
k
e
s
K
e
m
e
n
k
e
s
Su
ra
b
a
y
a
,
10
P
u
c
a
n
g
J
a
j
a
r T
i
m
u
r
,
S
u
r
a
b
a
y
a
,
J
a
w
a
T
i
m
u
r
,
In
d
o
n
e
s
i
a
2
Ph
y
s
i
c
a
l
M
e
d
i
c
i
n
e
a
n
d
Re
h
a
b
i
l
i
ta
ti
o
n
De
p
a
r
tm
e
n
t
Sc
h
o
o
l
o
f
M
e
d
i
c
i
n
e
,
Uni
v
e
rs
i
t
a
s
A
i
rl
a
n
g
g
a
-
Dr. S
o
e
t
o
m
o
Gen
e
ra
l
Ac
a
d
e
m
i
c
H
o
s
p
i
t
a
l
,
Su
r
a
b
a
y
a
,
J
a
w
a
T
i
m
u
r
,
In
d
o
n
e
s
i
a
*C
o
rre
s
p
o
n
d
i
n
g
a
u
th
o
r:
t
ri
w
i
y
a
n
to
1
2
3
@gm
a
i
l
.
c
o
m
Ab
strac
t
A
m
a
j
o
r
p
ro
b
l
e
m
i
n
a
n
e
x
o
s
k
e
l
e
to
n
b
a
s
e
d
o
n
e
l
e
c
tr
o
m
y
o
g
ra
p
h
y
(EM
G
)
c
o
n
tro
l
wi
th
p
a
tt
e
rn
re
c
o
g
n
i
ti
o
n
-
b
a
s
e
d
i
s
th
e
n
e
e
d
fo
r
m
o
re
t
i
m
e
to
tra
i
n
a
n
d
to
c
a
l
i
b
ra
te
th
e
s
y
s
t
e
m
i
n
o
rd
e
r
a
b
l
e
to
a
d
a
p
t
fo
r
d
i
ff
e
r
e
n
t
s
u
b
j
e
c
ts
a
n
d
v
a
r
i
a
b
l
e
.
Unf
o
rtu
n
a
t
e
l
y
,
th
e
i
m
p
l
e
m
e
n
t
a
ti
o
n
o
f
th
e
j
o
i
n
t
p
r
e
d
i
c
ti
o
n
o
n
a
n
e
m
b
e
d
d
e
d
s
y
s
te
m
fo
r
th
e
e
x
o
s
k
e
l
e
t
o
n
b
a
s
e
d
o
n
th
e
EM
G
c
o
n
tro
l
wit
h
n
o
n
-
p
a
tt
e
rn
r
e
c
o
g
n
i
t
i
o
n
-
b
a
s
e
d
i
s
v
e
ry
ra
r
e
.
Th
e
re
f
o
re
,
t
h
i
s
s
t
u
d
y
p
re
s
e
n
t
s
a
n
i
m
p
l
e
m
e
n
ta
t
i
o
n
o
f
e
l
b
o
w
-
j
o
i
n
t
a
n
g
l
e
p
r
e
d
i
c
ti
o
n
o
n
a
n
e
m
b
e
d
d
e
d
s
y
s
te
m
to
c
o
n
t
ro
l
a
n
u
p
p
e
r
l
i
m
b
e
x
o
s
k
e
l
e
to
n
b
a
s
e
d
o
n
t
h
e
EM
G
s
i
g
n
a
l
.
Th
e
a
r
c
h
i
te
c
tu
re
o
f
th
e
s
y
s
t
e
m
c
o
n
s
i
s
te
d
o
f
a
b
i
o
-
a
m
p
l
i
fi
e
r,
a
n
e
m
b
e
d
d
e
d
AR
M
ST
M
3
2
F4
2
9
m
i
c
r
o
c
o
n
tro
l
l
e
r,
a
n
d
a
n
e
x
o
s
k
e
l
e
to
n
u
n
i
t
d
r
i
v
e
n
b
y
a
s
e
rv
o
m
o
to
r.
Th
e
e
l
b
o
w
j
o
i
n
t
a
n
g
l
e
was
p
re
d
i
c
t
e
d
b
a
s
e
d
o
n
th
e
E
M
G
s
i
g
n
a
l
th
a
t
i
s
g
e
n
e
r
a
t
e
d
fro
m
b
i
c
e
p
s
.
Th
e
p
r
e
d
i
c
te
d
a
n
g
l
e
w
a
s
o
b
ta
i
n
e
d
b
y
e
x
tra
c
ti
n
g
t
h
e
EM
G
s
i
g
n
a
l
u
s
i
n
g
a
z
e
ro
-
c
ro
s
s
i
n
g
fe
a
t
u
re
a
n
d
fi
l
te
r
i
n
g
th
e
EM
G
fe
a
tu
re
u
s
i
n
g
a
B
u
tt
e
rwort
h
l
o
w
p
a
s
s
f
i
l
te
r
.
T
h
i
s
s
t
u
d
y
fo
u
n
d
t
h
a
t
th
e
ra
n
g
e
o
f
ro
o
t
m
e
a
n
s
q
u
a
r
e
er
ro
r
a
n
d
c
o
rre
l
a
ti
o
n
c
o
e
ff
i
c
i
e
n
ts
a
re
8
°
-
1
6
°
a
n
d
0
.9
4
-
0
.9
9
,
re
s
p
e
c
ti
v
e
l
y
whi
c
h
s
u
g
g
e
s
t
th
a
t
th
e
p
re
d
i
c
t
e
d
a
n
g
l
e
i
s
c
l
o
s
e
to
t
h
e
d
e
s
i
re
d
a
n
g
l
e
a
n
d
th
e
re
i
s
a
h
i
g
h
r
e
l
a
ti
o
n
s
h
i
p
b
e
twe
e
n
th
e
p
r
e
d
i
c
te
d
a
n
g
l
e
a
n
d
th
e
d
e
s
i
r
e
d
a
n
g
l
e
.
Key
w
ords
:
e
l
e
c
tr
o
m
y
o
g
r
a
p
h
y
,
e
m
b
e
d
d
e
d
s
y
s
t
e
m
,
e
x
o
s
k
e
l
e
t
o
n
,
fe
a
tu
re
e
x
t
ra
c
t
i
o
n
,
z
e
ro
c
ro
s
s
i
n
g
Copy
righ
t
©
2
0
1
9
Uni
v
e
rsi
t
a
s
Ahm
a
d
D
a
hl
a
n.
All
rig
ht
s
r
e
s
e
rve
d
.
1.
Int
r
o
d
u
ctio
n
E
l
ec
tr
om
y
og
r
a
ph
y
(
E
MG
)
s
i
gn
al
i
s
w
i
de
l
y
us
e
d
i
n
t
he
de
v
el
op
m
en
t
of
de
v
i
c
es
ba
s
ed
on
E
MG
c
o
ntro
l
s
uc
h
as
ex
os
k
el
eto
ns
,
pros
the
t
i
c
s
,
a
nd
t
el
e
op
era
ti
o
n.
A
n
ex
os
k
el
et
on
c
a
n
be
us
ed
as
an
as
s
i
s
ti
v
e
or
r
eh
a
bi
l
i
t
ati
v
e
de
v
i
c
e
to
h
el
p
h
um
an
s
to
ha
v
e
a
be
tt
er
l
i
f
e
[1
–
4]
.
T
he
d
ev
i
c
es
ba
s
ed
o
n
E
MG
c
on
tr
ol
are
c
l
as
s
i
f
i
ed
i
n
to
t
w
o
c
a
teg
o
r
i
es
:
pa
tt
ern
r
ec
og
n
i
t
i
on
an
d
no
n
-
p
att
ern
r
ec
og
ni
t
i
o
n
b
as
ed
[
1]
.
I
n
the
f
i
r
s
t
c
ate
g
or
y
,
the
c
on
tr
ol
s
i
g
na
l
i
s
d
i
s
c
r
i
m
i
na
ted
b
as
ed
on
the
p
att
erns
t
ha
t
ha
d
be
en
tr
ai
n
ed
t
o
th
e
c
l
as
s
i
f
i
er.
I
n
the
s
ec
o
nd
c
at
eg
or
y
,
th
e
c
on
tr
ol
s
i
gn
al
i
s
di
r
ec
t
l
y
ba
s
e
d
o
n
am
pl
i
t
ud
e
of
the
s
i
gn
a
l
b
y
hu
m
an
i
n
ten
t
i
o
n.
I
n
th
e
no
n
-
pa
t
t
ern
r
ec
og
n
i
t
i
on
m
eth
od
,
c
o
m
m
on
l
y
,
the
s
i
gn
al
i
s
pre
-
proc
es
s
ed
us
i
ng
ti
m
e
-
do
m
ai
n
f
ea
ture
ex
tr
ac
ti
o
n
an
d
po
s
t
-
proc
es
s
ed
us
i
n
g
a
d
i
g
i
ta
l
f
i
l
t
er
[5]
.
P
r
e
v
i
ou
s
s
tud
i
es
wer
e
m
ore
c
on
c
erned
i
n
de
v
el
op
i
ng
j
o
i
nt
an
g
l
e
p
r
ed
i
c
ti
on
b
y
ex
pl
or
i
ng
th
e
f
ea
ture
s
ex
tr
ac
ti
on
,
a
nd
m
od
i
f
y
i
ng
the
c
l
as
s
i
f
i
er
[6
–
9]
.
Cur
r
e
ntl
y
,
s
om
e
s
tud
i
es
pref
err
ed
us
i
ng
a
s
tan
d
ard
c
om
pu
ter
t
o
c
on
tr
o
l
a
n
ex
os
k
el
eto
n
b
as
ed
on
the
E
MG
s
i
g
na
l
.
Unf
ortu
na
te
l
y
,
r
es
e
arc
he
r
s
r
arel
y
i
m
pl
e
m
en
ted
a
j
oi
nt
-
an
gl
e
pred
i
c
ti
o
n
al
go
r
i
thm
,
ba
s
ed
on
p
att
er
n
r
ec
og
ni
t
i
on
an
d
n
on
-
p
att
ern
r
ec
og
n
i
ti
on
m
eth
od
, i
n t
he
em
be
dd
ed
s
y
s
t
em
.
