Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
9
, No
.
5
,
Octo
ber
201
9
, pp.
4099
~
41
13
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v9
i
5
.
pp4099
-
41
13
4099
Journ
al h
om
e
page
:
http:
//
ia
es
core
.c
om/
journa
ls
/i
ndex.
ph
p/IJECE
Dynami
c anal
ysis and
QF
T
-
bas
ed robust
contr
ol d
esig
n
of
a
coaxial m
icro
-
he
licopt
er
Aissa
Me
ksi,
Ah
med
H
ami
da
Bo
udin
ar,
B
enatma
n
K
oua
dri
El
e
ct
ri
ca
l
Eng
in
ee
ring
Depa
r
tment,
Fa
cul
t
y
of El
ec
tr
ic
a
l Engi
ne
er
ing,
Univer
sit
y
of
Sc
ie
nc
es
and
Tech
nolog
y
of
Or
an,
A
lg
eri
a
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Ma
r
9
, 201
9
Re
vised
A
pr1
9
, 201
9
Accepte
d
Apr
25
, 2
01
9
Thi
s
pape
r
pre
s
ent
s
the
d
y
namic
beha
v
ior
of
a
coa
xi
al
m
ic
ro
-
hel
i
copt
e
r
,
under
Quanti
t
ati
ve
Feedback
Th
eor
y
(QF
T)
con
t
rol
.
Th
e
fli
gh
t
dy
nami
cs
of
aut
onom
ous
ai
r
vehi
cles
(AA
Vs
)
with
rota
ting
rings
is
no
n
-
li
ne
ar
and
complex.
The
n
,
it
bec
om
es
nec
e
ss
ar
y
to
cha
r
ac
t
e
riz
e
th
ese
non
-
linear
i
ti
es
for
ea
ch
f
li
ght
conf
i
gura
ti
on
,
in
ord
er
to
provid
e
th
ese
aut
onom
ous
ai
r
veh
icle
s
(AA
Vs
)
with
aut
onom
ous
fli
ght
a
nd
navi
gati
on
ca
pabi
l
it
i
es.
The
n,
the
nonli
ne
ar
m
ode
l
is
l
ineari
z
ed
aro
und
the
o
per
ating
poin
t
using
som
e
assum
pti
ons.
Fi
nal
l
y
,
a
robust
QF
T
cont
rol
law
over
the
coaxia
l
m
ic
ro
-
hel
i
copt
er
is
app
li
ed
to
m
eet
some
spec
ifica
ti
ons.
QF
T
(qua
nti
tativ
e
fee
db
ack
th
eor
y
)
is
a
control
l
aw
design
m
et
hod
that
use
s
fre
quency
dom
ai
n
concept
s
to
m
ee
t
p
erf
orm
anc
e
spe
ci
f
ic
a
ti
o
ns
while
m
ana
gi
ng
unce
rt
ai
nt
y
.
Thi
s
m
et
hod
is
base
d
on
th
e
fee
db
ac
k
con
trol
when
th
e
pla
nt
is
unce
r
tain
or
when
unce
rt
ai
n
distur
banc
es
are
aff
e
ct
ing
th
e
pl
ant.
The
QF
T
design
appr
oac
h
invol
ves
conv
en
ti
onal
fre
qu
ency
response
loop
s
hapi
ng
b
y
m
anipulat
ing
th
e
gai
n
v
ari
ab
le
wi
th
the
pole
s
and
ze
ros
of
the
no
m
ina
l
tra
nsfer
func
ti
on
.
Th
e
design
proc
ess i
s
accom
pli
shed
b
y
using MATL
A
B
envi
ronm
ent software
.
Ke
yw
or
d
s
:
Coaxial
m
ic
ro
-
helic
op
te
r
MATLAB
t
oolbo
c
k
Non
-
li
near
Qu
a
ntit
at
ive
f
e
edb
ac
k
t
he
ory
Copyright
©
201
9
Instit
ute of
Ad
v
ance
d
Engi
ne
eri
ng
and
Sc
ie
n
ce
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
Aissa
Me
ksi
,
Ele
ct
rical
En
gi
neer
i
ng D
e
par
t
m
ent, F
acul
ty
of Elec
tric
al
E
ng
i
neer
i
ng,
Un
i
ver
sit
y o
f S
ci
ences a
nd Te
chnolo
gy
of Or
an
,
B.P 15
05, El
Mnao
uer,
Or
a
n, A
l
ger
ia
.
Em
a
il
:
ai
ssa.
m
eksi@
un
i
v
-
us
t
o.dz
1.
INTROD
U
CTION
Un
m
ann
e
d
aer
ia
l
veh
ic
le
s
(
U
AV
s
)
a
re
ai
rc
r
aft
capa
ble
of
fly
ing
a
nd
pe
rfor
m
ing
a
m
issio
n
with
o
ut
hu
m
an
prese
nc
e
on
bo
a
rd.
I
ni
ti
al
l
y
dev
el
ope
d
as
pa
rt
of
m
i
li
ta
ry
act
ivit
ie
s
,
ther
e
is
now
a
gr
e
at
pote
nti
al
for
ci
vil
act
ivit
ie
s
(survei
ll
ance,
carto
gr
a
ph
y.
..).
Ma
ny
unres
ol
ved
co
ns
trai
nts
rem
ai
n
fo
r
t
he
us
e
of
ci
vil
dron
e
s
in
the
pu
blic
sp
ace.
Am
on
g
t
h
e
ha
r
d
points
to
res
olv
e
a
re
e
m
bed
de
d
deci
sion
-
m
aking
a
utono
m
y,
al
l
-
weathe
r
per
ce
ptio
n
cap
abili
ty
,
safety
and
dep
e
ndabil
it
y.
These
featur
es
will
,
in
the
fu
tu
re,
be
s
ubj
ect
to
a
certi
f
ic
at
ion
process
b
ei
ng
dev
el
op
e
d, but
current
dro
nes suffer
fro
m
a lack
of
r
obus
t
ne
ss and a
uton
om
y.
In
the
li
te
ratu
r
e,
m
od
el
ing
an
d
de
velo
pm
ent
of
new
form
ulati
on
s
of
coa
xi
al
m
ic
ro
-
helic
op
te
rs
ha
ve
been
t
he
sub
j
e
ct
of
seve
ral
r
esearch
w
orks
su
ch
as:
Mo
ham
m
ad
Har
un
work
s
[1
]
,
a
nd
Che
n
[
2]
on
cl
assic
helic
op
te
rs,
Alvaro
[
3]
for
th
e
m
inidr
one
w
it
h
a
keeled
pro
pelle
r,
as
w
el
l
as
the
wo
r
ks
of
Sc
hafro
t
h
a
nd
Christi
an [4]
on c
oax
ia
l m
ic
ro
-
helic
opte
r (w
it
h
fixe
d
a
nd
va
riable ste
ps
).
