Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
n
g (IJ
E
C
E)
Vo
l.
10
,
No.
2
,
A
pr
il
2020, p
p. 18
68
~
1884
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v10
i
2
.
pp1868
-
18
84
1868
Journ
al h
om
e
page
:
http:
//
ij
ece.i
aesc
or
e.c
om/i
nd
ex
.ph
p/IJ
ECE
Elevati
on
, pitch
and tra
vel
ax
is
stabil
izati
on of
3DOF heli
copte
r
with hyb
rid contr
ol syste
m by
GA
-
LQ
R b
ase
d P
I
D cont
rolle
r
Ibra
him
K. M
ohammed
,
Abdul
la I. A
bdul
la
Depa
rtment
o
f
S
y
stems
and
Cont
rol
Eng
ine
er
ing,
Coll
ege of
Elec
t
ronic
s E
ng
ineeri
ng,
Ninev
ah
Uni
ver
sit
y
,
Ira
q
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
A
pr
29
, 201
9
Re
vised
A
ug
28
,
20
19
Accepte
d
Oct
9
, 2
0
19
Thi
s
rese
arc
h
work
int
roduc
e
s
an
eff
ic
i
ent
h
y
brid
con
trol
m
et
hodolog
y
through
combin
ing
the
tra
di
ti
o
nal
proport
ional
-
int
eg
ral
-
d
er
iv
ative
(PID
)
cont
roller
and
li
nea
r
qu
adr
a
ti
c
reg
u
lator
(
LQR)
opti
m
al
cont
rolher.
The
proposed
h
y
brid
cont
rol
ap
proa
ch
is
adopte
d
to
design
three
degr
e
e
of
fre
edom
(3DO
F
)
stabi
lizing
s
y
s
te
m
for
hel
ic
opt
er.
The
g
ai
n
pa
ramet
ers
of
the
c
la
ss
ic
PID
cont
roller
ar
e
det
ermined
usin
g
the
e
le
m
ent
s
of
the
LQR
fee
dba
ck
gai
n
m
at
rix.
The
d
ynamic
beha
viou
r
of
the
LQR
base
d
PI
D
cont
roller,
is
m
odel
ed
in
st
ate
spac
e
form
to
en
abl
e
ut
lizing
sta
te
fe
edba
c
k
cont
roller
techn
ique
.
Th
e
per
fo
rm
anc
e
of
the
proposed
LQR
base
d
LQR
c
ontrol
l
er
is
improved
b
y
usin
g
Gene
tic
Algo
rit
hm
opti
m
iz
a
tion
m
et
hod
which
ar
e
adop
te
d
to
obt
ai
n
o
pti
m
um
val
ues
for
LQR
controlle
r
gai
n
par
amete
rs
.
Th
e
LQR
-
PID
h
y
b
rid
cont
ro
ll
er
i
s
sim
ula
te
d
usi
ng
Matl
ab
envi
ronm
ent
an
d
it
s
per
form
an
ce
is
ev
al
ua
te
d
base
d
o
n
rise
tim
e,
sett
l
ing
ti
m
e,
over
shoot
and
stea
d
y
st
ate
err
or
par
amete
r
s
to
val
id
at
e
the
proposed
3DO
F
hel
ic
opte
r
bal
an
ci
ng
sy
stem.
Based
on
GA
tuni
ng
appr
oac
h
,
the
sim
ula
ti
on
r
esult
s
suggest
t
hat
the
h
y
br
id
LQR
-
PID
cont
roll
er
c
an
be
eff
ec
ti
ve
l
y
emplo
y
ed
t
o
stabili
ze
the
3DO
F he
l
ic
o
pte
r
s
y
stem
.
Ke
yw
or
d
s
:
3DOF heli
co
pt
er s
yst
em
Gen
et
ic
al
gorithm
Hybr
i
d
c
ontrol
LQR c
on
t
ro
ll
er
PI
D
contr
oller
Copyright
©
202
0
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
Ibrah
im
K
. M
oham
m
ed,
Dep
a
rtm
ent o
f Sy
stem
s an
d
C
on
t
ro
l E
nginee
rin
g,
Coll
ege
of
Ele
ct
ronics E
ng
i
ne
ei
ng
,
Nine
vah Unive
rsity
, Mo
s
ul, Ir
aq
Em
a
il
: i
br
ahim
.m
oh
a
m
m
ed@
uonin
e
va
h.
e
du.
iq
1.
INTROD
U
CTION
The
A
3D
OF
helic
op
te
r
m
od
el
is
a
m
ul
ti
-
input,
m
ult
i
-
outp
ut
syst
em
(M
IMO)
with
stron
gly
nonlinea
r
cha
r
act
erist
ic
s,
us
e
d
as
ex
per
im
e
ntal
platfo
rm
for
ve
rificat
io
n
of
va
rio
us
con
t
ro
l
al
gorit
hm
s
.
The
helic
op
te
r
is
a
com
plex
m
echan
ic
al
schem
e
and
has
un
sta
ble
dy
na
m
ic
s
wh
ic
h
m
ake
usi
ng
tra
di
ti
on
al
con
t
ro
l
m
et
ho
ds
to
sta
bili
ze
i
t
is
a
big
chall
en
ge.
In
con
t
ro
l
proc
edure
of
3D
OF
helic
opte
r
syst
e
m
the
no
nlinear
dynam
ic
s
is
lin
earise
d
a
bout
pr
es
el
ect
ed
operati
ng
c
ondi
ti
on
s.
Ba
se
d
on
li
near
ise
d
s
yst
e
m
m
od
el
a
cl
assic
al
pro
portiona
l
integ
ral
de
riva
ti
ve
PID
c
ontr
oller
te
ch
nique
is
wi
d
el
y
us
e
d
due
to
it
s
sim
p
li
ci
t
y
and
ea
sy
to
re
al
iz
e
[1
]
.
H
owever,
us
i
ng
c
onve
ntio
nal
PID
te
ch
nique
is
insuffici
ent
to
gu
i
de
the
helic
op
te
r
thr
ough
the
dem
and
tra
j
ect
or
y
as
the
re
gu
la
ti
on
of
th
e
co
ntr
oller
pa
ram
et
ers
is
hard
ta
s
k
e
ve
n
f
or
an
e
xp
e
rience
d
c
on
t
ro
l
e
ngineers
[
1
,
2].
I
n
a
no
t
her
co
nt
ro
l
m
et
ho
d,
a
n
optim
al
tracking
ap
proac
h
ui
li
zi
ng
LQR
c
on
tr
oll
er
to
sta
bili
ze
the
he
li
copt
er
syst
e
m
was
pro
posed
by
[
3].
I
n
th
e
pro
posed
s
yst
e
m
,
a
Ge
netic
Al
gorithm
(GA)
is
em
plo
ye
d
to
o
pti
m
iz
e
a
cl
assic
LQR
c
ontr
olle
r
f
or
po
sit
ion
c
on
tr
ol
of
DC
m
oto
r.