Z
ha
ng
us
ed
a
n
eu
r
a
l
m
ac
hi
ne
f
or
an
arti
f
i
c
i
a
l
l
eg
us
i
ng
a
f
i
el
d
-
pro
gram
m
ab
l
e
ga
te
arr
a
y
(
F
P
G
A
)
f
or
f
as
ter
c
l
as
s
i
f
i
e
r
an
d
a
m
i
c
r
oc
on
tr
ol
l
er
u
ni
t
(
MC
U)
f
or
an
al
o
g
i
nte
r
f
ac
e
[10
,
11]
.
T
he
pe
r
f
or
m
an
c
e
of
the
c
l
as
s
i
f
i
er
i
s
f
as
ter
th
an
the
i
m
pl
e
m
en
tat
i
o
n
i
n
t
he
s
tan
da
r
d
c
om
pu
ter
wi
th
a
h
i
g
h
-
l
e
v
e
l
pr
og
r
am
m
i
ng
l
an
gu
a
ge
,
bu
t
th
e
po
wer
c
on
s
um
pti
on
n
ee
ds
to
be
m
i
ni
m
i
z
ed
an
d
bu
i
l
t
i
nto
s
m
al
l
pa
c
k
ag
es
.
K
i
m
de
v
e
l
op
ed
a
f
i
n
ge
r
-
r
eh
ab
i
l
i
t
ati
on
r
o
bo
t,
a
no
n
-
pa
tt
ern
r
ec
og
n
i
ti
on
ba
s
ed
,
us
i
n
g
t
w
o
di
gi
t
al
s
i
gn
a
l
proc
es
s
ors
T
MS
32
0F
28
12
f
or
hi
gh
-
s
p
ee
d
c
o
ntr
ol
s
y
s
tem
[12]
.
In
hi
s
s
tu
d
y
,
the
f
i
n
ge
r
-
r
e
ha
b
i
l
i
tat
i
on
r
ob
ot
m
ov
e
d
ac
c
ordi
ng
to
th
e
pa
ti
e
nt’
s
f
i
ng
ers
when
the
E
MG
s
i
g
na
l
ex
c
e
ed
e
d
a
pre
-
de
f
i
n
ed
thres
h
ol
d
v
ol
ta
ge
.
In
f
utu
r
e
r
es
ea
r
c
h,
he
s
ug
ge
s
ted
tes
ti
n
g
th
e
de
v
el
op
e
d
f
i
n
ge
r
r
ob
ot
t
o
r
e
ha
b
i
l
i
tat
e
a
s
tr
ok
e
pa
ti
en
t
’
s
f
i
ng
ers
an
d
p
ati
en
t
’
s
pa
r
al
y
z
ed
f
i
ng
ers
.
Le
n
z
i
de
v
e
l
op
ed
a
po
w
er
-
as
s
i
s
ted
up
p
er
l
i
m
b
ex
os
k
e
l
eto
n
us
i
ng
a
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
E
mb
ed
d
ed
s
y
s
tem
for up
pe
r
-
l
i
mb
ex
os
k
el
e
ton
.
.. (
T
r
i
wi
y
an
to
T
r
i
wi
y
a
nto
)
2993
non
-
p
att
ern
r
ec
og
ni
t
i
o
n
ba
s
ed
on
the
E
MG
w
i
t
h
a
p
r
op
orti
on
a
l
c
on
tr
o
l
[8]
.
T
hi
s
proporti
o
na
l
al
g
orit
hm
i
s
proc
es
s
ed
us
i
n
g
i
nte
grate
d
E
MG
(
I
E
MG
)
a
s
f
ea
ture
ex
tr
ac
ti
on
,
an
d
th
en
the
f
ea
ture
was
f
i
l
tere
d
us
i
ng
a
ba
n
dp
as
s
f
i
l
ter
(
B
P
F
)
w
i
th
a
s
pe
c
i
f
i
ed
c
ut
-
of
f
f
r
eq
ue
nc
y
.
I
n
hi
s
s
tu
d
y
,
h
e
us
ed
a
s
tan
da
r
d
c
om
pu
ter
t
o
i
m
pl
em
en
t
the
m
eth
od
.
A
l
tho
u
gh
th
e
s
y
s
tem
ha
s
be
e
n
r
ep
orte
d
as
a robu
s
t
an
d s
tr
a
i
g
htf
orw
ard m
eth
od
, th
e a
c
c
urac
y
of
the
pre
di
c
t
i
on
ha
s
n
ot
b
ee
n
ev
a
l
u
ate
d
y
e
t.
G
en
eral
l
y
,
the
pre
di
c
ti
on
a
l
go
r
i
thm
w
as
i
m
pl
em
en
ted
us
i
ng
a
s
of
tware
ap
p
l
i
c
at
i
o
n
s
uc
h
as
M
A
T
LA
B
an
d
o
the
r
v
i
s
ua
l
pro
gram
m
i
ng
ap
p
l
i
c
a
ti
on
s
us
i
ng
a
s
tan
d
ard
pe
r
s
on
a
l
c
om
pu
ter.
It
w
as
r
ar
e
to
f
i
nd
i
n
th
e
l
i
teratur
e
tha
t
d
i
s
c
us
s
ed
th
e
i
m
pl
em
en
tat
i
o
n
of
th
e
e
l
b
o
w
-
j
oi
nt
an
gl
e
predi
c
t
i
on
on
a
n
em
be
dd
e
d
s
y
s
tem
.
T
he
r
ef
ore,
i
t
i
s
ne
c
es
s
ar
y
t
o
i
m
pl
em
en
t
the
j
oi
nt
an
g
l
e
predi
c
t
i
on
al
g
orit
hm
i
n
the
em
be
dd
ed
s
y
s
tem
to
r
e
du
c
e
the
s
i
z
e
of
the
s
y
s
t
e
m
,
to
m
i
ni
m
i
z
e
the
po
wer
c
on
s
um
pti
on
a
nd
to
ac
h
i
e
v
e
a
r
ea
l
-
ti
m
e
s
y
s
tem
.
T
hi
s
s
tud
y
pres
e
nts
a
r
ea
l
-
ti
m
e
m
y
o
el
ec
tr
i
c
c
on
tr
ol
wi
th
a
no
n
-
p
att
ern
r
ec
og
ni
t
i
o
n
m
eth
od
f
or
an
u
pp
er
l
i
m
b
ex
o
s
k
el
eto
n
us
i
ng
a
l
o
w
-
c
os
t
em
be
dd
ed
m
i
c
r
oc
on
tr
ol
l
er
A
RM
S
T
M3
2F4
29
.
T
hi
s
m
i
c
r
oc
on
tr
ol
l
er
h
as
l
o
w
p
o
w
er
c
on
s
um
pti
on
a
nd
a
h
i
g
h
-
s
pe
ed
c
l
oc
k
proc
es
s
or.
T
he
c
on
tr
i
bu
ti
o
n
of
thi
s
s
tud
y
i
s
t
ha
t
the
al
g
orit
hm
i
m
pl
em
en
ted
i
n
the
m
i
c
r
oc
on
tr
ol
l
er
s
y
s
t
em
c
an
ad
ap
t
to
s
om
e
v
ar
i
ab
l
e
s
uc
h
as
ne
w
s
ub
j
ec
t
an
d
s
pe
e
d
of
m
ov
e
m
en
t
wi
th
ou
t
a
l
ea
r
n
i
ng
proc
es
s
.
I
n
th
i
s
prop
os
ed
m
eth
od
,
t
o
ad
ap
t
wi
th
a
n
e
w
v
aria
bl
e
,
th
e
s
y
s
t
em
on
l
y
ne
e
ds
an
ad
j
us
tm
en
t
on
th
e
thre
s
ho
l
d
v
al
ue
of
the
f
ea
t
ure
an
d
g
ai
n
i
n
th
e
bi
o
-
am
pl
i
f
i
er
m
od
ul
e
.
F
i
na
l
l
y
,
th
e
pe
r
f
or
m
an
c
e
of
th
e
predi
c
t
i
on
w
as
ev
a
l
u
ate
d
us
i
n
g roo
t m
ea
n s
qu
are err
or (
RM
S
E
)
an
d P
ea
r
s
on
’
s
c
orr
e
l
at
i
on
c
o
ef
f
i
c
i
en
t (r
)
.
2.
M
ateri
al
and
M
eth
o
d
2.1.
S
y
stem
A
r
chit
ec
t
u
r
e
T
he
s
y
s
t
em
arc
hi
tec
ture
f
or
the
ex
os
k
el
eto
n
b
as
ed
o
n
th
e
el
ec
tr
om
y
og
r
a
ph
y
c
o
ntrol
i
s
s
ho
w
n
i
n
F
i
g
ure
1.
T
he
s
y
s
t
em
c
on
s
i
s
ts
of
f
ou
r
m
od
ul
e
s
:
a
bi
o
-
am
pl
i
f
i
er,
a
n
MCU,
a
s
erv
o
m
oto
r
driv
er
an
d
c
o
ntrol
l
er,
an
d
a
n
ex
os
k
el
eto
n
u
ni
t
w
i
t
h
on
e
de
gree
of
f
r
ee
do
m
(
DO
F
)
.
B
i
o
-
am
pl
i
f
i
er
m
od
ul
e:
T
he
bi
o
-
am
pl
i
f
i
er
c
on
s
i
s
ts
of
an
A
D
62
0
pre
-
am
pl
i
f
i
er,
ba
nd
pa
s
s
f
i
l
ter
(
B
P
F
)
,
5
0
H
z
no
tc
h
f
i
l
t
er, m
ai
n a
m
pl
i
f
i
er,
an
d s
um
m
i
ng
am
pl
i
f
i
er.