Am
on
g
these
m
ilit
ary
dr
on
e
s
a
m
ic
ro
coaxi
al
helic
op
te
r
presente
d
in
this
arti
cl
e
was
dev
el
oped
by
par
t
ner
s
I
SL,
CR
AN
an
d
H
EUDI
ASYC
Syst
e
m
s.Th
e
m
or
e
detai
le
d
m
od
el
s
of
so
m
e
con
fi
gurati
on
s
of
the
coax
ia
l
m
ic
ro
-
helic
op
te
r
ca
n
be
fou
nd
in
[5
,
6].T
he
c
oax
ia
l
m
ic
ro
-
helic
opte
r
with
swas
hpla
te
co
ns
ist
s
of
t
w
o
counter
-
r
otati
ng
r
ot
or
s
e
nsuri
ng
the
li
ft
a
nd
pro
pu
lsi
on
o
f
t
he
ai
rc
raf
t
a
nd
a
swa
shpla
te
m
ou
nted
at
the
lowe
r
ro
t
or
t
o
co
ntr
ol
the
cra
ft.
T
he
two
c
ounte
r
-
ro
ta
ti
ng
r
otors
sh
a
re
the
s
am
e
axis
of
ro
ta
t
ion
,
w
hich
m
akes
it
po
s
sible
to
c
om
pen
sat
e
m
utu
al
ly
fo
r
the
rea
ct
ion
to
rque
produce
d
by
eac
h
of
the
r
oto
r
s,
and
co
n
se
qu
e
ntly
the
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
5
,
Oct
ober
20
19
:
4099
-
4
1
1
3
4100
ste
ering
of
the
la
ce.
This
syst
e
m
is
al
so
a
com
pact
anti
-
t
orq
ue
so
luti
on,
wh
ic
h
sa
ves
t
he
au
xili
ary
syst
e
m
pr
ese
nt
on
a
c
onve
ntio
nal
helic
op
te
r
a
nd
the
r
efore
reduce
t
he
le
ng
t
h
of
the
fu
sel
a
ge.
The
m
ass
of
the
dr
on
e
is
li
fted
by
t
he
t
ot
al
thru
st
pro
duced
by
the
t
w
o
ro
t
or
s
.
A
ro
t
ary
wi
ng
dr
one
can
be
co
ns
i
der
e
d
as
a
rigi
d
body
with si
x degree
s of free
do
m
f
or whic
h
the
aer
od
y
nam
ic
f
or
c
es g
e
ner
at
e
d b
y i
ts act
uator
s
are a
dd
e
d
[7]
.
The
pro
blem
of
this
coa
xial
m
ic
ro
-
helic
op
t
er
is
non
-
li
nea
r,
com
plex,
qu
asi
sta
ti
on
naire
and
s
ubj
ect
to
wi
nd
disturbance
,
a
ff
ect
in
g
it
s
sta
ble
fli
gh
t
an
d
ea
sy
la
nd
i
ng
a
nd
the
dyna
m
ic
equ
at
ion
s
of
m
otion
for
a
coax
ia
l
m
ic
ro
-
helic
op
te
r
incl
ud
i
ng
str
uctured
a
nd
unstr
uc
ture
d
unce
rtai
nties.
The
ap
plica
ti
on
of
t
he
QFT
te
chn
iq
ue
f
or
t
raj
ect
or
y
trac
ki
ng
i
n
the
pres
ence
of
uncert
ai
nties
of
m
as
s
va
riat
ion
pa
r
a
m
et
ers
is
stu
died.I
n
this
pa
per,
t
he
QFT
Horowit
z
I.
M
m
et
hod
as
an
ap
pro
pr
i
at
e
rob
us
t
c
on
t
ro
l
m
et
ho
d
is
app
li
ed
to
a
c
oax
ia
l
m
ic
ro
-
helic
op
t
er, w
hich has
m
any adv
a
nta
ge
s over ot
he
r
re
le
van
t t
ec
hn
i
ques
[8
]
.
The
Q
FT
H
oro
witz
I.
M
desi
gn
proce
dure
is
no
t
on
ly
base
d
on
the
am
plit
ud
e
of
the
tra
ns
f
er
f
un
ct
io
n
in
the
fr
e
quen
cy
do
m
ai
n,
but
al
so
ta
kes
into
acco
un
t
the
ph
ase
in
f
or
m
at
ion
[
8].Th
e
qu
a
ntit
at
ive
fe
edb
a
c
k
theo
ry
QFT,
c
al
le
d
(La
t
hé
or
ie
qu
a
ntit
at
ive
de
la
ret
ro
act
i
on
TQR
)
in
Fre
nch
H
oro
witz
I.
M,
was
de
ve
lop
e
d
by
Isaac
Horowit
z.
It
is
a
frequ
e
ncy
te
ch
ni
qu
e
us
in
g
the
Nich
ols
diag
r
a
m
to
reali
ze
a
robust
co
ntr
ol
[9
]
.
The
desi
red
te
m
po
ral
responses
are
translat
ed
into
tolera
nc
es
in
the
fr
e
quency
dom
ai
n.
The
desig
n
pr
ocess
is
highly
trans
pa
ren
t;
al
lowi
ng
a
desig
ne
r
to
see
wh
at
tr
adeoff
is
nee
ded
t
o
ac
hiev
e
a
desire
d
le
vel
of
perform
ance.
In
this
m
anu
s
cript,
we
a
ddr
ess
the
three
m
ai
n
pr
oble
m
s
of
the
a
utom
at
ic
app
li
ed
to
the
coa
xia
l
helic
op
te
r
m
icr
o
-
ty
pe
dro
ne,
since
we
a
re
interest
ed
in
de
te
rm
ining
a
dy
nam
ic
m
od
el
of
the
syst
em
,
a
nd
aerona
utica
l
m
od
el
, as wel
l as cal
culat
ing
a s
ta
bili
zi
ng
input
. W
e are m
or
e
sp
eci
fical
ly
intereste
d
in ac
hi
evin
g
the ob
j
ect
ives t
hu
s
define
d:
-
Mod
el
in
g
the
transla
ti
on
an
d
ro
ta
ti
on
dy
na
m
ic
s
of
the
c
oa
xial
m
ic
ro
-
he
li
cop
te
r
dur
in
g
it
s
autonom
ou
s
fligh
t
phase
[10]
.