By
us
i
ng
LQR
c
ontr
oller
a
sig
nific
ant
im
pr
ov
em
ent
in
gu
i
dan
ce
of
helic
opte
r
syst
e
m
is
carried
ou
t.
Fu
rt
her
in
vesti
gations
in
co
ntro
l
strat
egies
f
or
he
li
cop
te
r
s
yst
e
m
are
st
il
l
the
go
al
of
m
a
ny
researc
hers.
A
ne
w
con
t
ro
l
a
ppr
oa
ch
base
d
on
hybri
d
c
ontrol
le
r
te
ch
nique
was
pro
po
se
d
by
m
any
res
earche
rs
t
o
c
on
t
ro
l
the
3DOF
hel
ic
op
te
r
syst
em
.
Fan
a
nd
J
oo
[4
]
pro
posed
a
hybr
i
d
co
nt
ro
l
m
e
tho
d
usi
ng
fu
zzy
an
d
PID
con
t
ro
ll
er
to
ba
la
nce
the
3D
OF
helic
opte
r
syst
e
m
at
desir
ed
e
qu
il
ib
r
ium
posit
ion.
In
th
is
stud
y,
an
ef
f
ic
ie
nt
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
El
ec &
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ele
vation,
pitch
and
tr
avel
ax
is st
ab
il
izati
on
of 3D
OF heli
cop
te
r
wi
th
…
(
I
br
ahim
K.
Mo
ham
me
d
)
1869
con
t
ro
l
m
et
ho
d
has
bee
n
a
dopt
ed
to
pro
pose
a
hybri
d
c
ontr
ol
strat
egy
us
i
ng
both
co
nve
ntion
al
a
nd
int
el
li
gen
t
con
t
ro
l
al
gorit
hm
s.
This
hybr
i
d
co
ntr
ol
appr
oach
us
es
PI
D
c
ontrolle
r
as
a
con
ve
nt
ion
al
co
ntr
ol,
wh
il
e
th
e
li
ner
quad
r
at
ic
reg
ulato
r
(
LQR)
c
on
tr
oll
er
base
d
on
G
A
us
e
d
as
an
i
ntell
igent
co
ntr
ol.
The
L
QR
optim
a
l
con
t
ro
ll
er
is
r
ecom
m
end
e
d
to
i
m
ple
m
ent
t
he
intel
li
gen
t
syst
e
m
du
e
to
it
s
hig
h
pr
ec
isi
on
in
m
ov
em
ent
app
li
cat
io
ns
[
5
]
.
I
n
this
sta
te
feedbac
k
co
ntr
oller
te
ch
nique
a
trade
off
bet
ween
t
he
pla
nt
char
act
erist
ic
s
and
the
in
put ef
f
ort
of the syst
em
can be ac
hieve
d
if
the c
ontr
oller g
ai
n param
et
ers
are
tu
ned
pro
pp
e
rly
[
6
-
8
]
.
T
he
LQR
c
ontrolle
r
desi
gn
can
be
im
pr
oved
by
em
plo
yin
g
s
of
t
com
puti
ng
an
d
opti
m
iz
at
i
on
te
chni
qu
es
,
wh
ic
h
a
re
a
do
pted
to
ac
hieve
tun
i
ng
proecss
of
the
c
on
tr
oll
er
gain
pa
ram
e
te
rs
,
s
uc
h
as
D
iffer
e
ntial
Ev
o
l
ution
(D
E
)
[9
]
,
Part
ic
le
Sw
arm
In
sp
ire
d
Ev
olu
ti
on
a
ry
Algorit
hm
(P
S
-
EA
)
[
10
]
,
Pa
rtic
le
Sw
arm
Op
ti
m
isa
ti
o
n
(P
S
O) [
11
],
Ge
netic
A
lg
ori
th
m
(
G
A)
[
12
]
,
m
e
m
e
ti
c algo
ri
thm
(
MA)
[
13
]
,
I
m
per
ia
li
st C
om
petit
ive A
lg
or
it
hm
(I
CA
)
[
14
]
,
A
nt
Colon
y
O
ptim
iz
at
ion
(A
C
O)
[
15
]
,
Ar
ti
fi
ci
al
Be
e
C
olo
ny
(A
BC
)
[
16
]
,
and
Ar
ti
fici
al
I
m
m
un
e
S
yst
e
m
s
(A
IS)
[
17
]
.
I
n
or
de
r
to
validat
e
the
pr
opos
e
d
he
li
cop
te
r
m
odel
,
the
hy
br
i
d
con
t
ro
l
syst
e
m
i
s
si
m
ulate
d
by
us
in
g
the
Ma
tl
ab
en
vir
onm
e
nt,
a
nd
it
s
bal
ancin
g
perfor
m
ance
is
then
evaluate
d
bas
ed
on
sta
bili
ty
crit
eri
a
par
am
et
ers
wh
ic
h
inclu
de
ri
se
and
set
tl
ing
tim
e,
ov
ersho
ot
and
co
ntr
ol
input.
T
he
pur
pose
of
the
pr
opos
e
d
hybri
d
G
A
-
L
QR
ba
sed
PID
c
on
t
ro
ll
er
i
s
to
desig
n
a
n
e
ff
ic
ie
nt
c
ontr
ol
syst
em
util
iz
ed
to
sta
bili
ze
a
pr
ot
otype
3DO
F
helic
op
te
r
syst
e
m
at
desired
ro
ll
an
d
pitch
po
sit
io
ns
as
w
el
l
as
angular
travel
sp
ee
d.
Durin
g
the
cu
rr
e
nt
dec
ade
the
re
has
be
en
a
c
onsider
able
intere
st
by
m
any
con
tr
ol
researc
he
rs
in
us
in
g
of
hy
br
id
co
ntr
ol
ap
proac
h
i
n
va
rio
us
e
ng
i
ne
erin
g
a
pp
li
cat
ion
s
.
V
en
doh
and
J
ov
it
ha
[
18
]
prese
nted
a
hybri
d
con
t
ro
l sy
ste
m
us
in
g
L
QR
bas
ed
P
ID co
ntr
ol
le
r
f
or
sta
bili
zat
ion
a
nd tra
j
ect
or
y t
rac
king
of m
agn
et
ic
le
vit
at
io
n
syst
e
m
.
Ar
ba
b
et
al
.
[1
9
]
pro
po
s
ed
a
hy
br
i
d
con
tr
ol
strat
e
gy
us
in
g
f
uzzy
based
L
QR
c
on
t
ro
ll
er
f
or
3
-
D
OF
helic
op
te
r
syst
e
m
.
