TELKOM
NIKA
, Vol.13, No
.3, Septembe
r 2015, pp. 9
30~939
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v13i3.1862
930
Re
cei
v
ed Ap
ril 15, 2015; Revi
sed
Jul
y
9, 2015; Accept
ed Jul
y
22, 2
015
Waypoint Navigation of AR.Drone Quadrotor Using
Fuzzy Logic Controller
Veronica Indra
w
ati*
1
, Ag
u
ng Pra
y
itno
2
, Thomas Ard
i
Kusuma
3
Electrical E
ngi
neer
ing D
e
p
a
rtment, Univers
i
ty
of Sura
ba
y
a
(
U
BAYA)
Jl. Ra
ya Kal
i
ru
ngkut – Sur
aba
ya
602
93, Ea
st
Java – Indo
ne
sia, T
e
l.+
62-31-298
115
7
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: veronic
a
@st
a
ff.uba
ya.ac.i
d
1
, pra
y
itn
o_a
gu
ng@staff.uba
ya.ac.id
2
,
blackr
ogu
3@g
m
ail.com
3
A
b
st
r
a
ct
In
th
i
s
pa
pe
r, a
n
AR
.D
ro
ne
is fl
o
w
n
a
u
t
on
om
ou
sl
y fro
m
t
h
e in
itial
pos
itio
n (x,y,
z
) to th
e
desir
e
d
positi
on cal
l
e
d
w
a
ypoint usi
n
g
F
u
zz
y
Lo
gic C
ontrol
l
er
(F
LC). T
he F
L
C consi
s
ts of three control lo
ops w
h
i
c
h
are
pitch c
ontr
o
l l
o
o
p
, rol
l
co
n
t
rol lo
op
an
d v
e
rtical r
a
te c
o
n
t
rol lo
op. F
o
r
e
a
ch c
ontrol
lo
o
p
, des
ired
p
o
sit
i
on
and r
eal
pos
iti
on ar
e us
ed a
s
inp
u
ts of the
F
L
C, w
h
ile
p
i
tch, roll a
nd v
e
rtical r
a
te are
used
as o
u
tp
ut
respectiv
e
ly. T
he
alg
o
rith
m is
realis
ed
in
thre
e fli
ght sc
he
me
s an
d th
e
navi
g
ation
d
a
ta is
re
corde
d
. T
h
e
fir
s
t
flight sche
m
e: a desir
ed x-p
o
s
ition
of AR.Dr
one w
ill
be re
a
c
hed first foll
o
w
ed by a des
ir
ed y-pos
itio
n, and
lastly a d
e
sire
d
z
-
positi
on. The seco
nd fli
g
ht sche
m
e: a
desir
ed x-p
o
siti
on an
d y-p
o
siti
on w
ill b
e
reac
he
d
simulta
neo
usly
follow
ed
by a
desir
ed
z
-
p
o
siti
on. T
he thir
d fli
ght sche
m
e: AR.Dron
e
flies t
o
w
a
rds to des
i
r
ed
positi
on s
i
mult
ane
ously. T
h
e
results
s
how
that the A
R
.Dron
e
ca
n re
ach th
e w
a
yp
oint w
i
th th
e thr
e
e
sche
m
es w
e
ll.
How
e
ver, the flight sch
e
m
e strai
ght tow
a
rds the w
a
ypo
i
nt w
i
th the F
L
C w
o
rki
n
g
simulta
neo
usly
is the most sat
i
sfying o
ne.
Ke
y
w
ords
: w
a
ypoi
nt navi
gati
on, AR.Dron
e
control, fu
zz
y
l
ogic co
ntroll
er
Copy
right
©
2015 Un
ive
r
sita
s Ah
mad
Dah
l
an
. All rig
h
t
s r
ese
rved
.
1. Introduc
tion
No
wad
a
ys, q
uadrotor i
s
n
o
t only u
s
ed
as a
ho
bby, but it ha
s al
so bee
n
widel
y used
for
variou
s activities, su
ch
a
s
news cove
ra
ge
in
the
affe
cted
area
s, traffic coverag
e
, the
sh
ootin
g of
a region,
pro
m
otional
eve
n
ts a
n
d
several
other
sh
o
w
s.
Gen
e
rall
y, this q
uad
rotor i
s
still fl
own
manually by usin
g remote
control. Research
at the university has develop
ed
a wide rang
e
of
c
ontrollers that c
a
n fly quadrotor automatic
a
lly.
Vari
o
u
s
kin
d
s of
co
ntrolle
rs have
bee
n d
e
si
gn
ed,
among oth
e
rs for tra
cki
ng
an obje
c
t, flying throug
h obsta
cle
s
, de
terminin
g formation-flig
ht and
tracking
the t
r
aje
c
tory. Th
e devel
opme
n
t of the va
ri
ous alg
o
rith
ms i
s
o
ne
of most i
n
teresting
fields of rese
arch in mo
st
of the leadin
g
univers
ities
worl
dwi
de. T
he develo
p
m
ent will be fa
ster if
the quad
roto
r is rea
d
y in hard
w
a
r
e.
One of
the most comm
only used q
uadrotors is the
AR.Drone.
AR.Drone i
s
a quad
rotor
made by Parrot, a Fr
en
ch
compa
n
y. At first, the AR.Dro
ne is
made a
s
a to
y for the sake of ente
r
tain
ment, whi
c
h
can
be pl
aye
d
with a
ppli
c
ations i
n
stall
e
d in
Androi
d devi
c
e
s
an
d iOS
device
s
th
ro
ugh
Wi-Fi.