-
Esta
blish
the
li
near
m
od
el
of
the
nonli
near
m
od
el
by
app
l
yi
ng
the
Tay
l
or
ap
pro
xim
ation
m
et
ho
d
ar
ou
nd
the ope
rati
ng point.
-
Determ
ine
a
st
abili
zi
ng
co
rr
e
ct
or
f
or
c
oa
xial
m
ic
ro
-
helic
op
te
r
in
the
case of
q
ua
si
-
sta
ti
onary
flig
ht,
f
r
om
a
su
f
f
ic
ie
ntly
sim
ple synthesis m
od
el
[11]
.
T
he
rem
ai
nd
er
of
t
his
pa
pe
r
is
organ
iz
e
d
as
fo
ll
ow.
I
n
Sect
ion2,
the
dynam
ic
m
od
el
of
coa
xial
m
ic
ro
-
helic
op
t
er
is
f
or
m
ulated
in
the
Ca
rtes
ia
n
sp
ace
.
I
n
S
ect
ion
3,
Q
FT
con
t
ro
ll
er
is
de
velo
ped
a
nd
a
pp
li
ed
to
the
direct
dy
nam
ic
m
od
el
of
dro
ne
in
Ca
rtesi
an
sp
ace.
Se
ct
ion
4
pr
ese
nts
si
m
ulati
on
resu
lt
s
of
the
pr
opose
d
con
t
ro
ll
er
an
d discussi
on
of
r
esults. Fi
nally
, s
om
e con
cl
us
i
on
s
are
prese
nted
in
t
he
cl
os
in
g
sect
io
n.
2.
MA
T
HEM
AT
ICA
L
MODE
L OF
COA
X
I
AL
MICRO
-
HEL
ICO
PTE
R
Kno
wing
that
coa
xial
m
ic
ro
-
helic
opte
r
is
co
ns
ide
red
a
s
a
ri
gid
body
with
fi
xed
m
ass
"
m
",
the
ge
ner
ic
m
od
el
with
six
degrees
of
fr
ee
do
m
(6
-
DOF)
ref
e
rs
to
it
s
three
translat
ion
s
and
it
s
th
ree
sp
at
ia
l
ro
ta
ti
ons.
It
th
eref
or
e
descr
i
be
s
the
dynam
i
cs
and
t
he
ki
nem
at
ic
s
of
ro
ta
ti
on
an
d
tra
ns
l
at
ion
in
the
refe
ren
ce
syst
e
m
li
nk
ed
to
the
"G"
bo
dy
.
The
re
a
re
t
wo
ways
of
e
xpressi
ng
the
e
qu
at
io
ns
of
m
otion
:
ei
ther
f
r
om
the
la
ws
of
Ne
wton
an
d
E
uler
th
a
t
we
will
us
e
or
by
refo
rm
ul
at
ing
the
New
t
on
ia
n
m
echan
ic
s
with
the
La
gr
a
nge
and
Ham
il
ton
equ
at
io
ns.
I
n
t
his
sect
io
n
a
re
def
i
ned
the
refe
ren
ce
points
,
axes
syst
em
s
us
ed
a
nd
the
ki
nem
at
ic
and
ki
netic
pa
ram
et
ers
app
e
arin
g
in
Fig
ur
e
1
,
as
well
a
s
al
l
the
relat
ion
s
c
om
po
sin
g
th
e
6
-
DOF
m
od
el
.
It is ass
um
ed
that t
he
i
ner
ti
al
eff
ect
s
of the
r
otors a
re
neg
li
gib
le
c
om
par
ed
to
th
os
e
of th
e m
a
in bod
y.
Figure
1.
La
nd
m
ark
s
us
e
d
to
descr
i
be
the
m
ov
em
ent of the
coa
xial m
ic
ro
-
helic
op
te
r
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Dyna
mic an
alysis a
nd QFT
-
base
d
r
obus
t c
ontrol
desig
n of
…
(
AissaMeksi
)
4101
2.1.
Ref
ere
nce
an
d
axes s
ystems
The
m
ov
em
ent of a
rigid b
od
y i
s m
a
inly
d
escribe
d by tw
o
r
efere
nces:
-
The
ine
rtia
l
ref
ere
nce
{J}:
boun
d
to
the
E
arth,
hav
i
ng
a
ref
ere
nce
po
i
nt
"O"
an
d
de
fine
d
by
the
ba
s
e
{x
e
; y
e
; z
e
},
where {
x
e
} poi
nts nor
t
h,
{y
e
}
po
i
nts east, a
nd {z
h
} points
up.
-
The
na
vig
at
io
n
m
ark
(g
ea
r
or
m
ob
il
e)
{
B
}:
relat
ed
t
o
t
he
ce
nter
of
gravit
y
"G"
of
t
he
str
uctu
re
of
the
dro
ne
an
d
de
fi
ned
by
the
ba
se
{x
C
;
y
G
;
z
G
}
wh
e
re
{x
G
}
is
the
longit
ud
inal
axis
point
ing
to
wa
rd
s
t
he
veh
ic
le
f
r
ont,
{
y
G
}
def
in
es
t
he
la
te
ral
axis
an
d
po
i
nts
t
o
the
v
eh
ic
le
r
ig
ht
a
nd
{z
G
} d
efi
ne
s
the
ve
rtic
al
axis
of the
ve
hicle
an
d p
oin
ts
to
t
he
top.
A
r
otati
on
m
a
trix
ide
ntifie
s
the
ori
entat
io
n
of
a
ny
c
oor
din
at
e
syst
em
in
the
t
hr
ee
dim
ension
s
of
s
pace.
It
is
al
so
cal
le
d
transiti
on
m
at
rix
or
at
ti
tud
e
m
atr
ix,
no
te
d
"R
η".
Con
se
quentl
y,
the
passa
ge
m
at
rix
Rη
of
the
ref
e
r
ence
a
pparatu
s
"
B"
to
the
re
fe
ren
ce
ine
rtia
l
"
J"
is
pa
ram
et
erized
by
the
Eul
er
a
ng
le
s
desc
r
ibing
the
t
hr
ee
r
otati
on
a
ng
le
s
(
r
ol
l),
θ
(
pitch)
a
nd
(yaw
).