Cho
ud
hary
[1
]
desig
n
an
oth
e
r
hybri
d
c
on
t
ro
l
schem
e
us
in
g
LQR
bas
ed
PID
co
ntr
olle
r
for
3
-
DOF
helic
op
te
r
syst
em
,
ho
w
eve
r,
t
he
i
ntr
odu
ce
d
hybri
d
con
t
ro
l
syst
em
s
are
not
ef
fici
ent
as
th
e
pa
ra
m
et
ers
of
the
LQR
fe
edb
ac
k
gai
n
m
at
rix
are
obta
ined
by
a
pply
ing
tria
l
a
nd
error
pro
ced
ure.
T
he
pa
per
pr
ese
nt
s
a
sim
ple
m
e
tho
d
f
or
the
a
pp
roxim
ation
of
PI
D
co
ntr
oller
gain
pa
ram
eter
s
f
r
om
feedback
gai
n
m
atr
ix
of
the
co
ntr
oller.
The
LQR
co
ntr
oller
ga
in
i
s
obta
ine
d
f
r
om
the
wei
gh
ti
ng
m
at
rices
Q
an
d
R
w
hich
their
el
e
m
ents v
al
ue
s ar
e t
un
e
d
e
ff
e
ct
ively
b
y t
he GA
opti
m
iz
at
i
on m
et
ho
d.
The
rest
of
th
e
pap
e
r
is
organ
ise
d
as
f
ol
lows
.
Sect
ion
2
prese
nts
co
nf
i
gu
rati
on
a
nd
dyna
m
ic
s
m
od
el
ing
of
th
e
he
li
cop
te
r
sy
stem
.
In
sect
io
n
3
,
co
ntr
oller
te
chn
iq
ues
of
helic
op
te
r
hy
bri
d
co
ntr
ol
syst
e
m
are
introd
uced.
T
unin
g
m
et
ho
d
is
desc
ribe
d
in
s
ect
ion
4.
Sect
ion
5
intr
oduce
s
helic
opte
r
c
ontr
ol
syst
em
de
sign
is
pr
ese
nted
.
In
s
ect
ion
6
sim
ulati
on
res
ults
of
the
hybri
d
co
ntro
l
syst
em
are
intro
du
ced
a
nd
f
ollo
wed
by
con
cl
us
io
ns
a
nd
pro
s
pecti
ve r
esearch
in se
ct
ion 7
.
2.
STRUCT
URE
AND D
YNAM
I
CS M
ODE
LING
OF H
EL
ICO
PTE
R
2
.
1.
Heli
copte
r
s
tu
c
tu
re
The
c
on
ce
ptu
a
l
platfo
rm
of
3
-
DOF
helic
opte
r
syst
em
is
sh
ow
n
in
Fig
ure
1.
It
c
on
sist
s
of
a
n
a
rm
m
ou
nted
on
a
base.
T
he
m
a
i
n
body
of
the
helic
op
te
r
com
po
s
ed
of
pro
pe
ll
ers
dr
ive
n
by
two
m
oto
rs
m
ounte
d
are
the
ei
the
r
ends
of
an
s
hort
bala
nce
ba
r.
The
whole
he
li
cop
te
r
body
is
fi
xed
on
one
en
d
of
the
ar
m
and
a
balance
bl
oc
k
instal
le
d
at
the
oth
e
r
en
d.
T
he
balance
a
rm
can
ro
ta
te
a
bo
ut
tra
vel
axis
a
s
well
as
sl
op
e
on
a
n
el
evati
on
a
xis.
The
body
of
he
li
cop
te
r
is
fr
e
e
to
r
oll
ab
ou
t
the
pitch
a
xis.
The
syst
em
is
pro
vid
e
d
by
e
nc
od
e
r
s
m
ou
nted
on
t
hese
a
xis
us
e
d
to
m
easur
e
the
tra
vel
m
oti
on
of
the
a
rm
and
it
s
el
evati
on
a
nd
pitch
ang
le
.
The
prop
el
le
r
s
with
m
oto
rs
can
ge
ne
rate
an
el
evati
on
m
echan
ic
al
f
orc
e
pro
portiona
l
to
the
vo
lt
age
powe
r
su
ppli
ed
to
the
m
oto
rs.
This
f
or
ce
ca
n
cause
the
helic
op
te
r
body
to
li
ft
of
f
the
gr
ou
nd.
It
is
wo
rt
h
co
ns
ider
i
ng
that t
he pur
po
s
e of
us
in
g balance
blo
c
k
is t
o red
uce t
he vo
lt
age
powe
r
s
upplied to
the
propell
ers
m
oto
rs.
2
.
2.
Heli
copte
r dynamics
an
d mathem
ati
cal m
od
el
ing
I
n
this
stu
dy,
the
no
nlinear
dy
nam
ic
s
of
3D
OF
helic
op
te
r
syst
e
m
is
m
od
el
ed
m
a
them
a
t
ic
al
ly
based
on
de
velo
ping
the
m
od
el
of
the
syst
e
m
beh
avi
or
for
el
evati
on,
pitch
and
tra
vel
axi
s.
The
def
i
niti
on
of
the sym
bo
ls an
d no
m
ecal
ture of
the
pro
pose
d heli
cop
te
r
sy
stem
is includ
e
d
in
Ta
ble 1.
2.2.1.
El
e
vatio
n axis m
od
e
l
The
f
ree
bo
dy
diagr
am
of
3DO
F
helic
opte
r
syst
e
m
based
on
el
evati
on
axis
is
sh
ow
n
in
Fig
ur
e
2.
The
m
ov
em
ent of the
elevat
io
n
a
xis is
gove
r
ned b
y t
he foll
ow
i
ng d
i
ff
e
rent
ia
l equ
at
io
ns
:
̈
=
+
cos
(
)
−
,
−
,
(1)
wh
e
re
is t
he
c
entrif
ug
al
t
orq
ue,
w
hich
is
fu
nction o
f
(
)
,
,
is fri
ct
ion
m
o
m
ent ex
erte
d on
the h
el
ic
op
te
r
el
evati
on
axis
,
w
hich
is
gen
e
rated
from
co
m
bin
ing
j
oi
nt
f
rict
ion
a
nd
ai
r
resist
ance,
an
d
,
is
t
he
e
ff
ect
i
ve
gr
a
vitat
ion
al
t
orq
ue d
ue
to
the
m
ass d
iffer
e
nt
ia
l
=
ℎ
about t
he
e
le
vation axis.
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.
10
, No
.
2
,
A
pr
i
l 202
0
:
1868
-
1884
1870
Fig
ure
1. 3D
O
F H
el
ic
opte
r
S
yst
e
m
[
20
]
Figure
2. Sc
he
m
at
ic
d
ia
gr
am
of ele
vatio
n
a
xi
s m
od
el
f
or
3DO
F
helic
op
te
r
syst
em
,
=
ℎ
(2)
If
the
el
evati
on
a
ngle
=
0
a
nd
th
e
r
otati
on
a
ng
l
e
of
pitch
a
xis
=
0
,
the
n
the
ce
ntr
ifu
gal
to
r
qu
e
(
)
wil
l
be
ze
ro
(
1) b
as
ed on E
uler’
s
s
econd la
w bec
om
es as f
ollo
w
s:
̈
=
−
,
+
,
(3)
̈
=
(
+
)
−
,
+
,
(4)
=
=
,
(5)
̈
=
(
+
)
−
,
+
,
(6)
̈
=
−
,
+
,
(7
)
2.2.2.