0
O
90
O
45
O
145
O
P
r
e
-
A
m
p
l
i
f
i
e
r
AD
620
B
a
n
d
P
a
ss
F
i
l
t
e
r
N
o
t
ch
F
i
l
t
e
r
P
o
st
-
A
m
p
l
i
f
i
e
r
S
u
m
m
i
n
g
A
m
p
l
i
f
i
e
r
F
e
a
t
u
r
e
E
xt
r
a
ct
i
o
n
Z
e
r
o
C
r
o
s
s
i
n
g
I
I
R
-
L
o
w
P
a
s
s
F
i
l
t
e
r
P
u
l
se
W
i
d
t
h
M
o
d
u
l
a
t
i
o
n
D
u
t
yC
yc
l
e
=
f
(
ZC
_
L
P
F
)
M
o
d
u
l
e
-
M
C
U
M
o
d
u
l
e
-
B
i
o
a
m
p
l
i
f
i
e
r
M
o
d
u
l
e
-
D
r
i
ve
r
A
/
D
C
o
n
ve
r
t
e
r
C
o
n
t
r
o
l
l
e
r
D
r
i
ve
r
1
2
3
Note
:
1
:
Po
ten
tio
mete
r
2
.
Se
rvo
Mot
or
3
.
Electrode
E
M
G
F
i
gu
r
e
1.
S
y
s
tem
arc
hi
tec
tu
r
e
Evaluation Warning : The document was created with Spire.PDF for Python.
◼
IS
S
N: 16
93
-
6
93
0
T
E
L
KO
M
NIK
A
V
ol
.
17
,
No
.
6,
D
ec
em
be
r
20
19
:
29
9
2
-
300
2
2994
W
e
us
ed
a
po
ten
t
i
om
ete
r
s
en
s
or
to
r
e
ad
t
he
r
ea
l
po
s
i
t
i
on
of
the
el
b
o
w
j
o
i
nt
.
MCU
m
od
ul
e
:
T
he
MC
U
was
us
ed
t
o
ac
q
ui
r
e
th
e
d
ata
f
r
om
the
E
MG
an
d
th
e
p
ote
nt
i
om
ete
r
s
en
s
or.
It
was
al
s
o
us
e
d
to
i
m
pl
em
en
t
the
al
go
r
i
thm
of
di
gi
t
al
s
i
g
na
l
proc
es
s
i
ng
(
D
S
P
)
.
S
erv
o
m
oto
r
driv
er
m
od
ul
e
:
T
hi
s
m
od
ul
e
was
us
ed
to
c
on
tr
o
l
w
h
eth
er
the
s
erv
om
oto
r
r
ota
t
es
i
n
a
c
l
oc
k
w
i
s
e
(
C
W
)
or
c
ou
nte
r
c
l
oc
k
w
i
s
e
(
CC
W
)
di
r
ec
ti
on
or
n
ot.
E
x
os
k
el
eto
n
:
t
he
ex
os
k
el
eto
n
c
on
s
i
s
ts
of
three
a
l
um
i
nu
m
be
a
m
s
.
T
he
ex
os
k
el
eto
n
h
as
two
j
oi
nts
,
the
el
b
o
w
,
a
n
d
s
ho
u
l
de
r
.
T
he
s
erv
o
m
oto
r
w
as
pl
ac
ed
i
n
th
e
el
b
o
w
j
oi
nt
p
os
i
t
i
on
f
or
on
e
D
O
F
m
oti
on
.
F
i
g
ure
2
(
a)
s
ho
w
s
a
pi
c
ture
of
a
m
an
uf
ac
tured
E
MG
a
m
pl
i
f
i
er
an
d
A
R
M
S
T
M3
2
m
i
c
r
oc
on
tr
ol
l
er
s
y
s
tem
.
T
he
a
m
pl
i
f
i
er
bo
ard
c
a
n
r
ec
ei
v
e
f
ou
r
E
M
G
s
i
gn
a
l
s
an
d
on
e
a
na
l
og
i
np
ut
i
n
the
r
a
ng
e
of
0
-
3V
.
T
he
ou
tpu
ts
of
the
b
i
o
-
am
pl
i
f
i
er
w
ere
i
n
te
r
f
ac
ed
to
A
/D
(
a
na
l
og
t
o
di
g
i
ta
l
c
on
v
erter
)
of
a
D
i
s
c
ov
er
y
bo
ard
A
RM
S
T
M3
2F42
9.
(
a)
(
b)
F
i
gu
r
e
2.
(
a) T
he
m
an
uf
ac
ture of
the
bi
o
-
am
pl
i
f
i
er and
the
i
nte
r
f
ac
e o
f
th
e
di
s
c
o
v
e
r
y
bo
ard
A
RM
S
T
M3
2F42
9,
(
b)
T
he
pl
ac
em
en
t o
f
di
s
p
os
ab
l
e e
l
ec
tr
od
es
(
A
g
/A
gCl
)
at
th
e b
i
c
ep
s
m
us
c
l
es
2.2.
E
xpe
r
imen
t
al
P
r
o
t
o
col
F
ou
r
he
al
t
h
y
m
al
e
s
u
bj
ec
ts
wer
e
i
nv
ol
v
e
d
i
n
thi
s
s
tud
y
.
B
ef
ore
pe
r
f
or
m
i
ng
the
ex
pe
r
i
m
en
t,
s
ub
j
ec
t
s
w
ere
gi
v
e
n
a
n
i
nf
orm
c
on
s
en
t
t
o
be
r
e
ad
an
d
s
i
gn
ed
.
T
hi
s
ex
pe
r
i
m
en
tal
protoc
o
l
ha
d
be
e
n
a
pp
r
o
v
e
d
b
y
the
E
t
hi
c
s
Com
m
i
tte
e
of
He
al
t
h
P
ol
y
t
ec
hn
i
c
of
S
urab
a
y
a
,
M
i
n
i
s
tr
y
of
Hea
l
th
In
do
ne
s
i
a
.
In
t
he
d
ata
c
ol
l
ec
t
i
ng
proc
es
s
,
the
s
u
bj
ec
ts
w
ere
i
n
a
s
tan
d
-
up
po
s
i
t
i
on
an
d
h
el
d
an
ex
os
k
el
eto
n
f
r
a
m
e.
T
he
E
MG
s
i
gn
a
l
was
c
ol
l
ec
t
ed
f
r
om
bi
c
ep
s
wi
th
a
bi
po
l
ar
c
o
nf
i
gu
r
at
i
on
us
i
n
g
t
wo
di
s
p
os
ab
l
e
el
ec
tr
od
es
(
A
g/A
gCl
,
s
i
z
e:
5
7x
4
8
m
m
,
A
m
bu
,
B
l
u
e
S
en
s
or
R,
M
al
a
y
s
i
a)
an
d
on
e
el
ec
tr
od
e
as
c
om
m
on
ground
(
F
i
gu
r
e
2
(
b))
.
T
he
el
ec
t
r
o
de
s
pl
ac
em
en
t
w
as
i
n
ac
c
orda
nc
e
w
i
th
S
E
NIA
M
(
S
urf
ac
e
E
l
ec
tr
om
y
og
r
ap
h
y
f
or
t
h
e
No
n
-
In
v
as
i
v
e
A
s
s
es
s
m
en
t
of
Mu
s
c
l
e)
r
ul
es
[
13
]
.
B
i
c
ep
s
an
d
tr
i
c
ep
s
are
r
es
p
on
s
i
bl
e
f
or
the
f
l
ex
i
o
n
a
nd
ex
ten
s
i
o
n
m
ov
em
en
t
r
es
pe
c
ti
v
el
y
as
de
s
c
r
i
be
d
i
n
th
e
l
i
terat
ure
[
14
]
.
Ho
w
e
v
er,
i
n
t
he
s
a
gi
t
tal
pl
a
ne
m
oti
on
,
L
en
z
i
pro
v
e
d
tha
t
th
e
E
MG
s
i
gn
al
g
e
ne
r
ate
d
f
r
o
m
tr
i
c
ep
s
s
ho
wed
i
ns
i
gn
i
f
i
c
an
t
ac
ti
v
i
t
i
es
[8]
.
T
he
r
ef
ore,
i
n
thi
s
w
ork
,
the
m
ov
em
en
t
of
the
ex
os
k
el
eto
n
w
as
on
l
y
c
on
tr
ol
l
ed
b
y
bi
c
ep
s
.
2.3.
Z
er
o
-
c
r
o
s
sing
Feat
u
r
e
Z
ero
c
r
os
s
i
ng
c
ou
nts
a
n
um
be
r
of
the
ti
m
e
of
th
e
s
i
gn
a
l
s
t
ha
t
c
r
os
s
es
pre
-
d
ef
i
ne
d
thres
ho
l
d a
m
pl
i
t
ud
e
. T
he
z
ero
-
c
r
os
s
i
ng
i
s
f
orm
ul
ate
d
as
s
ho
w
n
i
n (1)
[
15
]
:
=
∑
[
s
gn
(
.
+
1
)
∩
|
.
+
1
|
≥
ℎ
ℎ
]
−
1
=
1
s
gn
(
)
=
{
1
,
→
≥
ℎ
ℎ
0
,
ℎ
(
1)
where
x
i
i
nd
i
c
ate
s
t
he
i
-
th
E
MG
s
i
gn
a
l
s
,
N
i
n
di
c
at
es
t
he
n
um
be
r
of
s
a
m
pl
es
i
n
e
ac
h
s
eg
m
en
t,
an
d
th
e
t
hres
ho
l
d
i
n
di
c
at
es
the
l
e
v
el
of
am
pl
i
tud
e.
T
he
E
MG
s
i
gn
a
l
s
wer
e
ex
tr
ac
ted
f
or
ea
c
h
of
a n
um
be
r
of
th
e s
am
pl
e p
o
i
nts
to
g
et
a f
ea
ture
[
16
]
as
s
ho
w
n
i
n
F
i
gu
r
e
3.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
E
mb
ed
d
ed
s
y
s
tem
for up
pe
r
-
l
i
mb
ex
os
k
el
e
ton
.