AR.Drone
ha
s alre
ady ha
d
seve
ral
sen
s
ors,
su
ch a
s
: 3 axis accele
ro
meter, 3 axis gyrosc
op
e, a son
a
r altim
e
ter, and the
front and bo
ttom
came
ra
s. Mo
reove
r
, this
dron
e i
s
e
qui
pped
with
an
onb
oard
co
mputer that
can
be
u
s
ed
for
vertical take off, landing, hovering, and vi
deo streami
ng from two
cameras via Wi-Fi [1].
Parrot
ha
s
al
so
rele
ased
a
n
official
SDK
[2] that
can
help users
to
acce
ss
the
i
n
nerb
o
a
r
d of
the
AR.Drone. When the AR.
D
ro
ne is turn
ed on, the i
nnerb
o
a
r
d will
automaticall
y
act as a se
rver
whi
c
h i
s
com
p
lemented by
the facilities
of Dynami
c
Host Configuration
Protocol (DHCP
),
so
t
hat
use
r
s can
co
nne
ct to the
AR.Drone
wit
hout havin
g
to set
up a
n
Internet P
r
oto
c
ol
(IP) on th
eir
computers. By using the innerb
oard, users can control the main
flight (take-off, hovering,
landin
g
, and
emergen
cy st
op) a
nd ma
n
oeuvre th
e fli
ght by giving value within t
he ra
nge of -1 to
1 in th
e pit
c
h, roll, ya
w
rate an
d ve
rtical
rate
inpu
t. A value of
-1
an
d 1
wi
ll rep
r
e
s
e
n
t the
minimum an
d maximum
value of each input wh
ose value ca
n
be set from
the innerbo
ard
config
uratio
n. The value in
dicate
s the a
ngel
s pitch
a
ngle, roll, yaw rate an
d verti
c
al rate that a
r
e
prop
ortio
nal
with the mi
ni
mum an
d ma
ximum ra
nge
. Positive an
d neg
ative value
s
indicat
e
the
dire
ction
s
. Positive value
s
(+) in the pit
c
h
ca
u
s
e the
dron
e to mo
ve backward, while n
egati
v
e
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Wa
ypoi
nt Na
vigatio
n of AR.Drone Q
u
a
d
roto
r Usi
ng
Fuzzy Logi
c Controlle
r (V
eroni
ca Ind
r
a
w
ati)
931
values
(-) ca
u
s
e the d
r
on
e
to fly forward.
To mano
euv
re to the ri
ght
, roll input is
given a po
siti
ve
value, while to the left means giving th
e input
roll a
negative value. To mano
euvre the piv
o
t
clo
c
kwi
s
e, ya
w
rate i
nput
is give
n a
p
o
sitive valu
e, and
counte
r
-clo
ckwi
se
m
ean
s givin
g
t
h
e
input ya
w
rat
e
a
neg
ative
value. To
fly up, the ve
rt
ical rate i
s
giv
en a
po
sitive
value,
while
to
move do
wn, t
he verti
c
al
rat
e
is
given
a n
egative value
.
The point is this
, that to
c
ontrol the AR.
Dro
ne i
s
to
send
comm
an
ds to th
e inn
e
rbo
a
rd an
d
receive n
a
vig
a
tion data
(NavData) from
the
innerbao
rd via Wi-Fi. The
s
e comm
and
s are in the fo
rm of pitch, roll, yaw rate and vertical rate,
while the
Na
v–Data
s are
in the form o
f
actual
pit
c
h
value, forwa
r
d spee
d, actual roll valu
e,
side
wa
rd spe
ed, actual ya
w rate value,
yaw
value, vertical rate value and altitud
e
value
Figure 1. AR.Dro
ne send
comman
d
s
a
n
d
receive the navigation d
a
t
a
Based
on
thi
s
d
e
scription,
the A
R
.Dron
e
is ch
o
s
en
to be
the
platform
of this re
sea
r
ch.
The type of
AR.Drone u
s
ed in this re
search is
the
AR.Drone 2.
0 Elite Editio
n which ha
s the
followin
g
spe
c
ificatio
ns: 4
inrun
n
e
r
brushless
m
o
tors. 14.5W
28,5
00 RPM, 3
2
-bit ARM Cort
ex
A8 1G
Hz
pro
c
e
s
sor
with 8
00MHz
DSP
TMS320
DM
C64x video, 1
G
B DDR2 RAM at 200M
Hz, 3
-
axis gyro
sco
pes
200
0°/se
c
on
d preci
s
io
n, three ax
i
s
accele
rom
e
te
rs +/-
5
0
mg pre
c
isi
on,
three-
axis ma
gneto
m
eters 6
°
p
r
e
c
isi
on, Pressure
+/- 10
Pa
preci
s
ion
se
nso
r
s,
ultra
s
o
und
se
nsors f
o
r
measurement
of groun
d
altitude, 6
0
FPS
QV
GA vertical
gro
und
sp
eed
cam
e
ra
s for
measurement
, Linux 2.6.32, USB 2.0 high spee
d
for e
x
tension
s
, Wi
-Fi, HD Cam
e
ra. 720
p 30fps
[3].
Several
studi
es
usi
ng th
e
AR.Drone
a
s
a pl
atform
a
m
ong
othe
rs
are
de
scrib
e
d in thi
s
se
ction. Pierre-Jean
Brist
e
au et al. [4]
explained
in
detail that th
e navigatio
n
techn
o
logy a
n
d
control used
in the AR.Dro
ne in
clud
e the har
d
w
are de
scripti
on, vision al
gorithm, sen
s
or
calib
ration, al
titude estimat
i
on, velocity estimation
a
n
d
control architecture.