Pitch
ax
is model
Con
si
der
t
he
pi
tc
h
schem
at
ic
diag
ram
of
the
syst
e
m
in
Fig
ur
e
3.
It
ca
n
be
seen
from
the
fig
ur
e
t
hat
the
m
ai
n
torqu
e
act
i
ng
on
t
he
syst
e
m
pitch
axis
is
pro
duced
from
the
thru
s
t
fo
rc
e
ge
nerat
ed
by
t
he
pr
operll
er
m
oto
rs.
Wh
e
n
≠
0
, th
e
gravit
at
io
nal force
w
il
l a
lso
gen
e
rate a t
orq
ue
,
act
s on
t
he heli
cop
te
r p
it
ch
axis.
,
+
,
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
El
ec &
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ele
vation,
pitch
and
tr
avel
ax
is st
ab
il
izati
on
of 3D
OF heli
cop
te
r
wi
th
…
(
I
br
ahim
K.
Mo
ham
me
d
)
1871
Figure
3. Sc
he
m
at
ic
d
ia
gr
am
of p
it
ch
ax
is m
od
e
l
f
or
3DO
F
helic
op
te
r
sc
he
m
e
Table
1.
N
om
e
nc
la
ture
of t
he heli
cop
te
r
syst
e
m
No
m
en
clatu
re
Def
in
itio
n
,
,
ℎ
Ro
tatio
n
ang
le of
e
lev
atio
n
,
p
itch
and
tr
av
el axis
r
esp
ecti
v
ely
(
d
eg
re
e)
Distan
ce between
t
h
e pitch
axis
an
d
the h
elico
p
ter
cent
e
r
o
f
g
ravity
(
m
)
Distan
ce fro
m
axle
po
in
t to eith
er
p
rop
eller
m
o
to
r
(
m
)
Distan
ce fro
m
axle
po
in
t to th
e balan
ce blo
ck
(
m
)
Distan
ce between
t
h
e pitch
axis
and
the center of
pro
p
ellers (
m
)
Lif
t f
o
rce
o
f
the p
rop
erller
m
o
to
rs
(
N
)
,
,
Fo
rces of
f
o
rnt an
d
back
pro
p
erller
m
o
to
rs (
N)
Vo
tag
e app
lied
to th
e f
ron
t and
back
pro
p
erller
m
o
to
r
r
e
sp
ectiv
ely
(
v
)
,
,
,
,
,
,
,
,
,
,
Mass o
f
balan
ce blo
ck
(
k
g
)
Mass o
f
pro
p
erller
m
o
to
rs
(
k
g
)
Mass d
if
f
erential a
b
o
u
t the elev
atio
n
axis
(
k
g
)
Torq
u
e co
m
p
o
n
en
t gen
erate
d
f
ro
m
co
m
b
in
in
g
the jo
in
t f
riction
and
air
resi
stan
ce fo
r
elev
atio
n
,
p
itch
and
rate
ax
is
resp
ectiv
e
ly
(
N.
m
)
Cen
trif
u
g
al torq
u
e
abo
u
t the elev
atio
n
axis
(
N.
m
)
Ef
f
ectiv
e gravit
atio
n
al torq
u
e du
e to g
ravity
ab
o
u
t the e
lev
atio
n
axis
and
p
itch
axis
r
esp
ectiv
e
ly
(
N.
m
)
Inertia
m
o
m
en
t
of
the syste
m
abo
u
t e
lev
atio
n
,
p
itch
and
tr
av
el axis
r
esp
ecti
v
ely
(
k
g
.
m
2
)
Fo
rce
co
n
stan
t of
t
h
e
m
o
to
r
-
p
rop
eller
co
m
b
in
atio
n
The dynam
ic
s o
f
the
pitch a
xi
s can
b
e
m
od
el
ed
m
at
he
m
at
icall
y as fo
ll
ows:
̈
=
ℎ
−
ℎ
−
,
−
,
(
8)
Wh
e
re
,
is fr
ic
ti
on m
o
m
ent exe
rted o
n
the
p
it
ch
a
xis.
,
=
ℎ
sin
(
)
cos
(
)
(9)
Ba
sed on t
he
a
ssu
m
ption
that
the p
it
ch
a
ng
le
=
0
,
,
=
0
, th
e
n (8)
bec
om
es as f
ollo
w
s:
̈
=
ℎ
(
−
)
−
,
(10)
̈
=
(
−
)
−
,
(11)
̈
=
−
,
(12)
2.
2.3
.
Tr
avel
axis m
od
el
The
fr
ee
bo
dy
diag
ra
nm
of
the
he
li
cop
te
r
syst
e
m
dynam
ic
s
base
d
on
t
rav
el
a
xis
is
presente
d
i
n
Figure
4.
In
th
is
m
od
el
,
wh
e
n
≠
0
,
the
m
ai
n
forces
act
in
g
on
the
helic
opte
r
dynam
ic
s
are
the
th
rust
f
or
c
e
s
of
pro
pe
rlle
r
m
oto
rs
(
,
).
The
se
f
or
ce
s
hav
e
a
c
om
po
ne
nt
gen
e
rates
a
to
r
qu
e
on
the
tra
vel
a
xis.
Assu
m
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.
10
, No
.
2
,
A
pr
i
l 202
0
:
1868
-
1884
1872
that
the
helic
opte
r
bo
dy
has
r
oll
up
by
a
n
an
gle
as
s
how
n
i
n
Fig
ur
e
4.
T
he
n
t
he
dy
nam
i
cs
of
tra
vel
a
xi
s
f
or
3DOF heli
co
pt
er s
yst
em
is
m
od
e
le
d
as
foll
ows:
̇
=
−
(
+
)
(
)
−
,
(13)
Figure
4. Sc
he
m
at
ic
d
ia
gr
am
of travel
rate a
xis m
od
el
for 3
DOF
helic
op
te
r
sc
hem
e
The
t
hrust
f
oc
es
of
t
he
t
wo
pro
per
ll
er
m
oto
r
s
(
+
)
are
re
qu
ire
d
to
keep
the
helic
op
te
r
in
flig
ht
case an
d
is a
pp
roxim
at
e
ly
.
̇
=
−
(
)
−
,
(14)
Wh
e
re
,
is
fr
ic
ti
on
m
om
ent
exer
te
d
on
tra
vel
a
xis.