.. (
T
r
i
wi
y
an
to
T
r
i
wi
y
a
nto
)
2995
F
i
gu
r
e
3.
T
he
proc
es
s
of
th
e f
ea
ture ex
tr
ac
t
i
on
of
th
e
E
MG
s
i
gn
a
l
s
(
a) r
aw
E
MG
s
i
gn
al
s
,
a
r
ed
l
i
n
e i
s
a m
ea
s
ured a
n
gl
e
an
d a
b
l
ac
k
l
i
ne
i
s
th
e
E
MG
s
i
gn
a
l
,
(
b) EM
G
s
i
gn
a
l
s
are
ex
tr
ac
ted
e
v
er
y
20
0 s
am
pl
e
s
or 1
00
m
s
ec
(
f
sampli
n
g
=
2
00
0
H
z
)
, S
n
de
no
t
es
the
l
en
gth
of
th
e
wi
nd
o
w
,
a
nd
(
c
)
T
y
p
i
c
al
f
ea
t
ures
of
th
e E
MG
s
i
gn
al
s
f
or 10
00
m
s
ec
2.4.
Inf
init
e I
mp
u
lse
Respo
n
s
e
Inf
i
ni
t
e
i
m
pu
l
s
e
r
es
po
ns
e
(
IIR)
i
s
a
di
gi
t
al
f
i
l
ter
i
m
pl
em
en
ted
us
i
n
g
a
progr
a
m
m
i
ng
l
an
gu
a
ge
wi
t
h
a
p
arti
c
u
l
ar
al
g
orit
hm
.
T
he
IIR
i
s
a
r
ec
urs
i
v
e
f
i
l
ter
[1
7]
.
I
n
t
hi
s
s
t
u
d
y
,
the
IIR
f
i
l
t
e
r
was
de
s
i
g
ne
d
as
the
s
ec
on
d
-
order
l
o
w
-
pa
s
s
B
utt
er
w
or
th
f
i
l
ter
(
L
P
F
)
.
P
r
e
v
i
ou
s
r
es
ea
r
c
he
r
s
us
ed
a
d
i
g
i
ta
l
f
i
l
t
er
f
or
E
MG
s
i
g
na
l
proc
es
s
i
n
g
[
18
–
20]
.
T
he
tr
a
ns
f
er
f
un
c
ti
on
of
the
di
gi
t
al
f
i
l
ter
i
s
ex
pres
s
ed
i
n (2)
. T
he
l
i
ne
ar
f
un
c
ti
on
i
s
s
ho
w
n
i
n
(
3)
[17
]
.
=
(
)
(
)
=
0
+
1
−
1
+
2
−
2
+
.
.
.
+
−
1
+
1
−
1
+
2
−
2
+
.
.
.
+
−
(
2
)
[
]
=
∑
[
−
]
=
0
−
∑
[
−
]
=
1
(
3
)
where
b
k
an
d
a
k
are th
e (
M
+
1) n
um
erator and
N
de
no
m
i
na
tor c
oe
f
f
i
c
i
en
t, res
pe
c
ti
v
e
l
y
.
2.5.
S
y
stem
Imp
lem
ent
atio
n
T
he
pr
ed
i
c
t
i
on
a
l
g
orit
hm
o
f
the
e
l
b
o
w
j
o
i
nt
an
gl
e
wa
s
i
m
pl
em
en
ted
i
n
t
he
em
be
dd
e
d
s
y
s
t
em
A
RM
S
T
M3
2F42
9
m
i
c
r
oc
on
tr
ol
l
er
w
i
t
h
a
n
i
nt
erna
l
c
l
oc
k
s
pe
ed
of
18
0
MH
z
.
T
he
predi
c
ti
o
n
al
go
r
i
t
hm
c
on
s
i
s
ts
of
the
z
ero
-
c
r
os
s
i
n
g
f
ea
ture
an
d
l
o
w
p
as
s
B
u
tte
r
w
orth
f
i
l
t
er.
T
he
a
l
go
r
i
thm
of
the
z
ero
-
c
r
os
s
i
ng
f
ea
ture
was
i
m
pl
e
m
en
ted
as
s
ho
w
n
i
n
A
l
g
ori
thm
1
w
h
i
c
h
i
s
ac
c
ordi
ng
to
(
1).
S
om
e
pa
r
am
ete
r
s
s
uc
h
as
a
nu
m
be
r
of
s
am
pl
e
p
oi
nt
N
a
nd
t
he
Z
C
f
ea
ture
thres
ho
l
d
ne
ed
to
b
e
i
n
i
t
i
a
l
i
z
ed
wi
th
a
s
pe
c
i
f
i
c
v
a
l
ue
.
T
he
z
ero
-
c
r
os
s
i
ng
a
l
go
r
i
t
hm
r
ec
ei
v
ed
i
np
u
t
da
ta
f
r
om
the
E
MG
s
i
gn
al
s
E
MG
k
an
d
r
es
ul
ted
i
n
a
n
ou
tp
ut
na
m
ed
E
MG
ZC
.
T
he
ou
tpu
t
of
the
f
ea
ture
ex
tr
ac
ti
on
was
us
ed
as
a
n i
np
u
t o
f
th
e
B
u
tt
erw
orth
l
o
w
-
p
as
s
f
i
l
ter.
Evaluation Warning : The document was created with Spire.PDF for Python.
◼
IS
S
N: 16
93
-
6
93
0
T
E
L
KO
M
NIK
A
V
ol
.
17
,
No
.
6,
D
ec
em
be
r
20
19
:
29
9
2
-
300
2
2996
A
l
g
orit
hm
1:
A
l
g
orit
hm
of
th
e ZC f
ea
ture
Init:
N
= 200;
threshold
= 0.12
;
EMG
ZC
= 0;
Input:
EMG
[k]
Output:
EMG
ZC
1.
2.
3.
4.
5.
FOR k → 0 TO
N
-
1
DO
sgn
=
EMG
[k] *
EMG
[k+1]
diff
= ABS (
EMG
[k]
-
EMG
[k+1])
IF (
sgn
>=
threshold
AND
diff
>=
threshold
)
THEN
EMG
ZC
=
EMG
ZC
+ 1
End
A
l
g
orit
hm
2:
T
he
al
go
r
i
thm
of
th
e
l
o
w
-
pa
s
s
B
utt
er
wor
th
f
i
l
ter
Init:
a
[1],
a
[2],
b
[0],
b
[1],
b
[2]
Input
:
EMG
ZC
Output:
EMG
ZC_LPF
1.
2.
3.
4.
5.
6.
7.
x
0 =
EMG
ZC
x
1 =
x
0
x
2 =
x
1
y
1 =
y
0
y
2 =
y
1
y
0 =
b
[0]
x
0 +
b
[1]
x
1 +
b
[2]
x
2
-
a
[1]
y
1
-
a
[2]
y
2
EMG
ZC_LPF
=
y
0
End
A
l
g
orit
hm
2
i
s
the
i
m
pl
em
e
nta
t
i
on
of
the
i
nf
i
n
i
te
i
m
pu
l
s
e
r
es
po
ns
e
f
i
l
ter
i
n
th
e
l
o
w
p
as
s
B
utt
e
r
w
orth
f
i
l
ter
b
y
(
3).
T
he
c
oe
f
f
i
c
i
en
ts
of
the
i
nf
i
n
i
te
i
m
pu
l
s
e
r
es
po
ns
e
f
i
l
ter
we
r
e
c
al
c
ul
ate
d
us
i
ng
a
MA
T
LA
B
pr
og
r
am
(
S
tud
en
t
V
ers
i
on
,
M
ath
W
or
k
s
In
c
., M
as
s
ac
hu
s
ett
s
, U
S
A
)
. A
l
go
r
i
thm
2
de
s
c
r
i
be
s
t
he
s
ec
o
nd
-
ord
e
r
of
a
B
utt
erw
orth
l
o
w
p
as
s
f
i
l
ter.
T
hi
s
al
go
r
i
t
hm
r
ec
ei
v
ed
i
np
u
t
f
r
o
m
the
r
es
ul
t
of
th
e
f
ea
t
ure
ex
t
r
ac
ti
on
E
MG
Z
C
an
d
r
es
u
l
t
ed
i
n
ou
tpu
t
as
E
M
G
Z
C_L
P
F
.
T
he
ou
t
pu
t
of
the
B
utt
er
wor
th
l
o
w
-
p
as
s
f
i
l
ter
r
ep
r
es
e
nts
a
pred
i
c
ted
a
ng
l
e
of
th
e
e
l
b
o
w
j
oi
nt.
T
he
al
g
orit
hm
of
the
z
ero
-
c
r
os
s
i
ng
f
ea
tu
r
e
an
d
the
l
o
w
p
as
s
B
u
tt
erw
orth
f
i
l
ter
w
ere
c
od
ed
an
d
c
om
pi
l
ed
i
n
the
C
progr
am
m
i
ng
l
an
g
ua
ge
,
us
i
n
g
the
Co
oCox
-
C
oIDE
(
V
ers
i
on
1
.77
,
E
m
be
s
t
T
ec
hn
ol
o
g
y
Co.,
L
td.
,
W
uh
an
Un
i
v
ers
i
t
y
of
T
ec
hn
ol
og
y
, C
hi
n
a).
2.6.
T
imin
g
Co
n
t
r
o
l
M
anag
ement
T
he
MCU p
erip
he
r
a
l
was
c
on
tr
ol
l
e
d u
s
i
ng
T
i
m
er 5 an
d
T
i
m
er 2 as
s
ho
wn
i
n F
i
gu
r
e 4
(
a).
T
i
m
er
2
w
as
us
e
d
to
m
an
a
ge
th
e
s
am
pl
i
ng
r
a
te
of
th
e
an
a
l
og
to
d
i
g
i
ta
l
c
on
v
erter
(
A
DC)
.