Nick et al. [5] made
an A
R
.Drone
sim
u
lation
wi
th the
sen
s
o
r
and
motio
n
model
s. Th
ey also m
ade
a
visual
ma
p a
n
d
indoo
r envi
r
o
n
ment. Using
the visual m
ap, the
AR.d
rone
can l
o
cal
i
se itself. Michael Mo
gen
son
[1] made a
n
AR.Drone
La
bVIEW toolki
t to contro
l the AR.Dro
ne
1.0. Broadly
spe
a
ki
ng, th
is
softwa
r
e
con
s
ist
s
of seve
ral Virtual Inst
rume
nts (VI)
whi
c
h a
r
e the
main VI, vid
eo VI, NavDa
t
a
VI, supportin
g
VI’s and ad
ditional VI
’s. This softwa
r
e
is made to make it ea
sie
r
for re
sea
r
ch
ers
and tea
c
he
rs to learn a
b
out AR.Dron
e
. Krajnik
et al. [6] creat
ed a mod
e
l
stru
cture of the
AR.Drone
which
co
nsi
s
ts of 4 mod
e
l
s
: pitch,
roll,
yaw rate a
nd verti
c
al rate. The m
o
del
para
m
eter i
s
earn
ed from the estimatio
n
result
usi
ng the data from the experime
n
t. Agung et
al.
[7] impleme
n
ted fu
zzy l
ogi
c co
ntrolle
r i
n
t
he A
R
.Dr
one
2.0 for the t
r
a
j
ectory
tra
cki
ng a
ppli
c
atio
n.
Some form
s
of traje
c
tory t
r
ackin
g
h
a
ve
been
su
cce
s
sfully followe
d by the d
r
on
e. Sarah
Yifang
[8] obtained the dynami
cs
model of AR.Dro
ne that co
nsi
s
ts of internal cont
rolle
r model an
d the
physi
cal dyn
a
mics of the
dron
e. Som
e
co
ntrolle
r
algorith
m
is
applie
d to th
e dro
ne, such as
waypoi
nt navigation a
nd trajecto
ry, following
with
PID controlle
r an
d also vision
-based
control
l
er
for a variety of flight formation. Rab
ah
Abbas
et al. [9] prop
osed
a PID cont
rol
l
er an
d dire
ct
ed
lyapunov
con
t
roller fo
r formation tra
c
ki
ng of qu
adro
t
ors. PID
con
t
roller i
s
impl
emented
on t
he
leade
r qu
adrotor, whil
e directed
Lyapun
ov contro
ller
on the follo
wers.
Dynami
c
s optimi
s
atio
n of
the para
m
ete
r
s
controllers
is achieved u
s
ing a
n
ar
tificial fish swarm algorithm.
Emad Abba
si
et
al. [10] simul
a
ted two
cont
rol sch
e
me
s to cont
ro
l the
height of the
quad
roto
r. Th
e first sch
e
m
e
use
s
4
PID controlle
rs whi
c
h are
then
simulated usi
n
g
turb
ulen
ce
sign
al.
The
seco
nd schem
e
use
s
co
mbin
ation
fu
zzyPI
D cont
rolle
r whi
c
h are
al
so simul
a
ted
usi
ng
th
e same
tu
rbul
en
ce
sign
al. The result of the simulation sho
w
s that
the fuzzyPID com
b
ination is m
o
re suitable
with
the turbule
n
ce situation. Abba
si et al. [11]
comp
are
d
the classi
cal PID contro
ller and the fuzzy
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 3, September 20
15 : 930 – 939
932
sup
e
rvisory
controlle
r fo
r t
uning
the PID co
ntrolle
r to
stabili
se th
e q
uadrotor mod
e
lled
with Eul
e
r-
Ne
wton
equ
a
t
ion. The
resu
lt of the
simul
a
tion
sho
w
s
t
hat fuzz
yPID
is
better than
PID in the
c
a
s
e
of eliminating
overshoot an
d sho
r
teni
ng the settling ti
me. Santos e
t
al. [12] simulated fuzzy logi
c
to control the
model of the quad
roto
r. The input
is
the height, roll, pitch and
yaw value; the
output is the power of ea
ch of the four rotors.
The re
sult of the simulation sho
w
s the efficie
n
cy
from the co
ntrol strategy. Senthil Kuma
r et al. [
13] simulated fu
zzy logic to con
t
rol the mode
l of
the qua
droto
r
usin
g Fu
zzy
Logi
c Tool
bo
x Matlab. Th
e fuzzy used
has
3 input
s, whi
c
h a
r
e e
r
ror
(the differe
nce betwe
en th
e desi
r
ed val
ue and the
p
r
esent value
)
, derivative erro
r an
d integ
r
al
error. The o
u
tput is the co
n
t
rol value po
wer of ea
ch
motor.
Waypoi
nt navigation is a n
e
w te
chnolo
g
y
that allows for the dro
n
e
s
to fly from one point
to another.
With this techn
o
logy, the drone
s ca
n fl
y
at a certain h
e
ight, at a ce
rtain sp
eed,
with
certai
n fly p
a
tterns a
nd
hover at the
de
stinat
ion
point with
th
e
remote co
ntrol navigati
o
n
software. In the future thi
s
technology
will be
very helpful, especially
for busi
ness and social
missi
on
s. Fo
r example, it
can
be u
s
e
d
in the d
e
live
r
y of goo
ds f
o
r b
u
si
ness
or h
u
manita
ri
an
missi
on
s in d
i
sa
ster a
r
ea
s.
This te
chnol
ogy ty
pically utilise
s
GPS and a ma
p o
n
the co
mput
er
scree
n
for mo
nitoring a
nd control.
In this p
ape
r, waypoint
n
a
vigation te
c
hnolo
g
y will
be impl
emen
ted in the l
a
borato
r
y
using AR.
D
rone as a platf
o
rm. The AR.Drone
will
be designed t
o
fly from the initial positi
on
(x,
y
,z)
to
a de
s
i
re
d
w
a
ypoin
t
(x
des
,y
des
,z
des
)
with va
ri
ous sch
e
me
s. The
algo
rithm of th
e fu
zzy
logic
controll
er
will be used for
rem
o
te control navigation whic
h is
reali
s
ed using LabVIEW
softwa
r
e. The
implementati
on of the fuzzy algor
ithm
use
d
for co
ntrolling the A
R
.Drone h
a
s
not
yet been
d
o
ne by
many
resea
r
chers.