As
appr
oach
s
to
zero,
base
d
on
si
nc
functi
on,
(
)
=
, th
e a
bove
equ
at
ion
beco
m
es as foll
ows:
̇
=
−
−
,
(
15)
Ba
sed
on
the
a
ssu
m
ption
that
the
co
upli
ng d
ynam
ic
s,
gravi
ta
ti
on
al
tor
que (
,
)
an
d
fr
ic
ti
on m
o
m
ent
exer
te
d
on
el
evati
on,
pi
tc
h
an
d
tra
vel
axis
a
re
ne
glect
ed,
th
en
th
e
dynam
ic
s
m
od
el
ing
(7),
(
12)
and
(
15
)
f
or
3D
O
F
helic
op
te
r
sys
t
e
m
can
be
sim
plifie
d
as
in
(1
6), (1
7) an
d (18
)
r
es
pecti
vely
[1].
̈
=
(16
)
̈
=
(17
)
̇
=
(1
8)
2.3.
S
ys
te
m state
sp
ace m
odel
In
or
der
to
des
ign
sta
te
feedb
ack
co
ntr
oller
base
d
on
L
QR
te
chn
iq
ue
for
3DOF
helic
op
t
er
syst
e
m
,
the
dy
nam
ic
s
m
od
el
of
the
syst
e
m
sh
ou
ld
be
f
or
m
ulate
d
in
sta
te
sp
ac
e
form
.
In
this
stud
y,
the
pr
opos
e
d
hybri
d
c
on
tr
ol
al
gorithm
is
inv
est
igate
d
for
t
he
pu
rpose
of
con
t
ro
l
of
pitch
an
gle,
el
evat
ion
a
ng
le
a
nd
t
rav
el
rate
of
3D
OF
helic
opte
r
sc
hem
e
by
reg
ul
at
ing
the
vo
l
ta
ge
suo
pp
li
es
to
the
fron
t
and
bac
k
m
otors
.
Let
(
nx
1
)
=
[
1
,
2
,
3
,
4
,
5
,
6
,
7
]
=
[
,
,
̇
,
,
̇
,
ʓ
,
]
be
th
e
sta
te
vect
or
of
t
he
syst
em
,
the
sta
te
var
ia
bles
are
c
ho
s
en
a
s
the
a
ng
le
s
a
nd
rate
an
d
their
co
r
respo
nd
i
ng
an
gu
la
r
vel
ociti
es,
an
d
ʓ
̇
=
,
̇
=
.
The
volt
ages
s
upplied
to
the
fron
t
an
d
ba
ck
pr
op
el
le
rs
m
oto
rs
are
c
on
si
de
red
the
in
put’
s
vector
s
uc
h
t
hat,
(
)
(
mx1
)
=
[
1
,
2
]
=
[
,
]
an
d
the
el
evati
on
an
gle,
pitch
a
ngle
and
tra
vel
rat
e
are
ass
um
ed
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
El
ec &
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ele
vation,
pitch
and
tr
avel
ax
is st
ab
il
izati
on
of 3D
OF heli
cop
te
r
wi
th
…
(
I
br
ahim
K.
Mo
ham
me
d
)
1873
the
ou
t
pu
t’
s
ve
ct
or
su
c
h
t
ha
t,
(
)
(
px1
)
=
[
,
,
]
.
Ba
se
d
on
(13)
-
(
15)
,
c
hoos
in
g
t
hese
st
at
e
var
ia
bles
yi
el
d
s
the foll
owin
g
syst
em
st
at
e sp
ace m
od
e
l:
1
̇
=
=
2
2
̇
=
̇
=
3
3
̇
=
̈
=
(
+
)
4
̇
=
̈
=
(
−
)
(16)
5
̇
=
̇
=
2
6
̇
=
ʓ
̇
=
1
7
̇
=
̇
=
4
The
ge
ne
ral
sta
te
and
outp
ut
m
at
rix
equ
at
ion
s
desc
rib
in
g
the
dynam
ic
beh
a
vior
of
th
e
li
near
-
ti
m
e
-
inv
a
riant
(
LTI
)
helic
opte
r
syst
e
m
in
sta
te
sp
a
ce f
or
m
are
as
fo
ll
ows:
(
̇
)
=
(
)
+
(
)
(17)
(
)
=
(
)
+
(
)
(18)
Wh
e
re
(
n
x
n
)
is
the
s
yst
e
m
m
a
trix,
(
n
x
m
)
is
the
in
pu
t
m
at
rix,
(
p
x
m
)
is
the
outpu
t
m
at
rix,
a
nd
(
m
x
p
)
is
feed
f
orward
m
at
rix,
f
or
t
he
desig
ne
d
sy
stem
.
Ba
sed
on
(
16)
,
the
sta
te
sp
ace
re
pr
es
entat
ion
(
17)
a
nd
(
18
)
are r
e
w
ritt
en
a
s
in
(19
)
a
nd (2
0)
[
1
].
[
0
0
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
1
0
0
]
[
̇
̇
ʓ
]
+
[
0
0
0
0
−
0
0
0
0
0
0
]
[
]
(19
)
[
]
=
[
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
0
]
[
̇
̇
ʓ
]
+
[
0
0
0
0
0
0
]
[
]
(20)
[
̇
̇
̈
̈
̇
ʓ
̇
̇
]
=
[
0
0
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
.
0655
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
1
0
0
]
[
̇
̇
ʓ
]
+
[
0
0
0
0
5
.
8197
5
.
8197
63
.
9498
−
63
.
9498
0
0
0
0
0
0
]
[
]
(21)
In
t
his
stu
dy,
f
or
pur
pose
of
con
t
ro
l
syst
em
d
esi
gn,
the
m
od
el
of
the
sys
tem
is
fo
rm
ulate
d
in
sta
te
sp
ace
f
or
m
us
ing
the
phys
ic
al
par
am
et
er
s
values
sta
te
d
in
T
a
ble
2
[1
]
.
Ba
se
d
on
(19
)
an
d
us
in
g
the p
a
ram
et
ers
values
in
Ta
ble
1
,
the
sta
te
equati
on of
t
he
s
yst
e
m
is g
iven by (
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.
10
, No
.
2
,
A
pr
i
l 202
0
:
1868
-
1884
1874
Table
2.
Value
s of
ph
ysi
cal
pa
ram
et
ers
of
3DO
F
helic
op
te
r
syst
em
Sy
m
b
o
l
Ph
y
sical Un
it
Nu
m
eric
al Va
lu
es
k
g
.
m
2
1
.81
4
5
k
g
.
m
2
1
.81
4
5
k
g
.
m
2
0
.03
1
9
W
N
4
.25
9
1
m
0
.88
m
0
.35
m
0
.17
N/V
12
3.
CONTR
OLL
ER
TE
C
HNIQUES
In
this
sect
io
n,
the
theo
ry
of
t
he
PID
an
d
L
Q
R
con
tr
oller
te
chn
i
qu
e
s
us
e
d
for
desi
gn
i
ng
t
he
pro
posed
helic
op
te
r
c
ontrol
syst
em
a
re
intr
od
uced
.