T
i
m
er 3
was
us
ed
to
ge
ne
r
at
e
a
pu
l
s
e
wi
dt
h
m
od
ul
at
i
on
(
P
W
M).
T
i
m
er
1
an
d
T
i
m
er
4
wer
e
no
t
us
ed
i
n
thi
s
s
tud
y
.
T
he
P
W
M
r
e
gi
s
t
er
w
as
up
d
ate
d
e
v
er
y
m
i
l
l
i
s
ec
o
nd
us
i
ng
T
i
m
er
5
i
nt
err
up
ts
.
T
he
r
es
ul
ts
of
the
pre
di
c
t
i
o
n
a
l
g
orit
hm
(
E
MG
Z
C
-
L
P
F
)
wer
e
us
ed
to
up
da
t
e
t
he
P
W
M
r
eg
i
s
ter
s
o
tha
t
t
he
d
ut
y
c
y
c
l
e
(
DC)
of
the
P
W
M
f
i
t
the
po
s
i
ti
on
o
f
the
el
b
o
w
j
oi
nt.
C
om
m
un
i
c
ati
on
am
on
g
the
pe
r
i
ph
e
r
al
us
ed
d
i
r
ec
t
m
e
m
o
r
y
ac
c
es
s
(
DMA
)
to
g
et
r
e
al
-
ti
m
e
da
ta
proc
es
s
i
n
g.
U
S
A
RT
was
us
ed
to
s
en
d
the
A
DC
r
es
ul
ts
to
th
e
c
om
pu
ter f
or an o
f
fl
i
n
e d
ata
an
a
l
y
s
i
s
.
F
i
gu
r
e
4
(
b)
de
s
c
r
i
be
s
t
he
proc
es
s
i
ng
t
i
m
e
f
or
t
he
el
bo
w
j
oi
nt
an
g
l
e
pre
di
c
ti
on
.
T
he
proc
es
s
i
ng
ti
m
e
of
the
predi
c
t
i
on
a
l
g
orit
hm
c
on
s
i
s
ts
of
ti
m
e
to
bu
ff
er
a
nu
m
be
r
of
s
a
m
pl
es
po
i
nt
ac
c
ord
i
ng
to
wi
n
do
ws
l
en
g
th
t
w
,
ti
m
e
to
proc
e
s
s
the
f
ea
ture
ex
tr
ac
ti
o
n
t
FE
an
d
ti
m
e
to
proc
es
s
the
l
o
w
pa
s
s
B
ut
te
r
w
ort
h
f
i
l
ter
t
LPF
.
I
n
th
i
s
pa
p
er,
a
n
um
be
r
of
s
a
m
pl
es
an
d
th
e
ord
er
of
the
B
utt
er
wor
th
d
i
g
i
ta
l
f
i
l
te
r
wer
e
v
arie
d
to
g
et
th
e
op
ti
m
u
m
pe
r
f
or
m
an
c
e.
T
he
E
M
G
s
i
gn
a
l
s
ha
v
e
a
f
r
eq
ue
nc
y
r
a
ng
e
of
20
–
5
00
H
z
[
21
]
.
I
n
th
i
s
s
tud
y
,
t
he
E
MG
was
r
ec
orded
us
i
ng
A
D
C,
w
i
t
h
a
2,0
00
H
z
s
am
pl
i
ng
r
at
e,
whi
c
h
i
s
ac
c
ord
i
ng
to
t
he
N
y
q
ui
s
t
r
e
qu
i
r
em
en
t
[17]
.
A
no
the
r
s
tu
d
y
us
ed
a
s
am
pl
i
n
g
f
r
eq
ue
nc
y
of
20
48
H
z
t
o
c
ol
l
ec
t
t
he
E
MG
s
i
g
na
l
[2
2]
.
T
he
E
M
G
s
i
gn
a
l
s
w
ere
ex
tr
ac
ted
w
i
t
h
v
ar
y
i
ng
wi
n
do
w
l
en
gt
hs
,
f
r
o
m
10
0
to
20
0
s
am
pl
es
(
50
m
s
ec
to
10
0
m
s
ec
f
o
r
a 2
,
0
00
H
z
s
am
pl
i
ng
r
ate
)
.
Li
s
e
l
e
c
te
d a
10
0 s
am
pl
es
wi
nd
o
w
t
o
i
m
pl
em
en
t a
r
ea
l
-
ti
m
e c
on
tr
ol
i
n
up
pe
r
l
i
m
b
ex
os
k
el
eto
n
[
7]
.
E
d
w
ar
d
us
ed
a
wi
n
d
o
w
l
en
gt
h
of
24
8
m
s
ec
to
es
ti
m
ate
the
E
MG
am
pl
i
tud
e
[2
2]
.
O
s
k
oe
i
ex
pl
ore
d
a
wi
nd
o
w
l
en
gth
b
et
w
ee
n
50
t
o
50
0
m
s
ec
to
g
et
be
tt
er
pe
r
f
or
m
an
c
e o
f
th
e
c
l
as
s
i
f
i
er
[23]
. T
r
i
w
i
y
an
to
et
a
l
. s
tu
di
ed
t
ha
t
th
e
pe
r
f
orm
an
c
e o
f
th
e
pred
i
c
ti
o
n
w
as
af
f
ec
ted
b
y
wi
nd
o
w
l
e
ng
th
[
24
]
.
T
he
order
of
the
B
utt
er
w
orth
f
i
l
ter
w
as
v
ari
ed
f
r
o
m
tw
o
to
ei
g
ht
ord
ers
to
f
i
nd
an
op
t
i
m
u
m
order of
th
e f
i
l
ter.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
E
mb
ed
d
ed
s
y
s
tem
for up
pe
r
-
l
i
mb
ex
os
k
el
e
ton
.
.. (
T
r
i
wi
y
an
to
T
r
i
wi
y
a
nto
)
2997
2.7.
S
er
v
o
M
o
t
o
r
Co
n
t
r
o
l
T
he
E
MG
s
i
gn
a
l
w
as
ex
tr
a
c
ted
us
i
n
g
z
ero
-
c
r
os
s
i
ng
f
e
atu
r
e
ex
tr
ac
ti
on
,
a
nd
th
en
i
t
w
as
s
m
oo
the
ne
d
us
i
n
g
th
e
l
o
w
pa
s
s
B
utt
erw
orth
f
i
l
t
er.
T
he
ou
tpu
t
of
the
l
o
w
pa
s
s
B
ut
terwor
th
f
i
l
t
er
was
c
al
l
e
d
E
MG
ZC_LPF
.
T
he
s
erv
o
m
oto
r
w
as
c
on
tr
o
l
l
e
d
b
y
p
ul
s
e
-
wi
dt
h
m
od
ul
at
i
o
n
(
P
W
M
)
w
i
th
a
p
arti
c
u
l
ar
du
t
y
c
y
c
l
e
(
DC)
,
as
s
h
o
w
n
i
n
F
i
gu
r
e
5
.
T
he
du
t
y
c
y
c
l
e
r
a
ng
e
of
th
e
P
W
M
w
as
ba
s
ed
on
t
he
E
MG
ZC_LPF
v
al
ue
.
T
he
DC
w
as
c
al
i
brat
ed
s
o
t
ha
t
th
e
ex
os
k
el
eto
n’
s
r
an
g
e
o
f
the
m
oti
on
i
s
f
r
o
m
0
o
to 150
o
.
T
I
M
E
R
5
T
I
M
E
R
3
=
P
W
M
T
I
M
E
R
2
A
D
C
D
M
A
1
U
S
A
R
T
S
R
A
M
D
M
A
2
E
M
G
ZC
-
L
P
F
A
R
M
S
T
M
32
F
429
P
e
r
so
n
a
l
C
o
m
p
u
t
e
r
C
o
n
t
r
o
l
l
e
r
D
r
i
ve
r
S
e
r
vo
M
o
t
o
r
1
2
P
i
n
1
:
P
W
M
2
:
U
S
A
R
T
(
a)
F
e
a
t
u
r
e
e
xt
r
a
ct
i
o
n
D
i
g
i
t
a
l
l
o
w
p
a
s
s
f
i
l
t
e
r
(
2
nd
o
r
d
e
r
)
E
M
G
N
t
w
t
FE
t
L
P
F
B
U
F
F
E
R
N
th
P
r
e
d
i
ct
e
d
A
n
g
l
e
E
va
l
u
a
t
i
o
n
R
M
S
E
&
CC
D
a
ta
P
r
o
c
e
s
s
i
n
g
(
b)
F
i
gu
r
e
4.
(
a) R
ea
l
-
t
i
m
e c
on
tr
ol
of
th
e
pe
r
i
ph
er
al
i
n t
he
MCU m
od
ul
e,
an
d
(
b)
T
he
pro
c
es
s
i
ng
t
i
m
e f
or
the
pre
di
c
t
i
on
;
t
w
i
s
w
i
n
do
w
l
en
gth
,
t
FE
i
s
t
i
m
e f
or
f
ea
ture ex
tr
ac
ti
on
,
an
d
t
LPF
i
s
ti
m
e f
or di
gi
ta
l
f
i
l
ter
proc
e
s
s
i
ng
A
/
D
C
o
n
ve
r
t
e
r
F
e
a
t
u
r
e
E
xt
r
a
ct
i
o
n
D
r
i
ve
r
S
e
r
vo
M
o
t
o
r
B
i
o
-
a
m
p
l
i
f
i
e
r
L
o
w
P
a
ss
F
i
l
t
e
r
E
M
G
E
M
G
ZC
E
M
G
ZC
_
L
P
F
P
W
M
D
u
t
yC
yc
l
e
=
f
(
E
M
G
ZC
_
L
P
F
)
1
2
N
o
t
e
1
:
E
l
e
ct
r
o
d
e
2
:
S
e
r
vo
M
o
t
o
r
F
i
gu
r
e
5.
C
o
ntro
l
l
i
ng
th
e
e
x
os
k
el
eto
n u
s
i
ng
t
he
E
MG
s
i
gn
a
l
2.8.