T
he
product
of
t
hese
three
m
at
rices
m
akes
it
possible
to exp
ress
t
he passa
ge fr
om
t
he refe
ren
ce
"
B"
to t
he
r
efe
re
nce "
J"
,
that is
to say [
12]
:
=
[
−
+
+
−
−
]
(1)
w
he
re
is
the
ro
ll
an
gle,
θ
t
he
pitch
a
ngle
and
the
ya
w
ang
le
,
res
pect
ive
ly
.Th
e
kin
e
m
at
ic
equ
at
ion
of
or
ie
ntati
on
c
onnects
t
he
tim
e
der
i
vative
of
the
an
gles
of
r
ol
l
∅
,
pitc
h
θ
a
nd
ya
w
to
the
instanta
neous
s
pee
d
of rota
ti
on
by:
=
[
1
0
−
0
]
(2)
2.2.
Balance
of
fo
r
ces a
nd
m
ome
nt
s
appl
ie
d
Eq
uations
of
f
or
ces
a
nd
m
ome
nts
are
req
ui
red
to
com
plete
the
dynam
ic
m
od
el
ing
of
ae
ro
m
echan
ic
al
syst
e
m
s.
More
ov
e
r,
t
he
m
od
e
li
ng
of
the
se
f
or
ces
a
nd
m
ome
nts
(that
is
to
say
,
F
res
a
nd
M
)
rem
ai
ns
a
m
ajor
pro
blem
,
du
e
to
the
com
plexity
of
fl
uid
dynam
ic
s
and
inte
racti
on
s
bet
we
en
the
stu
died
veh
ic
le
(r
igi
d
body
equ
i
pp
e
d
with
m
ov
ing
act
ua
tors)
a
nd
the
su
r
rou
nd
i
ng
fl
uid
.
T
he
m
ai
n
f
or
ces
act
in
g
on
a
r
otorcra
ft
are
pr
ese
nted
[
12]
.
2.2.1.
Eff
ort
set
ting
s
The
e
ffor
t
pa
r
a
m
et
ers
are
si
m
pl
y
the
thre
e
f
or
ces
an
d
t
he
th
ree
m
ome
nts
that
i
nter
ven
e
in
the
dynam
ic
s
of
translat
ion
a
nd
ro
ta
ti
on
of
the
veh
ic
le
.
The
dev
el
op
m
ent
of
these
effo
rts
will
con
sti
tute
the
aerodynam
ic
m
od
el
p
rese
nte
d
in
the
f
ollow
i
ng secti
ons
[7
,
12
]
.
We
note
th
en:
=
[
]
,
the t
hr
ee
co
m
pone
nts
of
t
he fo
rce
vecto
r
e
xpress
ed
in
"
B
",
M
=
[
]
,
the t
hr
ee
co
m
pone
nts
of
t
he m
o
m
ent v
ect
or
expres
sed
in
"
B
".
a
nd
=
[
]
(3)
Figure
2.
T
he f
or
ces
and m
ome
nts
on the syst
e
m
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
5
,
Oct
ober
20
19
:
4099
-
4
1
1
3
4102
2.2.1.
1.
The f
orces
The fo
rces
w
hich
a
re acti
ng
on the
syst
em
are:
-
T
he wei
ght o
f
the
co
a
xial m
ic
ro
-
helic
opte
r
The
ve
hicle
is
su
bject
e
d
to
the
gr
avita
ti
onal
fiel
d
as
al
l
bodies
in
the
vicinit
y
of
the
Ear
th,
an
d
the
acce
le
rati
on
of
gravit
y
is
de
note
d
"
g".
Th
e
weig
ht
com
ponen
t
"
P"
is
the
force
to
w
hich
the
cra
ft
is
s
ubj
ect
ed,
and
whose
di
r
ect
ion
is
norm
al
to
the
su
rf
a
ce
of
the
Eart
h.
The
val
ue
of
"g"
is
exp
ress
ed
an
d
know
n
in
the
iner
ti
al
r
efe
re
nc
e fr
am
e "
B"
, t
hat is to
say [
12]
:
=
g
z
e
(4)
-
T
he pushi
ng for
ce
The
m
od
el
of
thr
us
t
of
a
ro
ta
r
y
wing
can
be
expresse
d
as
a
functi
on
of
th
e
sp
ee
d
of
r
ota
ti
on
of
t
he
ro
t
or
squa
red
Ω
p
2
at
a
ce
rtai
n
ae
r
od
y
nam
ic
coeffic
ie
nt
"
k"
.
T
he
li
ft
gen
e
rated
by
a
ro
ta
ti
ng
r
otor
t
her
e
fore
has
t
he
f
ollo
wing e
xpressi
on
:
=
Ω
2
(5)
-
T
he dra
g forc
e
The
aer
od
y
na
m
ic
dr
ag
force
of
the
ve
hicle
F
a
caused
by
the
fr
ic
ti
on
of
t
he
ai
r
on
the
f
us
el
age
ca
n
be
expresse
d
as:
[
13
]
F
a
=
1
Ω
2
(6)
Ω
: M
oto
r
rotat
io
n
s
pee
d.
2.2.2.
Forces
gener
ate
d by
th
e
coa
xia
l mi
cr
o
-
h
el
ic
op
ter
Thrust
is
the
m
ai
n
fo
rce
pro
du
ce
d
by
the
coax
ia
l
m
ic
ro
-
he
li
cop
te
r,
al
lo
wing
the
dr
one
to
m
ov
e
i
n
three
-
dim
ension
al
sp
ace.