The
gai
n
pa
ram
et
ers
of
the
P
ID
co
nt
ro
ll
er
are
obt
ai
ned
appr
ox
im
at
ely
from
the
LQR
gain
m
a
trix.
Con
tr
ol
eq
uations
of
el
evati
on
an
gle,
pitch
ang
le
,
a
nd
tra
ve
l
rate
for hel
ic
op
te
r
s
yst
e
m
u
sing
PID c
on
t
ro
ll
er a
r
e also
pr
e
sente
d
.
3.1
.
PI
D
c
ont
rolle
r
A
PID
is
the
m
os
t
po
pu
la
r
con
t
ro
ll
er
te
ch
ni
qu
e
that
is
wi
dely
us
ed
in
i
ndus
t
rial
app
li
c
at
ion
s
due
to
the sim
plicity
of
it
s
st
ru
ct
ur
e
and can
b
e
r
eal
iz
ed
easi
ly
f
or
var
i
ou
s
contr
ol
p
r
oble
m
s as the
gain
p
a
ram
eter
s
of
the
co
ntr
oller
a
re
relat
ively
in
dep
e
ndent
[
21
]
[
22
]
.
Ba
sic
al
ly
,
the
c
ontr
oller
pro
vid
es
co
ntr
ol
com
m
and
si
gn
al
s
(
)
base
d
on
the
e
rror
(
)
betwee
n
t
he
dem
and
in
put
a
nd
the
act
ua
l
outp
ut
of
t
he
syst
em
.
The
con
ti
nu
ous
tim
e stru
ct
ur
e
of the
PID c
on
t
ro
ll
er is
as
fo
ll
ow
s
:
(
)
=
(
)
+
∫
(
)
0
+
(
)
(22)
Wh
e
re
,
an
d
are
the
pro
port
ion
al
,
i
nteg
ral
an
d
diff
e
re
ntial
com
po
nen
t
s
of
the
co
ntr
oller
gain
.
These
co
ntr
oller
gain
pa
ram
et
ers
sh
ould
be
tun
ed
prop
e
rly
to
enab
le
the
ou
t
pu
t
sta
te
s
of
t
he
syst
e
m
to
eff
ic
ie
ntly
f
ollow
the
desire
d
input.
If
,
an
d
are
the
desire
d
el
evati
on
an
gl
e,
pitch
an
gle
and
tra
vel
rate
of
the
helic
op
te
r
syst
em
,
it
can
ex
press
t
he
form
of
P
ID
co
ntr
ollers
use
d
to
m
eet
the
desir
e
d
ou
t
pu
t
sta
te
s
a
s foll
ow
s
[1,
2
3
]
. In
t
his stu
dy
, f
or ele
vation an
gle, the
cont
ro
l e
quat
ion i
s
base
d on
t
he fo
ll
ow
in
g
P
ID co
ntr
ol
equ
at
io
n:
=
+
̇
+
∫
(23a)
=
(
−
)
−
̇
+
∫
(
−
)
(
23
b)
Wh
il
e th
e
pitch a
ng
le
is
contr
olled
by the
foll
ow
in
g PD
con
trol e
qu
at
io
n:
=
+
̇
(24a)
=
(
−
)
−
̇
(24
b)
The
tra
vel
rate
is g
ov
e
rn
e
d by
the foll
owin
g PI
c
o
ntro
l e
qua
ti
on
:
=
+
∫
(25a)
=
(
−
)
+
∫
(
−
)
(25
b)
wh
e
re
=
−
,
=
−
,
=
−
,
̇
=
−
̇
an
d
̇
=
−
̇
.
3.2
.
LQ
R
c
ontrolle
r
Linear
quad
rat
ic
regulat
or
is
a
c
omm
on
op
ti
m
al
con
trol
te
c
hn
i
qu
e
w
hich
has
been
wi
dely
util
iz
ed
in
var
i
ou
s
m
anip
ulator
syst
em
s
[24]
.
Using
L
QR
te
chn
i
qu
e
in
desig
n
co
ntr
ol
syst
e
m
req
ui
res
al
l
the
states
of
the
syst
e
m
plant
to
be
m
easur
a
ble
as
it
ba
ses
on
the
fu
ll
sta
te
feedbac
k
co
nce
pt.
T
he
refor
e
,
us
i
ng
LQR
con
t
ro
ll
er
t
o
st
abili
ze
the
3D
OF
helic
opte
r
syst
e
m
based
on
t
he
ass
um
ption
t
hat
the
sta
te
s
of
t
he
syst
em
are
consi
der
e
d
m
easur
a
ble.
LQR
appr
oach incl
udes a
pply
ing
t
he
optim
al
co
ntr
ol effo
rt:
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
El
ec &
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ele
vation,
pitch
and
tr
avel
ax
is st
ab
il
izati
on
of 3D
OF heli
cop
te
r
wi
th
…
(
I
br
ahim
K.
Mo
ham
me
d
)
1875
(
)
=
−
(
)
(26)
Wh
e
re
is
t
he
s
ta
te
feedbac
k
ga
in
m
at
rix
of
L
QR
c
on
t
ro
ll
er,
that
will
ena
bl
e
the
outp
ut
st
at
es
of
the
syst
e
m
to foll
ow the
tr
ajecto
ries
of
re
fer
e
nce inp
ut, while m
ini
m
izi
ng the
f
ollow
i
ng the
cost
fun
ct
ion
:
=
∫
(
(
)
(
)
∞
0
−
(
)
(
)
)
(27)
Wh
e
re
Q
a
nd
R
are
ref
e
rr
e
d
to
as
weig
htin
g
sta
te
and
c
ontr
ol
m
at
rices.
The
co
ntr
oller
feedback
gain
m
at
rix
can
be dete
rm
i
ned b
y
us
i
ng bel
ow
e
quat
io
n:
=
−
1
(28)
Wh
e
re
is (
nxn) m
at
rix
deter
m
ined
f
ro
m
t
he
so
l
ution o
f
th
e f
ollow
i
ng Ri
ccat
i
m
at
rix
eq
uation:
+
−
−
1
+
=
0
(29)
In this a
pp
li
cat
ion
,
=
[
k
11
k
12
k
13
k
14
k
15
k
16
k
17
k
21
k
22
k
23
k
24
k
25
k
26
k
27
]
and
(
)
=
[
1
2
]
.