Therefore,
th
e fuzzy
co
ntrol sch
e
me
fo
r the
waypoi
n
t
appli
c
ation wi
ll provide be
n
e
fits for the d
e
velopme
n
t of AR.Dron
e
control.
2. Rese
arch
Metho
d
In the research, three schemes
of way
point navigati
on AR.Drone will be implemented
usin
g the fu
zzy logi
c
controller m
ade
by the La
bV
IEW
softwa
r
e.
Waypoi
nt nav
igation i
s
a fl
ying
comm
and
of
the AR.
D
ro
n
e
from
its i
n
itial po
sition
(
x,y
,
z
) to the des
i
red
pos
i
tion (
x
des
,y
des
,z
des
)
whi
c
h i
s
kno
w
n a
s
th
e wa
ypoint. For th
e flying man
o
euvring,
we
u
s
e th
ree
co
ntrol sig
nal
s, wh
ich
are
pitch,
roll
and ve
rtical
rate an
d a
r
e
the
results of t
h
ree
fuzzy lo
gic
co
ntrolle
rs. Th
e d
e
si
gn
of
fuzzy l
ogic control i
n
thi
s
study
con
s
i
ders the
follo
wing
point
s.
The field
u
s
e
d
is
4 met
r
e
s
in
length, 4 met
r
es in wi
dth a
nd 4 met
r
e
s
i
n
heig
h
t. Assuming the i
n
itial po
sition
whilst flying is i
n
the centre of the field, th
e rang
e of the fuzz
ificati
on input po
si
tion and refe
ren
c
e po
sitio
n
is
betwe
en -2 to
2 metre
s
. Th
e ran
ge of e
a
c
h o
u
tput is b
a
se
d on a
n
e
m
piri
cal meth
od to dete
r
mi
ne
the value ra
n
ge so that the
spee
d is not
too slo
w
or to
o fast. Singleton is cho
s
en
for its sp
eed i
n
cal
c
ulatin
g the defuzzification pro
c
e
s
s. The det
ail
s
of each co
ntroll
er are de
scrib
ed belo
w
:
a) T
o
rea
c
h
t
he
coo
r
din
a
te
of the
de
si
re
d
x-
po
s
i
tion
(
x
des
),
a pitch
control
l
oop
and t
w
o
fuzzy in
puts
are d
e
si
gne
d
,
which
are the de
sired
coordi
nate
s
of
x
and th
e
x-
positio
n from
the
NavData. Th
e ra
nge
of th
e fuzzificatio
n is -2
to
2
metre
s
, which is state
d
in
the 5
trian
g
ular
membe
r
ship f
unctio
n
s. M
e
anwhile, the f
u
zzy output i
s
the
pitch va
lue in the
ra
n
ge of -0.5 to
0.5
whi
c
h is
state
d
in 5 singl
etons. Fu
rthe
r deta
ils of this
desi
gn are sh
own in Fig
u
re
2.
Figure 2. The
fuzzy co
ntrol
of pitch
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Wa
ypoi
nt Na
vigatio
n of AR.Drone Q
u
a
d
roto
r Usi
ng
Fuzzy Logi
c Controlle
r (V
eroni
ca Ind
r
a
w
ati)
933
b) To rea
c
h t
he co
ordinat
e of the desi
r
ed
y-
po
s
i
tion (
y
des
), a roll
control loo
p
and two
fuzzy in
puts
are d
e
si
gne
d
,
which
are the de
sired
coordi
nate
s
of
y
and th
e
y-
positio
n from
the
NavData. Th
e ra
nge
of t
he fu
zzifi
cati
on i
s
-2 to
2 metres,
which
is
state
d
in 5
tria
ng
ular
membe
r
ship
function
s. Me
anwhile, the f
u
zzy output
i
s
the roll val
ue in the
ran
ge of -0.3 to
0.3
whi
c
h is
state
d
in 5 singl
etons. Th
e deta
ils of the desi
gn are
sho
w
n
in Figure 3.
Figure 3. The
fuzzy co
ntrol
of roll
c) To
rea
c
h t
he co
ordi
nate
of the desire
d
z-
positio
n (
z
des
), a vertical cont
rol rate
and two
fuzzy in
puts
are d
e
si
gne
d
,
which
are the de
sired
coordi
nate
s
of
z
and th
e
z-
positio
n from
the
NavData. Th
e ra
nge
of t
he fu
zzifi
cati
on i
s
-2 to
2 metres,
which
is
state
d
in 5
tria
ng
ular
membe
r
ship functio
n
s. Me
anwhile, the fuzzy output is
the vertical rate value in t
he ran
ge of -0.7
to 0.7 which is stated in 5
singl
eton
s. Furthe
r
detail
s
of this desi
g
n
are sh
own in
Figure 4.
Figure 4. The
fuzzy co
ntrol
of vertical rat
e
Usi
ng the three FL
Cs, the
waypoint i
s
obtai
ne
d with
three flight
scheme
s
, wh
ich a
r
e
rea
c
hin
g
the
waypoi
nt in t
h
ree
sequ
en
ce
s, re
aching
waypoi
nt in
two sequ
en
ces a
nd
rea
c
h
i
ng
waypoi
nt in one se
que
nce. Surely for ea
ch of th
e
s
e flight schem
es
the coo
r
din
a
tes of ea
ch F
L
C
is nee
ded a
n
d
explained
b
e
low.