T
he
i
m
m
ob
il
e
up
pe
r
ro
t
or
pro
duce
s
on
ly
a
ver
ti
cal
thru
st
force
F
1
,
wh
ic
h
is
directl
y
pro
portion
al
to
the
s
quare
of
the
s
peed
of
r
ot
at
ion
of
it
s
bl
ades
Ω
1
an
d
the
a
erodynam
ic
coef
fici
ent
of
th
r
us
t
,
suc
h t
hat:
1
=
[
0
0
Ω
1
2
]
(7)
The
ex
pressi
on
of
the
f
orce
2
gen
e
rated
by
the
lowe
r
r
otor
de
pends
on
t
he
tw
o
an
gles
δ
cx
and
δ
cy
of
t
he
swas
hp
la
te
wh
i
ch fo
rm
the inputs
of
t
he
syst
e
m
. Th
is force i
s g
i
ven b
y:
2
=
Ω
2
2
[
−
−
]
(8)
Figure
3.Sum
m
ary of
t
he
el
e
m
ents r
el
at
ed
t
o
the
m
od
el
ing o
f
the
co
a
xial
m
ic
ro
-
helic
op
t
er
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Dyna
mic an
alysis a
nd QFT
-
base
d
r
obus
t c
ontrol
desig
n of
…
(
AissaMeksi
)
4103
The
t
otal
th
rus
t,
=
[
0
0
]
,
i
s
def
i
ned
by
the
s
um
of
the
in
div
id
ual
thr
us
ts
of
eac
h
of
the
ro
t
or
s
[
1
]
and
[
2
]
de
scribe
d by
(
5)
and (
6). T
he
e
xpressi
on
of
t
he t
otal t
hrust ca
n be
w
ritt
en:
=
[
Ω
1
2
+
Ω
2
2
]
(9)
Finall
y, the t
otal force F
g
e
ne
rated
by the
tw
o ro
t
or
s
can
b
e
w
ritt
en:
=
[
]
=
[
−
Ω
2
2
−
Ω
2
2
Ω
1
2
+
Ω
2
2
]
(10)
2.2.3.
Moment
s
act
i
ng
on the
c
oaxi
al micr
o
-
heli
copter
The
m
o
m
ent
M
p
gen
erate
d
by
the
total
fo
rc
e
F
app
li
ed
to
t
he
coa
xial
m
icr
o
-
helic
opte
r
a
t
it
s
center
of
gr
a
vity
G
is
th
eref
or
e
t
he
sum
of
the
m
o
m
e
nts
res
ulti
ng
from
the
fo
rce
of
the
upper
r
oto
r
F
1
,
that
of
th
e
lowe
r
ro
t
or
F
2
a
nd th
e w
ei
ght P
.
M
=
⃗
⃗
⃗
⃗
⃗
1
∧
1
+
⃗
⃗
⃗
⃗
⃗
2
+
∧
2
+
⃗
⃗
⃗
⃗
⃗
∧
(11)
As
a
co
nclusi
on, t
he
s
um
o
f
t
he
m
o
m
ents o
f
the
forces
F is
giv
e
n by:
M
=
[
−
Ω
2
2
0
]
(12)
Howe
ver,
a
m
om
ent
aro
un
d
t
he
z
a
xis
due
t
o
eac
h
of
the
r
otors
is
di
rectl
y
propor
ti
onal
to
the
s
qu
a
re
of
their
velocit
y wit
h t
he
ae
rodynam
i
c coe
ff
ic
ie
nts
1
>
0
,
2
>
0
,
su
c
h
as:
=
[
0
0
1
Ω
1
2
−
2
Ω
2
2
]
(13)
Finall
y, the t
otal
m
o
m
ent
M
p
is
giv
e
n by:
M
=
[
]
=
[
−
Ω
2
2
Ω
2
2
1
Ω
1
2
−
2
Ω
2
2
]
(14)
Accor
ding to
(2.
10)
a
nd (2.
14
),
t
he
e
xpr
essio
n of t
he
t
otal f
or
ce
F
ca
n be
wr
it
te
n
as:
=
(
0
,
0
,
)
+
∑
wh
e
re
∑
=
1
(
)
(15)
2.2.4.
Dev
el
op
me
nt
of t
he
m
at
he
mat
ic
al m
od
el
a
cc
ordin
g to
Newton
-
Eul
er
Af
te
r
prese
nting
t
he
diff
e
re
nt
eq
uatio
ns
we
ca
n
no
w
de
velo
p
the
m
od
el
m
at
he
m
a
tical
us
in
g
th
e
New
t
on
-
E
uler fo
rm
ulati
on
. T
he
c
orres
ponding eq
uatio
ns
a
r
e wri
tt
en
in t
h
e
foll
ow
i
ng form
[
7
,
12]
:
̇
=
̇
=
+
+
̇
=
Ω
Ω
̇
=
−
Ω
×
Ω
+
+
(16)
w
he
re:
:
Is
the
posit
io
n vecto
r of
t
he c
oax
i
al
m
ic
ro
-
helic
op
te
r
=
[
,
,
]
,
m:
the total
m
a
ss of the
c
oax
i
al
m
ic
ro
-
he
li
co
pter,
J:
sym
m
et
ric m
at
rix
inerti
a
of d
im
ension
(
3x3),
giv
e
n by:
=
[
0
0
0
0
0
0
]
,
(17)
Ω
:
The
ang
ular v
el
ocity
ex
press
ed
in
the
fixe
d i
ner
ti
al
r
efe
re
nc
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
5
,
Oct
ober
20
19
:
4099
-
4
1
1
3
4104
2.2.5.
Dyna
mi
c
simu
lation
of
th
e
s
ys
te
m in
M
A
TL
AB
-
Simul
i
nk
en
viro
nme
nt
MATLAB
Si
m
ul
ink
s
of
t
wa
re
is
us
ed
to
si
m
ulate
the
pro
pu
lsi
on
s
yst
e
m
.
In
Fi
gure
4,
asi
t
is
m
entioned,
a
s
ect
ion
of
bo
dy
m
od
el
ing
,
cal
l
ed
rigid
bo
dy,
is
show
nin
Si
m
ul
ink
e
nv
i
ronm
ent.
In
this
syst
e
m
,
inputs,
f
or
ces
and
th
e
total
dr
ag
f
orce
are
e
xer
te
d
on
the
center
of
m
ass
of
the
ve
hicle
.
Total
dr
a
g
f
orces,
befor
e
enteri
ng
the
body
dy
nam
ic
cal
culation
s
syst
em
,
wh
ic
h
are
w
ritt
en
base
d
on
New
t
on
Eule
r’s
la
w,
are
tra
ns
fe
rred
from
body
c
oor
d
inate
syst
e
m
to
inerti
a
c
oor
din
at
e
syst
e
m
.
In
this
syst
e
m
,
pro
pu
lsi
on
syst
e
m
and
t
he
tra
ns
it
ion
of
f
or
ces
a
nd
m
o
m
ents
fr
om
bo
dy
c
oord
i
nate
syst
e
m
to
inerti
a
co
ordin
at
e
syst
e
m
are
add
e
d
to
the
rigi
d
body
sub
-
syst
e
m
.