Ba
sed on t
his,
the contr
ol e
ffor
t
of the
syst
em
sta
te
d
in
(26
)
ca
n be
wr
it
te
n
as
foll
ows:
[
1
2
]
=
−
[
k
11
k
12
k
13
k
14
k
15
k
16
k
17
k
21
k
22
k
23
k
24
k
25
k
26
k
27
]
[
̇
̇
ʓ
]
(30)
Fo
r
the
purp
ose
of sim
plici
t
y of co
ntr
ol pr
oblem
the w
ei
ght
ing
m
at
rices Q a
nd R are
chos
en
as
the
diag
onal
m
at
rices so
tha
t t
he
co
st f
unct
ion
(
27)
ca
n be
r
ef
orm
ulate
d
as b
el
ow
:
=
∫
(
11
1
2
+
22
2
2
+
33
3
2
+
44
4
2
+
55
5
2
+
66
6
2
+
77
7
2
+
11
1
2
+
22
2
2
)
∞
0
(31)
Wh
e
re
11
,
22
,
33
,
44
,
55
,
66
and
77
denote
the
weig
hting
el
e
m
ents
of
flig
ht
an
gles
an
d
their
corres
pondin
g
angular
velocit
ie
s
of
t
he
pro
pose
d
3DOF
helic
op
te
r
syst
em
resp
ect
ively
,
w
hile,
11
and
22
ar
e
the
wei
gh
ti
ng
el
e
m
ents
of
c
ontr
ol
inputs.
It
is
worth
c
ons
iderin
g
that,
in
this
stu
dy,
ba
sed
on
t
he
sta
te
an
d
con
t
ro
l
L
QR
weig
hting
m
atr
ic
es
Q
an
d
R
,
the
fee
db
ac
k
gain
m
at
rix
(
)
ca
n
be
cal
c
ulate
d
by
util
iz
ing
the Mat
la
b
c
om
m
and
“
lqr
”.
The
optim
al
con
t
ro
l
a
ppr
oac
h
L
QR
is
hi
ghly
rec
omm
en
ded
f
or
sta
bili
zi
ng
the
3DO
F
helic
op
te
r
syst
e
m
as
it
basical
ly
loo
ks
f
or
a
com
pr
om
i
sing
bet
ween
t
he
best
co
ntr
ol
per
f
or
m
ance
and
m
ini
m
u
m
c
on
t
rol
effor
t.
Ba
sed
on
L
QR
c
on
t
ro
l
le
r,
a
n
op
ti
m
um
trackin
g
perform
ance
can
be
in
vestigat
e
d
by
a
pro
per
se
tt
ing
of
the
fee
db
a
ck
c
on
t
ro
ll
er
gain
m
at
rix.
To
ach
ie
ve
this,
t
he
LQR
c
on
tr
olle
r
is
o
ptim
iz
ed
by
us
i
ng
G
A
t
un
i
ng
m
et
ho
d w
hich i
s adop
te
d
to
obtai
n
a
n op
ti
m
um
ele
m
ents v
al
ues
f
or
Q
a
nd R w
ei
gh
ti
ng
m
at
rices.
4.
TUNING
M
E
THOD
In
this
st
ud
y,
GA
t
un
i
ng
a
pp
ro
ac
h
is
ad
opte
d
to
op
ti
m
iz
e
the
LQR
gain
m
at
rix
us
ed
t
o
appr
ox
im
at
e
the
gai
n
pa
ra
m
et
ers
of
P
I
D
con
t
ro
ll
er
for
3DOF
helic
opte
r
syst
em
.
GA
is
a
glob
al
search
opti
m
iz
at
ion
te
chn
iq
ue
base
s
on
the
strat
egy
of
gen
et
ic
s
an
d
natural
sel
ect
ion
[2
5][
26
]
.
This
opti
m
iz
at
ion
m
et
ho
d
is
util
iz
ed
to
ob
ta
in
an
opti
m
u
m
glo
bal
s
olu
ti
on
f
or
m
or
e
cont
ro
l
an
d
m
anipu
la
ti
ng
prob
le
m
s.
The
pr
oce
dure
of
GA
ap
proac
h
include
s
of
th
ree
basic
ste
pe
s:
Sele
ct
ion
,
Cros
s
over
an
d
Muta
ti
on.
A
pp
ly
in
g
t
hese
m
ai
n
op
e
rati
ons
cre
at
es
new
in
divi
du
al
s
w
hich
c
ou
l
d
be
bette
r
than
thei
r
pa
re
nts.
Ba
sed
on
the
requirem
ents
of
desire
d
respo
nse
,
the
se
qu
e
nc
e
of
GA
opti
m
i
zat
ion
te
c
hn
i
que
is
re
peated
f
or
m
any
it
eratio
ns
a
nd
fin
al
ly
sto
ps
at
gen
e
rati
ng
optim
u
m
so
luti
on
el
em
ents
for
the
ap
plica
ti
on
pro
blem
s
[
27
]
.
T
he
se
qu
e
nce
of
the
GA
tun
i
ng
m
et
ho
d
is
pres
ented
i
n
Fi
gure
5
[
3]. T
he
ste
ps o
f
the
GA l
oop are
d
e
fine
d as f
ollows:
1.
In
it
ia
l set
of
popu
la
ti
on.
2.
Sele
ct
ing
i
nd
i
vi
du
al
s
for
m
at
i
ng.
3.
Ma
ti
ng
the
po
pula
ti
on
t
o
c
reate p
roge
ny.
4.
Muta
te
p
r
ogen
y.
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.
10
, No
.
2
,
A
pr
i
l 202
0
:
1868
-
1884
1876
5.
In
se
rtin
g new
gen
e
rated
in
div
id
uals int
o popu
la
ti
on.
6.
Ar
e
the syste
m
f
it
ness
fun
ct
io
n
sat
isfie
d
?
7.
En
d
sea
rch p
ro
cess f
or so
l
utio
n.
In
this
st
udy,
t
he
ai
m
of
us
i
ng
GA
op
ti
m
iz
a
ti
on
m
et
ho
d
is
to
tu
ne
t
he
PID
c
ontrolle
r
pa
ram
et
ers
by
m
ini
m
iz
ing
a
sel
ect
ed
fitness
functi
on
w
hich
will
be
discusse
d
in
detai
l
in
the
sect
ion
4.
The
im
ple
m
ent
at
ion
proce
dure
of
the
GA
tu
ning
m
et
ho
d
be
gin
s
with
t
he
def
i
ni
ti
on
ste
p
of
t
he
ch
ro
m
os
om
e
re
pr
ese
ntati
on.
Eac
h
chrom
os
om
e
i
s
rep
re
sente
d
by
nin
e
cel
ls
wh
ic
h
corres
pond
to
the
we
igh
t
m
a
tric
es
el
e
m
ents
of
the
LQR
con
t
ro
ll
er
as
s
how
n
in
Fi
gur
e
6.
By
this
re
pr
ese
ntati
on
it
can
a
dju
st
the
LQR
e
le
m
ent
s
in
orde
r
to
a
chiev
e
the
re
qu
ire
d
pe
rfor
m
ance.
Th
ese
cel
ls
are
f
orm
ed
by
real
posit
ive
num
bers
an
d
cha
racter
iz
e
the
ind
i
vidual
to
be
e
valuate
d
[
27]
.
Figure
5. Proce
ss lo
op of
GA.
Figure
6. De
fin
it
ion
of
G
A
c
hrom
os
om
e
5.