E
v
aluat
ion
of
t
h
e P
er
f
o
r
man
ce
T
he
ev
a
l
ua
ti
o
n
pr
oc
es
s
wa
s
pe
r
f
or
m
ed
of
f
-
l
i
ne
us
i
n
g
a
c
om
pu
ter
(
Int
e
l
Cor
e
i
3
-
32
17
U
CP
U
at
1
.80
G
H
z
,
8
G
B
of
RA
M,
W
i
nd
ow
s
8)
.
T
he
E
MG
an
d
th
e
m
ea
s
ured
a
ng
l
e
wer
e
s
en
t
to
the
c
om
pu
ter
us
i
ng
a
US
A
RT
pe
r
i
ph
era
l
f
or
f
urther
da
ta
an
al
y
s
i
s
.
T
he
pred
i
c
ti
o
n
al
go
r
i
t
hm
w
as
Evaluation Warning : The document was created with Spire.PDF for Python.
◼
IS
S
N: 16
93
-
6
93
0
T
E
L
KO
M
NIK
A
V
ol
.
17
,
No
.
6,
D
ec
em
be
r
20
19
:
29
9
2
-
300
2
2998
an
a
l
y
z
e
d
an
d
v
i
s
u
al
i
z
ed
of
f
l
i
ne
us
i
ng
L
a
z
arus
progr
am
m
i
ng
(
V
ers
i
on
1.8
.
4)
.
T
he
predi
c
ted
an
g
l
e
was
ob
ta
i
n
ed
af
ter
the
f
ea
ture
ex
tr
ac
t
i
on
an
d
l
o
w
-
pa
s
s
f
i
l
ter
proc
es
s
.
T
he
p
erf
or
m
an
c
e
of
the
el
bo
w
j
oi
nt
an
gl
e
pred
i
c
ti
on
w
as
e
v
a
l
ua
t
ed
us
i
ng
a
P
ea
r
s
on
’
s
c
orr
el
at
i
o
n
c
oe
f
f
i
c
i
en
t
(
CC)
,
an
d
r
oo
t
m
ea
n
s
qu
are
err
or
(
RMS
E
)
.
T
he
CC
an
d
RMS
E
w
ere
c
a
l
c
ul
ate
d
us
i
ng
(
4)
an
d
(
5),
r
es
pe
c
ti
v
el
y
[
6,
2
5]
:
=
√
1
∑
(
−
)
2
=
1
(
4
)
=
∑
−
(
∑
)
(
∑
)
√
(
∑
2
−
(
∑
)
2
)
(
∑
2
−
(
∑
)
2
)
(
5
)
where
x
n
i
s
the
m
ea
s
ured
an
g
l
e,
y
n
i
s
th
e
pred
i
c
ted
a
ng
l
e,
an
d
N
i
s
th
e
nu
m
be
r
of
the
s
am
pl
e
s
po
i
nt.
T
he
CC
an
d
RM
S
E
wer
e
of
ten
us
ed
b
y
prev
i
o
us
r
es
ea
r
c
he
r
s
[7,
9
,
26
,
2
7]
to
ev
al
u
ate
the
p
erf
or
m
an
c
e
of
the
predi
c
t
i
on
.
T
he
C
C
an
d
RM
S
E
v
a
l
ue
s
wer
e
c
al
c
u
l
at
ed
us
i
ng
La
z
arus
pro
gram
m
i
ng
.
T
he
pro
po
s
e
d
m
eth
od
was
t
es
ted
w
i
t
h
t
hree
pe
r
i
od
s
of
m
oti
on
(
12
s
ec
,
8
s
ec
an
d
6
s
ec
)
to
s
ee
th
e
c
o
ns
i
s
ten
c
y
of
th
e
m
eth
od
.
In
th
i
s
s
t
ud
y
,
t
he
p
-
v
al
u
es
wer
e
us
e
d
to
ex
am
i
ne
whet
h
er
th
ere
w
ere
an
y
s
i
gn
i
f
i
c
an
t
d
i
f
f
erenc
es
be
twee
n
the
pe
r
i
od
s
of
the
m
oti
on
or
no
t
.
T
he
bi
c
ep
s
w
ere
us
e
d
to
predi
c
t
t
he
e
l
bo
w
j
oi
nt
an
g
l
e
.
In
th
i
s
s
tud
y
,
the
r
a
ng
e
of
the
m
ov
em
en
t
of
th
e e
l
bo
w j
oi
nt
an
gl
e
w
a
s
f
r
o
m
0
o
–
150
o
.
3.
Re
sult
s
and
D
isc
u
s
sio
n
T
he
el
bo
w
j
oi
nt
an
g
l
e
pr
e
di
c
ti
on
w
as
i
m
pl
em
en
ted
i
n
the
em
be
dd
e
d
s
y
s
t
e
m
us
i
ng
an
A
RM
S
T
M3
2F4
29
m
i
c
r
oc
on
tr
ol
l
er.
T
he
pre
di
c
ti
on
al
go
r
i
t
hm
c
ou
l
d
c
o
ntrol
t
he
ex
os
k
el
et
on
s
m
oo
thl
y
ac
c
ord
i
ng
t
o
t
he
hu
m
an
el
bo
w
j
oi
nt.
T
he
pr
op
os
ed
m
eth
od
c
an
pe
r
f
orm
i
n
r
ea
l
-
ti
m
e
wi
th
th
e
proc
es
s
i
n
g
ti
m
e
of
10
0.3
98
m
s
ec
(
the
nu
m
be
r
o
f
the
s
a
m
pl
e
po
i
n
t
s
are
20
0
w
i
t
h
the
s
ec
on
d
-
or
de
r
of
the
B
u
tte
r
w
orth
l
o
w
-
pa
s
s
f
i
l
ter)
.
T
he
pred
i
c
ti
o
n
al
go
r
i
t
hm
o
f
t
he
el
bo
w
j
oi
nt
an
g
l
e
wor
k
s
w
el
l
f
or
the
three
pe
r
i
od
s
of
m
oti
on
(
p
-
v
a
l
ue
s
>
0.0
5).
I
n
t
h
i
s
ex
pe
r
i
m
en
t,
the
proc
es
s
i
n
g
t
i
m
e
f
or
f
ea
ture
ex
tr
ac
t
i
on
i
s
6
4
µ
s
.
Li
n
r
ep
orted
t
he
proc
es
s
i
n
g
t
i
m
e
was
5
m
s
ec
us
i
ng
an
em
be
dd
e
d
F
r
ee
s
c
al
e
m
i
c
r
oc
on
tr
ol
l
er
(
M
P
C556
6)
[28]
.
I
n
th
i
s
prop
os
ed
m
eth
od
,
the
t
i
m
e
of
the
em
be
dd
e
d
s
y
s
tem
to
proc
es
s
the
predi
c
t
i
on
wh
i
c
h
c
on
s
i
s
ts
of
the
f
ea
ture
ex
tr
ac
ti
on
an
d
t
he
d
i
gi
tal
f
i
l
ter
of
th
e
s
ec
on
d
order
i
s
1
00
.3
9
m
s
ec
.
T
an
g
r
ep
orte
d
tha
t
th
e
t
i
m
e
to
pr
oc
es
s
the
pre
di
c
t
i
on
al
go
r
i
thm
c
on
s
i
s
ts
of
th
e
f
ea
ture
ex
tr
ac
t
i
on
a
nd
c
l
as
s
i
f
i
er
us
i
ng
a s
ta
nd
ard c
om
pu
ter
w
as
2
50
m
s
ec
[26]
.
3.1.
T
ime
E
xe
cut
ion
o
f
t
h
e
A
lg
o
r
it
h
m
F
i
gu
r
e
6
(
a)
s
h
o
w
s
th
e
pr
oc
es
s
i
ng
t
i
m
e
f
or
s
i
ng
l
e
f
ea
ture
ex
tr
ac
ti
o
n.
T
he
proc
es
s
i
ng
ti
m
e
of
the
f
ea
ture
ex
tr
ac
ti
on
w
as
te
s
te
d
f
or
s
ev
eral
nu
m
be
r
s
of
the
s
am
pl
e
po
i
nts
(
N
=
1
00
,
N
=
15
0,
an
d
N
=
20
0).
T
he
proc
es
s
i
ng
ti
m
e
of
th
e
l
o
w
p
as
s
B
utt
er
wor
th
f
i
l
ter
was
al
s
o
m
ea
s
ured i
n s
ev
eral
orders
(
2
nd
, 4
th
an
d 8
th
order
)
, a
s
s
ho
w
n
i
n F
i
g
ure 6
(
b).
(
a)
(
b)
F
i
gu
r
e
6.
(
a) T
h
e p
r
oc
es
s
i
n
g t
i
m
e o
f
th
e f
ea
ture
ex
tr
ac
ti
on
wi
t
h a
v
ar
i
ou
s
n
um
be
r
of
s
a
m
pl
e
po
i
nts
, a
n
d
(
b)
t
h
e p
r
oc
es
s
i
ng
t
i
m
e o
f
th
e l
o
w
pa
s
s
B
u
t
terwor
th
f
i
l
t
er
wi
th
di
f
f
erent o
r
de
r
f
i
l
ter
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
E
mb
ed
d
ed
s
y
s
tem
for up
pe
r
-
l
i
mb
ex
os
k
el
e
ton
.
.. (
T
r
i
wi
y
an
to
T
r
i
wi
y
a
nto
)
2999
3.2.
E
v
aluat
ion
of
t
h
e
P
r
edict
i
o
n
A
lgo
r
it
h
m
T
he
E
MG
s
i
gn
al
s
r
es
p
on
d
to
th
e
c
ha
ng
e
of
t
h
e
el
b
o
w
j
o
i
nt
an
gl
e
,
as
s
h
o
w
n
i
n
F
i
gu
r
e
7
(
a),
w
i
t
h
v
ar
y
i
ng
a
nd
r
an
d
om
a
m
pl
i
tud
es
(
0
to
3
m
V
pe
ak
to
pe
ak
)
[21]
.