The
physi
cal
pr
ope
rtie
of
the
c
oa
xial
m
ic
r
o
-
hélic
o
pter
are
c
ons
idere
d
as in Ta
ble
1
.
Figure
4
.
N
on
li
near sy
ste
m
o
f
the co
a
xial m
icr
o
-
helic
opte
r
Table
1
.
Pa
ram
et
ers
m
od
el
of
coax
ia
l m
ic
ro
-
helic
op
te
r
Para
m
eter
Descripti
o
n
Valu
e
Un
it
m
I
x
I
y
I
z
d
m
a
ss
(kg
)
Mo
m
en
t of
inertia
ab
o
u
t X
ax
is
Mo
m
en
t of
inertia
ab
o
u
t Y
ax
is
Mo
m
en
t of
inertia
ab
o
u
t Z
axis
Distan
ce c
en
ter
of
grav
ity
aerod
y
n
a
m
i
c coef
f
icien
t of
thru
st
aerod
y
n
a
m
i
c coef
f
icien
t of
thru
st
Aerod
y
n
a
m
i
c
co
ef
f
icien
t y
aw
o
f
u
p
p
er
roto
r
Aerod
y
n
a
m
i
c
co
ef
f
icien
t y
aw of
lo
w
er
roto
r
0
.29
0
1
.38
3
×
1
0
−3
1
.38
3
×
1
0
−3
2
.72
×
1
0
−4
0
.06
7
6
3
.68
×
1
0
−5
3
.77
×
1
0
−5
1
.47
×
1
0
−6
1
.32
×
1
0
−6
Kg
k
g
m
2
k
g
m
2
k
g
m
2
m
N/rad
2
s
2
N/rad
2
s
2
N.
m
.
rad
-
2
.s
-
2
N.
m
.
rad
-
2
.s
-
2
2.3.
Expressi
on
of
li
nea
ri
z
ed t
angen
t
m
od
el
The
nonlinea
r
syst
e
m
descr
ib
ed
by
eq
uatio
n
(2.16
)
will
ser
ve
as
a
ba
sis
f
or
t
he
desig
n
of
the
c
ontr
ol
al
gorithm
s
by
nonlinea
r
a
ppr
oach
e
s,
but
it
is
abs
olu
te
ly
not
s
uitable
f
or
the
sy
nth
esi
s
of
li
near
al
gor
it
h
m
s.
Howe
ver,
thes
e
appro
ac
h
es
are
base
d
on
a
dynam
ic
m
od
el
li
near
iz
ed
arou
nd
a
certai
n
point
of
ope
rati
on,
in
this
case
hove
rin
g.
The
li
near
iz
at
ion
as
su
m
ption
s
of
t
he
nonlinea
r
m
od
el
relat
ing
to
the
ho
ver
are
a
s
fo
ll
ows:
-
The
tra
ns
la
ti
on
sp
ee
d
v
of the
veh
ic
le
is sm
al
l:
=
(
)
,
-
The p
os
it
ion
ξ
of the
ve
hicle
is any,
-
The rotat
io
n
s
pe
ed
Ω
of the
v
e
hicle
is lo
w
Ω
=
o (e),
-
Roll
angle
φ
a
nd p
it
ch
angle
θ
are l
ow
=
= o (e)
,
-
The
ya
w
a
ng
le
ψ
is
regulat
ed
to
ze
ro.
The
li
near
iz
e
d
m
od
el
ta
ng
e
nt
to
the
vicinit
y
of
th
e
hove
r
i
s
com
po
sed
of
the
f
ollow
i
ng
four
SI
S
O
(S
in
gle In
pu
t
-
Sing
le
O
utput)
ch
a
nn
el
s:
-
A
n al
ti
tud
e c
hain ha
ving it
s
sta
te
[
]
:
[
̇
̇
]
=
[
0
1
0
0
]
[
]
+
[
0
1
]
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Dyna
mic an
alysis a
nd QFT
-
base
d
r
obus
t c
ontrol
desig
n of
…
(
AissaMeksi
)
4105
-
A
roll
ing chai
n hav
i
ng it
s stat
e
[
]
[
̇
̇
]
=
[
0
1
0
0
0
0
0
0
0
0
0
0
1
0
0
]
[
]
+
[
0
0
0
1
]
-
A
p
it
c
h
c
hain
h
a
ving it
s stat
e
[
]
[
̇
̇
̇
̇
]
=
[
0
1
0
0
0
0
−
0
0
0
0
0
0
1
0
0
]
[
]
[
0
0
0
1
]
-
A
stri
ng of la
ce ha
ving as
stat
e
[
]
:
[
̇
̇
]
=
[
0
1
0
0
]
[
]
+
[
0
1
]
3.
QFT C
ONT
R
OLL
ER OF
COA
X
IAL
M
I
CRO
-
HEL
IC
OPTE
R
Qu
a
ntit
at
ive
F
eedb
ac
k
The
ory
is
a
rob
us
t
c
on
t
ro
l
desig
n
t
echn
i
qu
e
that
us
es
feedbac
k
to
achie
ve
respo
ns
es
t
ha
t
m
eet
sp
eci
fied
s
pecifica
ti
on
s
des
pite
struct
ur
e
d
plant
unc
ertai
nty
an
d
plant
disturba
nces
[
14
,
15]
.
This
te
chn
i
qu
e
has
been
a
pp
li
ed
t
o
m
any
c
la
s
ses
of
pro
blem
s
su
c
h
as
Sing
le
In
put
Sing
le
Outp
ut (
SISO
),
M
ulti
pl
e
Inp
ut
Sin
gle
O
utput
(M
ISO
)
a
nd
Mult
iple Inp
ut
Mult
iple Ou
t
pu
t (MIM
O)
f
or
bo
t
h
co
ntinuo
us
and
disc
rete
cases.
For
this
researc
h,
a
SISO
syst
em
is
a
ssu
m
ed
fo
r
c
ontr
ol
la
w
desig
n
an
d
seq
uen
ti
al
loop
cl
os
ures
are
util
iz
ed
as
wit
h
the
po
le
plac
e
m
ent
con
tr
oller.
The
Q
FT
de
sign
m
et
ho
do
log
y
is
qu
it
e
tra
nspare
nt,
al
lo
wing
t
he
desi
gn
e
r
t
o
see
the
necess
ary
tra
de
-
offs
to
ac
hieve
t
he
cl
os
e
d
-
l
oop
s
yst
e
m
sp
eci
ficat
io
ns
[16
,
17]
.