HEL
ICO
PTE
R
H
YBR
ID
CO
NTR
OLS
Y
STE
M DESI
GN
In
t
his
stu
dy,
a
co
ntr
ol
syst
e
m
us
in
g
L
Q
R
base
d
P
ID
con
t
ro
ll
er
is
de
sign
e
d
t
o
sta
bili
ze
3DOF
helic
op
te
r
syst
e
m
.
B
ased
on
s
te
p
input,
a
hybr
i
d
co
ntro
ll
er
is
design
e
d
f
or
the
fo
ll
owin
g
desire
d
pe
rform
ance
par
am
et
ers:
ri
se
ti
m
e
(
)
le
s
s
tha
n
10
(m
s),
set
tl
ing
tim
e
(
)
le
ss
t
han
30
(m
s),
m
axim
u
m
ov
ers
hoot
per
ce
ntage
,
(
)
le
ss
than
5%
.
T
he
bl
ock
dia
gra
m
of
the
pro
pose
d
helic
opte
r
con
t
ro
l
syst
e
m
based
on
L
QR
con
t
ro
ll
er is
pr
esented
in
Fig
ure
7.
Figure
7. LQR
co
nt
ro
ll
er
base
d on G
A for
3DO
F
helic
op
te
r
syst
em
11
q
22
q
33
q
44
q
55
q
66
q
77
q
11
r
22
r
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
El
ec &
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ele
vation,
pitch
and
tr
avel
ax
is st
ab
il
izati
on
of 3D
OF heli
cop
te
r
wi
th
…
(
I
br
ahim
K.
Mo
ham
me
d
)
1877
The
c
on
tr
ol
sy
stem
is
analy
s
ed
m
at
he
m
atic
al
ly
and
the
n
si
m
ulate
d
us
in
g
Ma
tl
ab/sim
u
li
nk
to
ol
to
validat
e
the
propose
d
LQR
c
on
t
ro
ll
er.
Ba
se
d
on
the
desire
d
perf
or
m
ance
par
am
et
ers
wh
ic
h
inclu
de
ri
se
an
d
set
tl
ing
tim
e,
ov
e
rs
hoot
a
nd
erro
r
ste
ady
sta
te
par
am
et
e
rs
the
fitnes
s
functi
on
of
t
he
co
ntr
ol
pro
bl
e
m
is
form
ulate
d
as foll
ows:
=
0
.
3
.
+
0
.
3
.
+
0
.
2
.
+
0
.
2
.
(32)
Wh
e
re,
is
cl
ose
d
l
oop
t
ran
s
f
er
f
unct
io
n
of
the
helic
op
te
r
syst
e
m
,
.
,
,
,
are
th
e
rise
ti
m
e,
set
tl
ing
tim
e,
m
axi
m
um
ov
er
s
hoot
and
e
rro
r
ste
ad
y
sta
te
of
the
cl
os
ed
-
lo
op
c
ontr
ol
syst
e
m
.
It
is
wo
rt
h
co
nsi
der
in
g
that
the
con
tr
ol
inp
ut
effo
rt
is
con
si
der
e
d
in
the
evaluati
on
process
of
the
pro
po
se
d
s
ta
bili
zi
ng
helic
op
te
r
syst
e
m
.
In
this
stu
dy,
t
he
des
ign
of
c
ontr
oller
is
ef
fecti
vely
optim
iz
ed
by
us
i
ng
GA
tu
ni
ng
m
et
ho
d
w
hi
ch
is
adopted
to
obt
ai
n
op
ti
m
u
m
ele
m
ents
value
s
for
L
QR
weig
hting
m
at
rices
Q
a
nd
R.
Th
es
e
opti
m
iz
ed
m
at
rices
are
us
in
g
to
ca
lc
utate
op
ti
m
u
m
con
trolle
r
ga
in
m
at
rix
by
us
in
g
(
29)
an
d
(3
0).
H
oweve
r,
the
gain
m
atr
ix
is
determ
ined
by
us
in
g
t
he
Ma
tl
ab
c
omm
and
‘
lqr
’.
5.1
.
PI
D appr
ox
im
at
i
on
In
t
his
subsect
ion,
the
gain
pa
ram
et
ers
,
,
of
t
he
PI
D
co
ntr
oller
are
cal
culat
ed
appr
ox
im
at
ely
from
the
feed
ba
ck
gai
n
m
at
ri
x
of
the
L
QR
con
t
ro
ll
er
base
d
on
GA
optim
iz
at
ion
te
ch
nique
.
An
al
yz
in
g
(30)
yi
el
ds
th
e f
ollo
wing
[4
]
:
[
1
2
]
=
−
[
k
11
k
12
k
13
k
14
k
15
k
16
k
17
k
11
−
k
12
k
13
−
k
14
−
k
15
k
16
−
k
17
]
[
̇
̇
ʓ
]
(33)
5.1.1.
El
e
vatio
n contr
ol us
in
g
PI
D
c
ontr
oller
Su
m
m
ing
the
row
s
of
(33
)
r
es
ults the
fo
ll
ow
i
ng [1]:
1
+
2
=
−
(
2
11
+
2
13
̇
+
2
16
ʓ
)
=
−
(
2
11
+
2
13
̇
+
2
16
∫
)
(34)
The
a
bove
equ
at
ion
ca
n be
wri
tt
en
as
=
−
2
11
(
−
)
−
2
13
̇
−
2
1
6
∫
(
−
)
(35)
It
is
ob
vi
ous
that
(23
b)
an
d
(35)
hav
e
the
sam
e
structur
e
,
this
m
eans
that
the
gain
para
m
et
ers
of
pitch
P
I
D
con
t
ro
ll
er
can
b
e o
btai
ned
f
rom
the
gai
n
el
e
m
ents
of
L
QR co
nt
ro
ll
er.
T
hus,
c
om
par
ing
(
23b) w
it
h
(
35)
,
yi
el
ds
the foll
owin
g g
ai
n
relat
io
nships:
=
2
11
=
2
13
=
2
16
(36)
The
blo
c
k
dia
gr
am
of
cl
os
ed
-
l
oop
co
ntr
ol
syst
e
m
fo
r
3DO
F
helic
opte
r
syst
e
m
base
d
on
PID
con
t
ro
ll
er
i
s
sh
ow
n
in
Fi
gur
e 8
.
Ta
king La
place t
ra
ns
f
orm
f
or
elevat
io
n axi
s m
od
el
(1
3) yi
el
ds
the
f
ol
lowing e
qu
at
i
on
:
(
)
.
2
=
1
(
)
(37)
The
tra
nsfer
fu
nction o
f
t
he
el
evati
on ax
is
p
l
ant is
giv
e
n by:
(
)
(
)
=
2
(38)
The
tra
nsfer
fu
nction o
f
t
he
P
ID
co
n
tr
oller is
as foll
ows:
(
)
(
)
=
2
+
+
(39)
Evaluation Warning : The document was created with Spire.PDF for Python.