In
thi
s
s
tu
d
y
,
the
pe
r
f
orm
an
c
e
of
the
propos
ed
m
eth
od
was
ev
al
u
ate
d
us
i
ng
t
he
thr
ee
d
i
f
f
er
en
t
pe
r
i
od
s
of
m
oti
on
.
T
he
pe
r
i
od
of
th
e
m
ov
e
m
en
t
was
s
y
nc
hr
on
i
z
e
d
us
i
n
g
the
Me
tr
o
n
om
e
(
Mi
c
r
os
o
ft
W
i
nd
ow
s
ap
p
l
i
c
at
i
o
n).
(
a)
(
b)
F
i
gu
r
e
7.
(
a) T
he
E
MG
s
i
g
n
al
f
or one
c
y
c
l
e
of
m
oti
on
(
f
l
ex
i
on
an
d
ex
ten
s
i
on
)
,
(
b) T
he
E
MG
f
ea
tures
a
nd
predi
c
te
d a
ng
l
e f
or on
e p
er
i
od
of
m
ov
em
en
t;
t
he
r
ed
l
i
ne
i
s
th
e
ex
p
ec
ted
an
g
l
e,
an
d
the
bl
ac
k
l
i
ne
i
s
th
e
m
ea
s
ured
an
gl
e
T
he
pe
r
i
od
s
us
e
d
i
n
th
i
s
wo
r
k
wer
e
12
s
,
8s
,
a
nd
6s
.
T
he
m
ea
n
c
orr
el
ati
on
c
oe
f
f
i
c
i
en
t
f
or
the
thre
e
of
di
f
f
erent
pe
r
i
o
ds
(
12
s
,
8s
,
an
d
6s
)
s
ho
w
s
a
hi
g
h
c
orr
el
at
i
o
n,
w
h
i
c
h
i
s
0.968
±
0.
01
2
,
0.9
7
9±
0.0
0
6
an
d
0
.98
5±
0.
00
3
r
es
pe
c
t
i
v
el
y
,
as
s
ho
w
n
i
n
F
i
gu
r
e
8(a)
.
T
he
ac
c
urac
y
(
RM
S
E
)
of
the
pred
i
c
ti
o
n
a
l
g
orit
hm
i
s
s
ho
w
n
i
n
F
i
gu
r
e
8(b)
.
T
he
RM
S
E
v
a
l
ue
v
arie
s
wi
t
h
a
v
arie
t
y
of
s
tan
da
r
d
d
ev
i
at
i
o
ns
b
y
th
e
pe
r
i
o
ds
of
the
m
oti
on
of
the
el
bo
w
j
o
i
nt
.
T
he
m
ea
ns
of
the
RM
S
E
f
or
the
m
oti
on
p
erio
ds
of
12
s
ec
,
8
s
ec
,
an
d
6
s
ec
are
1
4.0
5
3
o
±
1.6
63
o
,
11
.9
5
o
±
2.3
14
o
an
d
10
.30
o
±
1.4
53
o
, res
pe
c
t
i
v
el
y
.
(
a)
(
b)
F
i
gu
r
e
8.
T
he
c
om
pa
r
i
s
on
o
f
(
a) Pe
ars
on
’
s
c
orr
el
ati
on
c
oe
f
f
i
c
i
en
t a
n
d (b)
r
oo
t m
ea
n s
qu
ared
err
or of
th
e e
l
bo
w j
oi
n
t a
n
gl
e p
r
ed
i
c
ti
on
f
or thr
ee
di
f
f
erent p
er
i
od
s
(
1
2 s
ec
, 8
s
ec
,
a
nd
6
s
ec
)
Evaluation Warning : The document was created with Spire.PDF for Python.
◼
IS
S
N: 16
93
-
6
93
0
T
E
L
KO
M
NIK
A
V
ol
.
17
,
No
.
6,
D
ec
em
be
r
20
19
:
29
9
2
-
300
2
3000
A
s
s
ho
wn
i
n
T
ab
l
e
1,
the
p
-
v
al
ue
s
of
t
he
CC
f
or
t
he
m
oti
on
pe
r
i
od
s
of
12
s
,
8s
,
an
d
6s
are
0.
25
0,
0.7
6
2
an
d
0.
66
9,
r
es
pe
c
t
i
v
el
y
wh
i
c
h
t
he
p
-
v
a
l
ue
s
are
hi
gh
er
tha
n
t
he
al
p
ha
v
al
u
e
(
0.0
5).
T
he
p
-
v
al
u
es
of
t
h
e
RM
S
E
f
or
th
e
pe
r
i
od
s
of
the
m
oti
on
12
s
,
8s
,
an
d
6s
are
0.7
63
5
,
0.3
0
1
an
d
0
.84
7,
r
es
pe
c
t
i
v
el
y
i
n
w
h
i
c
h
t
he
p
-
v
a
l
ue
s
s
h
o
w
h
i
g
he
r
v
al
ue
s
t
ha
n
th
e
al
ph
a
(
0.0
5)
s
ug
ge
s
t
th
at
the
r
e
are
no
s
i
gn
i
f
i
c
an
t
di
f
f
erenc
es
be
t
we
en
t
he
pe
r
i
od
s
of
the
m
oti
o
n.
F
i
g
ure
9
s
ho
w
s
the
ef
f
ec
t
of
the
order
of
the
l
o
w
pa
s
s
B
u
tte
r
w
ort
h
f
i
l
ter
to
the
R
MS
E
v
a
l
ue
.
I
n
thi
s
ex
pe
r
i
m
en
t,
the
l
o
w
p
as
s
B
utt
er
w
orth
f
i
l
t
er
was
v
ar
i
ed
f
r
o
m
the
f
i
r
s
t
to
s
i
x
th
-
order
to
ob
tai
n
the
l
o
w
es
t
RM
S
E
. T
he
r
es
p
on
s
e o
f
th
e
RM
S
E
to
t
he
f
i
l
t
er or
de
r
i
s
no
t
a l
i
n
ea
r
f
un
c
ti
on
.
T
he
c
om
pa
r
i
s
on
be
t
ween
t
he
pred
i
c
ted
an
d
m
ea
s
ured
a
ng
l
e
f
or
d
i
f
f
erent
f
i
l
ter
o
r
de
r
i
s
s
ho
w
n
i
n
F
i
gu
r
e
1
0.
W
he
n
the
f
i
l
ter
order
i
s
to
o
l
o
w
,
i
t
af
f
ec
ts
the
ou
t
pu
t
of
th
e
e
x
pe
c
ted
an
gl
e
wi
th
a
r
i
pp
l
e
n
oi
s
e
(
F
i
gu
r
e
10
(
a)
)
.
W
he
n
the
f
i
l
ter
ord
er
i
s
too
h
i
gh
,
t
he
pre
di
c
te
d
an
gl
e
w
i
l
l
s
h
i
f
t
to
th
e
r
i
gh
t
(
F
i
gu
r
e
1
0
(
c
)
)
.
It
s
ho
w
s
tha
t
th
e
or
de
r
of
the
f
i
l
ter
af
f
ec
ts
the
p
erf
or
m
an
c
e
o
f
the
pred
i
c
ti
o
n
be
c
a
us
e
of
the
l
a
g
t
i
m
e
of
the
f
i
l
t
er.
S
i
nc
e
a
h
i
g
he
r
-
order
of
th
e
f
i
l
t
er
c
a
us
es
m
ore
de
l
a
y
t
i
m
e
to
proc
es
s
the
s
i
g
na
l
.
In
th
i
s
ex
pe
r
i
m
en
t,
w
e
f
ou
nd
th
at
th
e
l
o
we
s
t
RMS
E
i
s
i
n
the 3
rd
order
f
i
l
ter. T
he
z
ero
-
c
r
os
s
i
ng
f
ea
ture i
s
al
s
o
v
er
y
es
s
en
t
i
a
l
i
n t
h
i
s
pro
po
s
ed
m
eth
od
.
A
t
hres
ho
l
d
v
ol
t
ag
e
of
t
he
z
ero
-
c
r
os
s
i
ng
f
ea
ture,
as
w
r
i
tt
en
i
n
(
1),
i
s
v
er
y
i
m
po
r
tan
t
to
de
term
i
ne
a
n
op
ti
m
u
m
predi
c
ted
an
gl
e.
A
s
s
h
o
w
n
i
n
F
i
gu
r
e
11
,
the
l
o
wes
t
R
MS
E
v
al
ue
i
s
6.1
4
o
when
the
thres
h
ol
d
v
al
ue
i
s
0.0
9
m
V
.
In
th
i
s
s
tud
y
,
the
ex
os
k
el
eto
n
ba
s
ed
on
el
ec
tr
om
y
og
r
a
ph
y
c
on
tr
ol
i
s
ab
l
e
to
ad
a
pt
f
or
di
f
f
erent
s
ub
j
ec
t
an
d
v
aria
b
l
e
b
y
a
dj
us
ti
ng
t
he
f
ea
t
ure
thres
ho
l
d
an
d
ga
i
n
i
n
t
he
bi
o
-
am
pl
i
f
i
er.
L
en
z
i
r
e
po
r
te
d
s
i
m
i
l
ar
r
es
u
l
t
s
i
n
t
he
de
v
e
l
op
m
en
t
of
th
e
ex
os
k
el
eto
n
ba
s
ed
on
m
y
oe
l
ec
tr
i
c
c
on
tr
ol
,
w
h
i
c
h
the
f
i
nd
i
n
g
s
ho
w
s
tha
t
i
t
was
ea
s
y
to
i
m
pl
em
e
nt
an
d
d
i
d
n
ot
ne
ed
a
c
al
i
brat
i
o
n
[
8]
.