Th
e
basic
ste
ps
of
the
pr
oce
dure
a
s
sh
ow
n
in
Fig
ur
e
5
are
prese
nted
in
the
f
ollow
i
ng
su
b
-
sect
io
ns
.
T
hey are:
[18
-
20]
-
Plant m
od
el
(w
it
h
uncertai
nty)
, Tem
plate
s g
e
ner
at
io
n
a
nd
nom
inal plant se
le
ct
ion
P
o
(j
ω
).
-
Sp
eci
fy
acce
pt
able
trackin
g
m
od
el
s,
wh
i
ch
the
cl
os
ed
-
lo
op
res
pons
e
sat
isfie
s,
T
R
L
≤
T
R
≤
T
RU
,
and d
et
e
rm
ine trackin
g b
ound
s
-
Determ
ine
disturba
nce
re
j
ect
i
on
m
od
el
s,
T
D
,
base
d
on
disturba
nce
re
j
ect
i
on
sp
eci
ficat
ion
s,
a
nd
determ
ine
disturba
nce
bounds
-
L
oop
-
s
hap
i
ng the c
on
t
ro
ll
er
G
(j
ω
).
-
Synthesize
no
m
inal l
oo
p t
ra
nsfer
fun
ct
io
n,
L0 (j
ω
)
-
Pr
e
-
filt
er s
ynt
he
sis
F
(j
ω
).
-
Si
m
ulati
on
a
nd D
esi
gn V
al
ida
ti
on
.
Figure
5.QFT
con
t
ro
l l
oop
3.1.
Genera
tin
g
pl
an
t
t
e
mpla
tes
Am
on
g
t
he
fir
s
t
ste
p
in
Qu
a
nt
it
at
ive
Feedb
ac
k
T
h
eo
ry
(
QFT)
desig
n
proc
ess
are
ge
ner
at
ed
the
plant
tem
plate
s.
The
pla
nt
te
m
plate
s
ca
pture
t
he
uncertai
nties
in
the
plant
tra
ns
f
er
functi
on,
(
)
a
nd
a
re
pl
otted
as
boun
dar
ie
s
of
m
agn
it
ud
e
an
d
ph
ase
va
riat
ion
s
on
t
he
Nich
ols
ch
art
at
sp
eci
fic
desig
n
fr
e
q
ue
ncies
.
These tem
plate
s ar
e t
hen use
d t
o
c
reate b
ound
s
on the
N
ic
hols c
har
t as
sho
wn in
Fi
gure
5 [19
,
21]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
5
,
Oct
ober
20
19
:
4099
-
4
1
1
3
4106
3.2.
Genera
tin
g
pe
rfo
rm
an
ce
b
ounds
Give
n
the
te
m
plate
s,
QF
T
c
onve
rts
the
cl
os
e
d
lo
op
m
a
gn
it
ude
sp
eci
fi
cat
ion
s
i
nto
m
agn
it
ude
a
nd
ph
a
se
co
ns
trai
nt
on
the
open
loop
tra
ns
fe
r
f
un
ct
io
n
for
the
no
m
inal
plant
(
0
=
(
(
)
0
(
)
)
.
Us
ually
syst
e
m
perform
ance is d
esc
ribe
d
as
re
fer
e
nce trac
ki
ng, ro
bust st
abi
li
ty
an
d
re
j
ect
ion t
o
i
nput
disturba
nce[20
,
21]
.
-
Re
fer
e
nce T
rac
king
To
m
eet
the tra
ckin
g per
form
a
nce th
e
contr
ol
le
r
sho
uld
sat
isfy t
o
t
he
f
ollo
wing:
|
|
≤
|
1
+
|
≤
|
|
(18)
w
he
re
t
he
uppe
r
an
d
lo
we
r
boun
ds
a
re
de
fined
us
in
g
ti
m
e
dom
ai
n
fig
ur
e
of
m
erit
su
c
h
as
pe
ak
ove
rsho
ot
and sett
li
ng
ti
m
e.
=
(
ω
n
2
a
⁄
)
(
s
+
a
)
s
2
+
2
Ɛ
ω
n
s
+
ω
n
2
/
ω
n
=
1
/
, a=1
,
Ɛ
=
0
.
6
(19)
=
s
+
1
s
2
+
1
.
2s
+
1
(20)
=
K
(
+
1
)
(
+
2
)
(
+
3
)
/
1
=
0
.
5
,
2
=
1
,
3
=
2
,
K
=
1
(21)
=
1
(
s
+
0
.
5
)
(
s
+
1
)
(
s
+
2
)
(22)
Nu
m
erical
v
al
ues (
ω
n
,
a
,
Ɛ
,
1
,
2
,
3
,
K
)
are take
n
f
ro
m
the sp
eci
ficat
ion
s
.
-
Ro
bust Sta
bili
ty
Robust
sta
bili
ty
in
QF
T
am
ou
nts
to
c
hec
kin
g
sta
bili
ty
us
ing
no
m
inal
plant
(
0
(
)
=
(
)
(
)
)
and the
n dem
on
strat
in
g
sta
bili
ty
o
f
the
wh
ole set pla
nts b
y
assigni
ng a se
nsi
ti
vity
r
at
ing
1
giv
en
b
y
[
19
]
:
|
1
+
|
≤
1
=
(23)
wh
e
re:
μ
is t
he
ci
rcle M
sp
eci
ficat
io
n
in
m
agn
it
ud
e:
M
m
= 2
0 l
og10(
μ
)
Φ
=2
c
os
-
1(
0
.
5
μ
⁄
)
ϵ
[
0,
180º
]
In
[
20]
,
O
dedYan
i
v
translat
ed
this
co
nd
it
i
on
to
desire
d
phase
m
arg
in
(
)
s
how
n
in
(
24
)
a
nd
gain
m
arg
in
(
)
s
ho
w
n
in
(
25
).
=
2
−
1
(
1
2
(
1
)
)
,
PM
≥
180º
-
Φ
(
deg)
(24)
=
20
log
(
1
+
1
1
)
,
GM
≥
1
+
1
μ
⁄
(
m
agn
it
ud
e
)
(25)
-
Dist
urba
nce rejecti
on at
plan
t ou
t
pu
t
An
uppe
r
lim
it is set to the s
ensiti
vity
f
un
ct
i
on
t
o
lim
it
the p
eak
value o
f
disturba
n
ce am
plific
at
ion
t
o
the p
la
nt out
put as f
ollow
:
[
19
-
21]
|
1
1
+
|
≤
(26)
=
=
1
1
+
≤
2
+
2Ɛ
ω
n
2
+
2Ɛ
ω
n
+
ω
2
/
ω
n
=
1
/
a=1
,
Ɛ
=
0
.