2.1. Reach
Wa
y
point in
Three Sequ
e
n
ces
In this
schem
e, the AR.Drone
will rea
c
h the de
si
red
waypoi
nt co
ordin
a
te (
x
des
,y
des
,z
des
)
by reaching t
he de
sire
d
x
-positio
n (
x
des
) first, followe
d by the desired y-po
sition
(
y
des
), and la
stly
the de
sired
z
-position (
z
de
s
). The three FLCs (pit
ch, roll, ve
rti
c
al rate)
will
work together in
depe
nden
ce.
The co
ntroll
er in the flight schem
e wo
rks this way:
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 3, September 20
15 : 930 – 939
934
a)
The FL
C pit
c
h syste
m
ma
ke
s the A
R
.Drone
move to
wards the
x
des
coo
r
din
a
te, and
shut
s
down the FL
C roll
and F
L
C vertical rat
e
system.
Wh
en the AR.
D
rone rea
c
he
s
x
des
, the pitch
FLC system will
send
logi
c signal
s to activate the rol
l
FLC sy
stem.
b)
The a
c
tivation of this FL
C roll sy
stem
make
s the A
R
.Drone m
o
ve towa
rd
s the
y
des
position
and
stop
s th
e FLC pitch
and verti
c
al
rate
syste
m
. Whe
n
the A
R
.Drone
re
aches
y
des
, the
FLC roll syste
m
will sen
d
lo
gic si
gnal
s to activate the vertical
rate FL
C syste
m
.
c
)
The ac
tivation of this FLC
vertic
al
rate
sys
tem mak
e
s the AR
.
D
r
one move tow
a
rds
the
z
des
positio
n and
stop
s the FL
C pitch a
nd ro
ll system.
Whe
n
the AR.Drone
rea
c
hes
z
des
, the
FLC roll syst
em will se
nd
logic
signal
s
to acti
vate the FLC pitch system back a
nd rep
eats
the seq
uen
ce
above.
d)
This process
is d
one
be
ca
use
while
swi
t
chin
g
the
FL
C, a
ch
ang
e i
n
the
po
sition
may o
c
cur.
In ord
e
r th
at the AR.Dro
ne is always on
waypoi
nt (
x
des
,y
des
,z
des
), the FL
C mu
st be
con
d
u
c
ted u
s
ing the se
que
nce a
bove so
that it can hover.
The ord
e
rs o
f
flight and diagra
m
blocks of
the cont
rolled
system
to finish the fligh
t
scheme a
r
e
shown in Figu
re 5 and 6.
Figure 5. Sch
e
me of rea
c
h
waypoi
nt in three sequ
en
ce
s
Figure 6. Con
t
rol archite
c
tu
re of the sche
me of rea
c
h
waypoi
nt in three sequ
en
ce
s
2.2. Reach
Wa
y
point in
T
w
o S
e
que
n
ces
In this scheme, the AR.Drone will reach t
he desired coordinate
of the
waypoi
nt
(
x
des
,y
des
,z
des
) with flying towa
rd
s the (
x
des
,y
des
) coo
r
dinate first and then flying towa
rd
s the
desi
r
ed
z-
po
sition (
z
des
). Th
ree FL
C
s
(pit
ch, roll, v
e
rtic
al rate)
wor
k
t
h
is w
a
y
:
a)
The F
L
C pitch and roll
system
will be
turned
on at
the same tim
e
so that the AR.Drone
moves
toward the
x-
y
field
and to the (
x
des
,y
des
) coord
i
nate dire
ctly. After that, th
e two FL
C
system
s will send logi
c si
gn
als to acti
vate
the FLC verti
c
al rate
syste
m
.
b)
Once the FL
C vertical rat
e
system i
s
activated, the
AR.Drone
wi
ll move towa
rds the
z
des
coo
r
din
a
te an
d stop the FL
C pitch an
d roll system. After rea
c
hin
g
the
z
des
, the vertical rate
system
will send logi
cal
signal
s to activate the FLC p
i
tch and roll system ba
ck.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Wa
ypoi
nt Na
vigatio
n of AR.Drone Q
u
a
d
roto
r Usi
ng
Fuzzy Logi
c Controlle
r (V
eroni
ca Ind
r
a
w
ati)
935
c)
The po
sition
of the AR.Dro
ne on
waypoint (
x
des
,y
des
,z
des
) should always be maintained
usin
g the co
n
t
rol seq
uen
ce
above.
The ord
e
rs o
f
flight and diagra
m
blocks of
the cont
rolled
system
to finish the fligh
t
scheme a
r
e
shown in Figu
re 7 and 8.
Figure 7. Sch
e
me of rea
c
h
waypoi
nt in two sequ
en
ce
s
Figure 8. The
control a
r
chit
ecture of sch
e
me of rea
c
h
waypoi
nt in two sequ
en
ce
s
2.3. Reach
Wa
y
point in
One Seque
n
c
e
In this
schem
e, the AR.Drone
will rea
c
h the de
si
red
waypoi
nt co
ordin
a
te (
x
des
,y
des
,z
des
)
by flying dire
ctly towa
rd
s
those
coordi
nates. T
h
re
e
FLC
s
(pit
ch,
roll, v
e
rti
c
al
rate
) will
w
o
rk
simultan
eou
sl
y and each wi
ll be respon
si
ble for its po
sition.
The ord
e
rs o
f
flight and diagra
m
blocks of
the cont
rolled
system
to finish the fligh
t
scheme a
r
e
shown in Figu
re 9 and 10.