T
hi
s
i
s
c
on
tr
ar
y
whe
n
t
he
s
y
s
tem
us
es
a
pa
tt
ern
r
ec
o
gn
i
t
i
o
n
m
eth
od
whi
c
h
r
e
qu
i
r
es
a
tr
a
i
n
i
ng
s
tep
s
o
th
a
t
t
he
s
y
s
t
em
c
an
a
da
pt
to
t
he
pred
i
c
ti
o
n
al
g
orit
hm
f
o
r
di
f
f
erent s
ub
j
ec
ts
, a
s
m
en
ti
on
ed
b
y
pre
v
i
ou
s
r
es
ea
r
c
h
ers
[7,
26
, 2
9
]
.
T
ab
l
e 1
. T
he
S
um
m
ar
y
of
t
he
Cor
r
e
l
at
i
on
Co
ef
f
i
c
i
en
t
a
nd
RM
S
E
of
the
P
r
ed
i
c
ted
A
n
gl
e
f
or the
T
hree
P
erio
ds
of
th
e
Mo
t
i
o
n
12
s
ec
,
8
s
ec
,
a
nd
6
s
ec
P
e
r
iod
s
(
s
)
C
o
r
r
e
lat
ion
c
o
e
f
f
i
c
ien
t
R
M
S
E
(
°)
12
8
6
12
8
6
M
e
a
n
0
.
9
6
8
0
.
9
7
9
0
.
9
8
5
1
4
.
0
5
3
1
1
.
9
5
1
0
.
3
0
SD
0
.
0
1
2
0
.
0
0
6
0
.
0
0
3
1
.
6
6
3
2
.
3
1
4
1
.
4
5
3
M
e
d
ian
0
.
9
7
0
0
.
9
8
1
0
.
9
8
6
1
4
.
3
5
3
1
2
.
2
5
9
.
9
4
5
M
a
x
im
u
m
0
.
9
8
3
0
.
9
8
7
0
.
9
9
0
1
6
.
4
9
0
1
5
.
1
1
1
3
.
2
0
M
inim
u
m
0
.
9
4
0
0
.
9
6
9
0
.
9
7
9
1
1
.
2
3
2
9
.
0
4
5
8
.
1
8
6
p
-
v
a
lue
0
.
2
5
0
0
.
7
6
2
0
.
6
6
9
0
.
7
6
3
5
0
.
3
0
1
0
.
8
4
7
a
lph
a
0
.
0
5
0
.
0
5
0
.
0
5
0
.
0
5
0
.
0
5
0
.
0
5
F
i
gu
r
e
9.
T
he
ord
er of
th
e
d
i
gi
tal
f
i
l
ter
ef
f
ec
ts
th
e RM
S
E
of
th
e
ac
c
urac
y
of
the
pre
di
c
t
i
on
,
RM
S
E
i
n
de
grees
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
E
mb
ed
d
ed
s
y
s
tem
for up
pe
r
-
l
i
mb
ex
os
k
el
e
ton
.
.. (
T
r
i
wi
y
an
to
T
r
i
wi
y
a
nto
)
3001
F
i
gu
r
e
10
. T
he
or
de
r
of
th
e
f
i
l
ter ef
f
ec
ts
th
e p
erf
or
m
an
c
e o
f
th
e
predi
c
t
i
o
n
(
a) l
o
w
pa
s
s
B
ut
ter
w
ort
h f
i
l
t
er 1
st
ord
er, (
b) 3
rd
order
a
n
d (c
)
6
th
order
;
t
he
r
ed
l
i
ne
i
s
th
e p
r
e
di
c
te
d
an
g
l
e,
an
d
th
e
b
l
ac
k
l
i
ne
i
s
th
e m
ea
s
ured a
n
gl
e
F
i
gu
r
e
11
. T
he
ef
f
ec
t o
f
th
e
th
r
es
ho
l
d
v
o
l
ta
ge
i
n t
he
z
e
r
o
-
c
r
o
s
s
i
ng
f
ea
ture to
th
e
R
MS
E
-
v
a
l
u
e
4.
Co
n
clusion
In
th
i
s
s
tu
d
y
,
a
r
ea
l
-
ti
m
e
predi
c
t
i
on
o
n
t
he
el
bo
w
j
oi
nt
an
gl
es
ba
s
e
d
o
n
m
y
oe
l
ec
tr
i
c
c
on
tr
ol
i
s
i
m
pl
em
en
ted
i
n
an
em
be
dd
ed
s
y
s
tem
us
i
n
g
a
n
A
RM
S
T
M3
2F4
29
m
i
c
r
oc
on
tr
ol
l
er
to
c
on
tr
ol
th
e
ex
os
k
el
eto
n.
In
ac
c
ordanc
e
wi
th
the
ex
p
e
r
i
m
en
t,
a
b
i
c
ep
s
c
an
be
u
s
ed
to
c
on
tr
ol
the
ex
os
k
el
eto
n
s
m
oo
thl
y
ac
c
ordi
ng
t
o
the
el
bo
w
j
oi
n
t
an
g
l
e.
T
he
pred
i
c
ti
o
n
a
l
g
orit
hm
c
on
s
i
s
t
s
of
the
z
ero
-
c
r
os
s
i
ng
f
ea
t
ure
ex
tr
ac
ti
o
n
an
d
t
he
s
ec
on
d
-
ord
er
c
on
s
i
s
ts
of
the
l
o
w
pa
s
s
B
utt
er
w
orth
f
i
l
ter
w
h
i
c
h
p
l
a
y
s
as
the
s
m
oo
the
ni
n
g
f
un
c
ti
on
.
T
he
n
ov
el
t
y
of
the
pr
op
os
ed
m
eth
od
he
r
e
i
s
a
s
i
m
pl
e
a
l
go
r
i
th
m
to
predi
c
t
the
el
bo
w
j
o
i
nt
an
gl
e
us
i
ng
t
he
z
ero
-
c
r
os
s
i
ng
f
ea
tur
e
ex
tr
ac
ti
on
an
d
the
l
o
w
p
as
s
B
utt
er
wor
th
f
i
l
t
er.
T
hi
s
al
go
r
i
thm
i
s
c
ate
g
ori
z
ed
as
a
no
n
-
pa
tt
ern
r
ec
og
ni
t
i
o
n
m
eth
od
.
T
he
r
ef
ore,
w
e
d
o
no
t
n
ee
d
a
l
ea
r
ni
ng
s
tag
e
to
tr
a
i
n
th
e
s
y
s
tem
s
i
nc
e
the
E
MG
s
i
gn
a
l
s
are
ex
te
ns
i
v
el
y
r
ec
og
n
i
z
ed
as
a
n
on
-
s
tat
i
on
ar
y
s
i
gn
al
a
nd
r
an
do
m
w
h
i
c
h
the
c
ha
r
ac
teri
s
ti
c
s
d
ep
e
nd
on
th
e
s
ub
j
ec
ts
.
T
hu
s
,
th
e
propos
e
d
m
eth
od
c
an
be
a
da
pt
ed
s
i
m
pl
y
b
y
ad
j
us
ti
n
g t
h
e t
hr
es
ho
l
d
v
ol
ta
ge
i
n t
he
z
ero
-
c
r
os
s
i
ng
f
ea
ture an
d t
h
e a
m
pl
i
f
i
er g
a
i
n.
In
the
ne
x
t
w
ork
,
an
au
to
-
tun
i
n
g
of
the
thres
ho
l
d
v
ol
t
ag
e
i
n
the
z
ero
-
c
r
os
s
i
ng
al
g
orit
hm
wi
l
l
b
e
de
v
e
l
o
pe
d
s
o
th
at
th
e
s
y
s
tem
c
an
ad
ap
t
au
tom
ati
c
al
l
y
when
the
s
ub
j
ec
t
i
s
c
ha
ng
e
d.
T
hi
s
w
ork
c
an
al
s
o
be
i
m
pl
em
en
ted
f
or
oth
er
s
tud
i
es
i
n
s
i
m
i
l
ar
area
s
s
uc
h
as
f
or
pros
the
t
i
c
de
v
i
c
es
an
d
ex
o
s
k
el
eto
n
-
ba
s
ed
m
y
o
el
ec
tr
i
c
c
on
tr
ol
f
or
r
eh
ab
i
l
i
tat
i
on
p
urpos
e.
In
t
hi
s
s
tud
y
,
a
r
ea
l
-
ti
m
e
predi
c
t
i
on
on
th
e
e
l
bo
w
j
oi
nt
a
ng
l
es
ba
s
ed
on
m
y
o
el
ec
t
r
i
c
c
on
tr
ol
i
s
i
m
pl
em
en
ted
i
n
an
em
be
dd
ed
s
y
s
tem
us
i
ng
a
n
A
RMS
T
M3
2F4
29
m
i
c
r
oc
on
tr
ol
l
er
to
c
o
ntrol
the
ex
os
k
el
eto
n
.
In
ac
c
ord
an
c
e
w
i
t
h
the
ex
pe
r
i
m
en
t,
a
bi
c
ep
s
a
l
on
e
c
an
be
u
s
ed
to
c
on
tr
o
l
the
ex
os
k
el
eto
n
s
m
oo
thl
y
ac
c
ordi
ng
t
o
the
el
bo
w
j
oi
n
t
an
g
l
e.
T
he
pred
i
c
ti
o
n
a
l
g
orit
hm
c
on
s
i
s
t
s
of
the
z
ero
-
c
r
os
s
i
ng
f
ea
t
ure
ex
tr
ac
ti
o
n
an
d
t
he
s
ec
on
d
-
ord
er
c
on
s
i
s
ts
of
the
l
o
w
pa
s
s
B
utt
er
w
orth f
i
l
ter
w
h
i
c
h p
l
a
y
s
as
th
e s
m
oo
th
en
i
ng
f
un
c
t
i
on
.
Ref
er
en
ce
[1
]
O
s
k
o
e
i
M
A,
Hu
H
.
M
y
o
e
l
e
c
tri
c
c
o
n
tro
l
s
y
s
te
m
s
-
A
s
u
rv
e
y
.
Bi
o
m
e
d
.
Si
g
n
a
l
Pr
o
c
e
s
s
.
C
o
n
tro
l
.
2
0
0
7
;
2
(4
)
:
2
7
5
–
2
9
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.