6
(27)
=
1
1
+
≤
2
+
1
.
2
2
+
1
.
2
+
1
(28)
These
boun
ds
are the
n use
d f
or lo
op sh
a
ping to
d
e
sig
n
a c
on
t
ro
ll
er.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Dyna
mic an
alysis a
nd QFT
-
base
d
r
obus
t c
ontrol
desig
n of
…
(
AissaMeksi
)
4107
3.3.
Contr
oller
d
es
ign
The
co
ntr
oller
is
design
e
d
by
add
i
ng
po
le
s
a
nd
ze
ro
es
to
th
e
no
m
inal
transf
er
f
unct
ion
by
sat
isfyi
ng
al
l
bo
un
ds
at
each
f
reque
ncy.
Durin
g
this
sta
ge,
the
desig
ne
r
co
n
side
rs
a
trade
-
off
betw
een
the
s
pecifica
ti
on
,
con
t
ro
ll
er
com
plexity
and t
he
co
st
of
fee
dba
ck
in
the
ba
nd
width [
19
-
21]
.
(
)
=
3
.
072
(
−
12
)
(
−
350
)
(
−
350
)
(29)
3.4.
Pre
-
filter
d
esi
gn
A
pre
-
filt
er
is
need
e
d
to
br
in
g
the
res
pons
e
within
the
re
f
eren
ce
trac
king
tolera
nces,
and
an
d
is
done by a
dd
i
ng
po
le
s
and ze
r
oes
[19
-
21]
.
(
)
=
1
(
−
0
.
30
0
66
)
(30)
Si
m
ulati
on
a
nd
validat
io
n of
the
desig
n
is
show
n
i
n
Fi
gure
6 [22]
.
Figure
6.QFT
desig
n proce
dure
4.
RESU
LT
S
AND A
N
ALYSIS
The
f
re
qu
e
ncy
range
f
r
om
0.
01
to
1000
r
ad/s
was
sel
ect
ed
f
or
t
he
Q
FT
c
ontr
oller
desi
gn.
This
ra
nge
cov
e
rs
t
he
fu
ll
range
of
f
re
quencies
f
or
a
ty
pi
cal
coa
xial
m
ic
ro
-
helic
op
t
er.
T
he
first
ste
p
of
the
QFT
de
sign
is
to
ge
ner
at
e
the
plant
te
m
plates.
These
plant
t
e
m
plate
s
wer
e
gen
e
rated
us
in
g
the
Ma
tl
ab
s
of
t
war
e
[
23]
and
a
re
sh
ow
n
in
Fi
gur
e 7
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
9
, N
o.
5
,
Oct
ober
20
19
:
4099
-
4
1
1
3
4108
Figure
7.Pl
ant
tem
plate
s
Crit
erion
s
n
e
ed
ed fo
r
the
d
e
si
gn
i
ng of t
he
c
ontr
oller are:
-
T
rack
i
ng crit
erio
n
The
up
per
bo
und
f
or
the
tr
ackin
g
crit
eri
on
was
ge
ner
at
ed
by
c
on
si
de
rin
g
1.1
9
%
oversho
ot
an
d
set
tl
ing
tim
e
o
f
4.4
8
s.
T
he
trans
fer
functi
on
us
e
d
f
or
th
e
uppe
r
boun
d
s
how
n
in
(
19).
Si
m
il
arly
,
the
lowe
r
bound
was
ge
ne
rated
by
c
on
s
iderin
g
no
ove
r
sho
ot
an
d
set
t
li
ng
ti
m
e
of
9.
77
s
.
T
he
tra
nsf
er
f
unct
ion
use
d
f
or
the
l
ow
e
r
bound
s
how
n
i
n
(
22)
.T
hese
bounds
i
n
ti
m
e
do
m
ai
n
and
fr
e
quency
do
m
ai
n
are
sho
wn
i
n
Figure
8.
T
he
trac
ki
ng
bounds are
s
hown in t
he Nic
hols
plo
t i
n
Fig
ure
9.
Figure
8.
T
rac
ki
ng
boun
ds
; (a
)
tim
e d
om
ai
n; (
b) freq
ue
ncy dom
ai
n
Figure
9
.
I
nters
ect
ion
bounds
(stabili
ty
, trac
ki
ng
a
nd
disturb
ance
rej
ect
io
n)
-
3
5
0
-
3
0
0
-
2
5
0
-
2
0
0
-
1
5
0
-
1
0
0
-
5
0
0
-
1
5
0
-
1
0
0
-
5
0
0
50
100
N
i
ch
o
l
s
p
l
o
t
O
p
e
n
-
l
o
o
p
p
h
a
se
(
d
e
g
)
O
p
e
n
-
l
o
o
p
m
a
g
n
i
t
u
d
e
(
d
B
)
0
.
0
1
0
.
0
5
0
.
1
0
.
5
1
5
10
50
100
500
1000
-
1
0
0
-
8
0
-
6
0
-
4
0
-
2
0
0
20
M
a
g
n
i
t
u
d
e
(
d
B
)
10
-2
10
-1
10
0
10
1
10
2
-
1
8
0
-
1
3
5
-
9
0
-
4
5
0
P
h
a
s
e
(
d
e
g
)
B
o
d
e
D
i
a
g
r
a
m
F
r
e
q
u
e
n
c
y
(
r
a
d
/
s
)
Tu
Tl
0
2
4
6
8
10
12
14
0
0
.
2
0
.
4
0
.
6
0
.
8
1
1
.
2
1
.
4
S
t
e
p
R
e
s
p
o
n
s
e
T
i
m
e
(
s
e
c
o
n
d
s
)
A
m
p
l
i
t
u
d
e
Tu
Tl
-
3
5
0
-
3
0
0
-
2
5
0
-
2
0
0
-
1
5
0
-
1
0
0
-
5
0
0
-
2
0
-
1
0
0
10
20
30
40
50
60
70
O
p
e
n
-
l
o
o
p
m
a
g
n
i
t
u
d
e
(
d
B
)
O
p
e
n
-
l
o
o
p
p
h
a
se
(
d
e
g
)
N
i
ch
o
l
s
p
l
o
t
Evaluation Warning : The document was created with Spire.PDF for Python.