Figure 9. Sch
e
me of rea
c
h
waypoi
nt in one se
que
nce
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 3, September 20
15 : 930 – 939
936
Figure 10. Th
e control arch
itecture of
sc
heme of re
ach waypoi
nt in one sequ
en
ce
To impl
ement
the flight
sch
e
mes ab
ove,
several
subVI
and
fro
n
t pa
nels in th
e L
a
bVIEW
softwa
r
e
are
made. Seve
ral main
su
bVI, such
a
s
th
e su
bVI used
for flying an
d su
bVI used
to
read th
e Na
vData, modif
y
the subVI i
n
the
AR.Drone L
abVIEW tool
kit whi
c
h
wa
s mad
e
by
Micha
e
l Mog
enson [1, 14] for AR.Dro
n
e
1.0 so t
hat it could be u
s
ed for the AR.Dron
e
2.0. The
inputs of thi
s
AR.Drone
sy
stem
are
the
pitch val
ue,
roll, yaw
rate
and ve
rtical
rate who
s
e va
lues
are i
n
the
ran
ge of
-1 to
1.
Mean
while, t
he vari
able
s
t
hat could
be t
a
ke
n fro
m
th
e AR.Dro
ne
a
r
e
actual
pitch v
a
lue, fo
rwa
r
d
spe
ed, a
c
tual
roll,
sid
e
ward spee
d, a
c
tu
al yaw rate, yaw, ve
rtical
rate
and altitud
e
. To obtai
n th
e po
sition
s o
f
x
an
d
y
, th
e su
bVI po
sition e
s
timatio
n
is m
ade. T
h
e
inputs from
the bl
ock
po
si
tion e
s
timatio
n
are the
forward
spe
ed
(
v
x
),
side
wa
rd
sp
eed
(
v
y
) a
nd
time stamp
(
t
). The e
quati
on of this
est
i
mation of
x
and
y
po
sitio
n
is
stated a
s
in Equ
a
tion
(1)
and (2
) bel
ow:
x
n
=x
n-1
+v
xn
(t
n
-t
n-1
)
(
1
)
y
n
=y
n-1
+v
yn
(t
n
-t
n-1
)
(
2
)
Whe
r
ea
s
n
is the present sampl
e
data
and
z
po
sitio
n
is the dire
ct resu
lt of the ultrasoni
c se
nso
r
onbo
ard.
The
s
e
equ
ation
s
result in
the
subVI po
sition
estimation.
T
he F
L
C blo
ck is re
alised i
n
to
the subVI Fu
zzifi
cation, su
bVI Infe
rence
,
and subVI Defuzzificatio
n.
3. Results a
nd Analy
s
is
The al
go
rithm
of the
FL
C i
s
implem
ented
in
the
AR.
D
rone,
whi
c
h i
s
flown
auton
o
m
ously
in a clo
s
ed
sp
ace u
s
in
g La
bVIEW, Figure
11. The pro
c
ed
ure
s
for te
sting are:
a)
Thro
ugh th
e
front pa
nel
software, th
e
AR.Dr
one i
s
flown in
hove
r
mod
e
1 m
e
tre from th
e
grou
nd. That
point is called
the initial pos
ition with the coo
r
din
a
te value (0,0,1
).
b)
Next, the de
si
red
wayp
oint
coo
r
din
a
te i
s
inse
rted th
rou
gh the front p
anel. By switching off the
hoverin
g mo
d
e
, the AR.
D
rone
will fly a
u
tonomo
u
sly
with the
mad
e
FL
C
contro
l towa
rd
s the
waypoi
nt spot
.
c)
While
flying f
r
om th
e initi
a
l po
sition to
the
waypoin
t
coo
r
din
a
te, the a
c
tual
x-
po
sition,
y-
positio
n and
z-
p
o
sitio
n
values a
r
e recorded.
d)
After rea
c
hin
g
the wayp
oi
nt coo
r
din
a
te
,
indicate
d with hover mo
de, the AR.Drone
will b
e
lande
d towa
rds the g
r
oun
d
station.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Wa
ypoi
nt Na
vigatio
n of AR.Drone Q
u
a
d
roto
r Usi
ng
Fuzzy Logi
c Controlle
r (V
eroni
ca Ind
r
a
w
ati)
937
Figure 11. Th
e front panel
and blo
c
k dia
g
ram of wayp
oint navigatio
n
The re
sult o
f
the FLC algorithm for
the
flight scheme “rea
ch
waypoint in
three
seq
uen
ce
s” i
s
sh
own in Figure 1
2
.
Figure 12. Th
e experim
ent results of sch
e
me re
ach waypoint in three se
que
nce
s
Figure 12
sh
ows the resul
t
s of
the expe
riment d
one t
h
ree tim
e
s, from the initial
positio
n
towards th
e
waypoi
nt. Ge
nerally, the
AR.Drone
ca
n do
control
comm
and
s
made fo
r it to fly
towards
the
x-
pos
i
tion first, followed by the
y-
po
sition, and to the heig
h
t of the desired
z-
positio
n.
It can be seen t
hat whe
n
the
x-
p
o
sitio
n
is rea
c
he
d an
d it is moving towa
rd
s the
y-
positio
n, there is a
shift of the
x-
po
si
tion away from the setp
oint. Exactly the sam
e
thing
happ
en
s whe
n
the
y-
p
o
siti
on is re
ache
d and the drone is movin
g
towards th
e
z-
po
s
i
tion
. T
h
is
happ
en
s be
cause of the
switchi
ng e
n
a
b
le p
r
o
c
e
s
s
a
nd di
sabl
eme
n
t of the th
re
e FL
Cs that
are
being
u
s
ed. The pro
b
lem
can
al
so occur be
cau
s
e
the valu
es of
the x an
d y
positio
ns are
the
result of the
estimated
out
put of the
blo
c
k po
siti
on
calcul
ation, no
t the se
nsor readin
g
s dire
ctly.
Ho
wever, ge
nerally ea
ch
positio
n can
be rea
c
h
ed a
t
around 4
seco
nd
s while
the waypoin
t
is
rea
c
he
d at around 1
5
se
co
nds.
The next testing is for the
flight schem
e
“Re
a
ch Wa
ypoint in Two Sequen
ce
s”, which
wa
s also don
e three time
s. The re
sult is
sho
w
n in Fig
u
re 13.
-0
.
5
0
0.
5
1
1.
5
-0
.
5
0
0.
5
1
1.
5
0
0.
5
1
1.
5
2
2.
5
x
-
po
si
t
i
on[
m
]
Wayp
o
i
n
t
Navig
a
t
i
on
with
3
se
qu
en
ce
s: [
x
]-->
[y]-->
[z]
y-
pos
i
t
i
on[
m
]
z-po
si
ti
on[m
]
0
5
10
15
20
0
0.
5
1
1.
5
x-po
si
ti
on[m
]
Res
p
ons
e
o
f
E
a
c
h
P
o
s
i
t
i
on
Co
ord
i
na
te
0
5
10
15
20
-0
.
5
0
0.
5
1
1.
5
y
-
pos
i
t
i
o
n[m
]
0
5
10
15
20
0.
5
1
1.
5
2
z-po
si
ti
on[m
]
T
i
m
e
[s
e
c
o
n
d
s
]
I
n
it
i
a
l P
o
s
i
t
i
o
n
Wa
y
p
o
i
n
t
1
s
t
E
x
pe
ri
m
e
nt
2nd
E
x
per
i
m
ent
3r
d E
x
p
e
r
i
m
e
n
t
I
n
it
ia
l P
o
i
n
t
(0
,0
,
1
)
W
ayp
oi
n
t
(
1
,1
,2
)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 3, September 20
15 : 930 – 939
938
Figure 13. Th
e experim
ent results of sch
e
me re
ach waypoint in two
sequ
en
ce
s
The re
sult sh
ows that the positio
ns (
x
de
s
,y
des
), can b
e
rea
c
he
d well simultan
e
ously, but
whe
n
it
mov
e
s t
o
w
a
rd
s t
h
e
z
des
, the shift of the
x
des
an
d
y
des
from th
e setp
oint ca
n be seen. T
h
is
also ha
ppe
ns beca
u
se the control pit
c
h
and roll switche
s
off whe
n
the vertical
rate cont
rol i
s
wor
k
in
g.
Agai
n, the effects of the
estim
a
ted positions
x and y
are still visi
ble. T
he
waypoi
nt
can
be rea
c
h
ed at
aroun
d 9 se
cond
s.
The la
st testi
ng is d
one fo
r the flight schem
e “Re
a
ch
Waypoint in
One Seq
uen
ce”,
whe
r
e
the AR.Drone
flies towards
the waypoi
nt (
x
des
,y
des
,z
des
) directly. The
result is sh
own in Figure 14
.
Figure 14. Th
e experim
ent results of sch
e
me re
ach waypoint in on
e seq
uen
ce
The results
of the experi
m
ents
(don
e
3 ti
mes) sh
ow
that
the AR.Drone ca
n
rea
c
h
waypoi
nt (
x
de
s
, y
des
,z
des
)
with the settling time less than 4 secon
d
s. The resp
onse wh
en it is
steady
sho
w
s a rel
a
tively better re
sult th
an the tw
o p
r
evious flig
ht scheme
s
. Thi
s
is
ca
used b
y
the 3 FLCs worki
ng si
mult
aneo
usly.
4. Conclusio
n
Gene
rally, the three flight
scheme
s
can
be im
pleme
n
ted usi
ng th
ree FL
Cs (FL
C
pitch,
FLC roll,
FL
C vertical
rate) for th
e
wayp
oint navig
atio
n. The
results of
the
s
e te
sts
sho
w
th
at the
-0.
5
0
0.
5
1
1.
5
-0.
5
0
0.
5
1
1.
5
0
0.
5
1
1.
5
2
2.
5
x
-
p
o
s
it
io
n
[
m
]
Waypoint navigation with
2
sequ
ences:
(x,y)-->
(z)
y-
p
o
si
t
i
on [
m
]
z
-
p
o
s
i
ti
o
n
[m
]
0
5
10
15
0
0.
5
1
1.
5
x
-
posi
t
i
on[
m
]
Response
of ea
ch
coor
dinate
0
5
10
15
-1
0
1
2
y
-
pos
i
t
i
on[
m
]
0
5
10
15
0.
5
1
1.
5
2
z-
pos
i
t
i
on[
m
]
T
i
m
e
[
s
econd
s
]
I
n
it
i
a
l P
o
s
i
t
i
o
n
Way
P
oi
n
t
1s
t
E
x
pe
ri
m
e
nt
2n
d E
x
p
e
ri
m
ent
3rd
E
x
pe
ri
m
e
nt
In
it
ia
l P
o
s
i
t
i
o
n
(
0
,0
,1
)
W
a
y
P
oi
nt
(1
,
1
,
2
)
-0
.
5
0
0.
5
1
1.
5
-1
0
1
2
0
0.
5
1
1.
5
2
2.
5
x-
p
o
si
t
i
o
n
[
m
]
Wa
ypoint Navig
a
ti
on:only in 1 seque
nce
y-
posi
t
i
on[
m
]
z-po
si
ti
on
[
m
]
0
2
4
6
8
10
0
0.
5
1
1.
5
x-p
o
si
ti
o
n
[m
]
Respo
n
se o
f
each
po
s
i
ti
on
coordin
a
te
0
2
4
6
8
10
-1
0
1
2
y-po
si
ti
o
n
[
m
]
0
2
4
6
8
10
1
1.
5
2
2.
5
z-po
si
ti
on
[
m
]
T
i
m
e
[s
e
c
o
n
d
s
]
I
n
it
ia
l P
o
s
i
t
i
o
n
W
a
y
poi
nt
1
s
t
E
x
per
i
m
ent
2
nd E
x
p
e
r
i
m
ent
3
nd E
x
p
e
r
i
m
ent
I
n
itia
l P
o
s
i
tio
n
(0
,0
,1
)
W
a
y
p
oi
nt
(1
,
1
,
2
)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Wa
ypoi
nt Na
vigatio
n of AR.Drone Q
u
a
d
roto
r Usi
ng
Fuzzy Logi
c Controlle
r (V
eroni
ca Ind
r
a
w
ati)
939
flight sche
m
e
straight toward
s the way
point with
the FLC worki
ng simulta
n
e
ously is the
most
satisfying
co
mpared to
th
e othe
r t
w
o fli
ght sch
e
me
s.
Cal
c
ul
ation
o
f
the po
sition
s
(
x
an
d
y
) i
s
st
i
ll
su
sceptible t
o
noise. Use can b
e
made
of a com
pen
sator o
n
the side of the pit
c
h an
d roll to
get
better re
sult
s.
Referen
ces
[1]
Micha
e
l M. T
he AR.Dr
one
L
abVIEW
T
oolki
t: A Soft
w
a
re
F
r
ame
w
ork for
the
Contro
l
o
f
Lo
w
C
o
st
Quadrotor Aer
i
al Ro
bots. Master of Sc
ience
T
hesis.
T
U
FT
S Universit
y
; 20
12.
[2]
Stepha
ne P, Ni
colas B. AR.Dr
one D
e
vel
o
p
e
r Guide.
Parrot. SDK 1.6
. 201
1.
[3]
http://ardrone2.parrot.com acc
e
ssed on 11 A
ugust 2014.
[4]
Pierre-Je
an B,
F
r
ancois C, David V, Nico
l
as P.
T
he Navigati
on an
d Co
ntrol T
e
chn
o
lo
gy Inside th
e
AR.Drone Micro UAV
. 18th IF
AC W
o
rld Co
n
g
ress. Mila
no, Ital
y
. 2
011.
[5]
Nick Dijks
h
o
o
r
n
, Arnou
d Viss
er. Integratin
g
S
ensor
and M
o
tion Mo
de
ls to Loc
aliz
e an
Autonom
ous
AR.Drone.
Internatio
nal Jo
urn
a
l of Micro Air Vehic
l
e.
20
11; 3(4): 183-
20
0.
[6]
Krajnik T
,
Vonasek V, F
i
ser D, F
a
igl J.
AR-Dron
e
as a Pl
a
tform for Rob
o
t
i
c Rese
arch a
n
d
Educati
o
n
.
Rese
arch an
d
Eductatio
n
in R
obotics: EURO
BOT
.
Heidelb
e
r
g. 2011.
[7]
Agun
g Pra
y
itn
o
, Veron
i
ca Ind
r
a
w
ati, Gabri
e
l
Ut
omo. T
r
ajector
y
T
r
acking
of AR.Drone Qu
adrotor Us
in
g
F
u
zz
y
Lo
gic C
ontrol
l
er.
T
E
LK
OMNIKA T
e
lec
o
mmunic
a
tio
n
Co
mp
uting
El
e
c
tronics
and
C
ontrol.
20
14;
12(4): 81
9-8
2
8
.
[8]
Sarah Y
i
fang
T
ang. Vision-B
a
sed C
ontro
l for
Auton
o
mou
s
Quadrotor. F
i
nal
Re
port: Under
grad
uat
e
d
Seni
or T
hesis.
Dep
a
rtment of Mecha
n
ica
l
an
d Ae
rosp
ace E
ngi
neer
in
g, Princeton U
n
iv
ersit
y
; 20
13.
[9]
Rab
ah A
b
b
a
s, Qingh
e W
u
.
Improved
Le
a
der F
o
l
l
o
w
e
r
F
o
rmation
Co
ntrol for M
u
lti
p
le Qu
adr
otor
s
Based AFSA.
T
E
LKOMNIKA
T
e
leco
mmunic
a
tion C
o
mputi
n
g Electron
ics a
nd Co
ntrol
. 20
15; 13(1): 85-
92.
[10]
Emad A
b
b
a
si
Seid
aba
d, S
a
e
ed V
a
n
daki,
A
li V
ahi
din
Kam
y
a
d
. D
e
sig
n
i
n
g
F
u
zz
y PID
C
ontrol
l
er fo
r
Quadrotor.
Int
e
rnati
ona
l Jou
r
nal of Adv
a
nced R
e
se
ar
c
h
in Co
mput
er Scienc
e &
T
e
chnol
ogy
(IJARCST)
. 2014; 2(4): 22
1-2
27.
[11]
E Abbasi, MJ Mahj
oob. Co
ntro
lli
ng of Quad
rotor UAV Usin
g a F
u
zz
y
S
y
st
em for
T
uning the PID Gain
s
in Hov
e
rin
g
Mo
de. 201
5.
[12]
Matild
e Sant
os
, Victoria L
o
p
e
z
, F
r
anciso M
o
rata.
Intell
ig
e
n
t F
u
zz
y
Cont
roller
of a Qu
adrotor
. IEEE
Intelli
gent S
y
stems and Kn
o
w
led
ge En
gin
eer
ing C
onf (ISKE). 2010.
[13]
K Senth
i
l
Kum
a
r, Moh
a
mma
d R
a
she
ed,
R
Muthu
Mad
h
a
v
a Kum
a
r.
De
sign
an
d I
m
p
l
ementati
o
n
of
F
u
zz
y
Lo
gic
C
ontrol
l
er for
Quad
Rotor
UA
V
. 2nd
Intern
ation
a
l
Confer
en
ce o
n
R
e
searc
h
in
Sci
ence
,
Engi
neer
in
g an
d T
e
chnolo
g
y
(
I
CRSET
’2014).
Duba
i. 201
4.
[14]
https://ardronelabv
ie
w
t
oolkit.
w
ordpress.com
Evaluation Warning : The document was created with Spire.PDF for Python.