TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol. 15, No. 2, August 201
5, pp. 270 ~
276
DOI: 10.115
9
1
/telkomni
ka.
v
15i2.810
1
270
Re
cei
v
ed Ma
y 14, 201
5; Revi
sed
Jul
y
1, 2015; Accept
ed Jul
y
16, 2
015
Flight-Path Tracking Control of an Aircraft Using
Backstepping Controller
Laban
e
Chrif*
1
, Zemalache Meguen
n
i Kada
2
, Tah
a
r
Mohamed
2
1
Departem
ent of Electtotechn
i
c, Universit
y
of
Dr Moula
y
T
ahar, Said
a, Alg
é
rie
2
LDEE Lab
orat
or
y
UST
O, MB
Oran, Algeri
a
*Corres
p
so
ndi
ng auth
o
r, e-mail: c_l
aba
ne
@
hotmai
l
.fr
A
b
st
r
a
ct
For transportat
i
on
aircraft, the primary cont
rol ob
jectiv
e for an autopilo
t system
engaged
durin
g
appr
oach
an
d l
and
ing
is rel
a
ti
ve to t
he flig
ht path tracki
ng o
n
the bas
is of
hig
h
ly si
mp
lifie
d lin
ear
mo
del
s of
flight dy
na
mics
. T
he dyn
a
m
ic
s gover
ni
ng th
e flig
ht p
a
th
of
an
aircraft ar
e
in
gen
era
l
h
i
g
h
ly n
o
n
lin
ear
a
n
d
involv
e co
mp
le
x physics for w
h
ich no acc
u
rate mode
ls ar
e avai
lab
l
e. In
this pap
er a non
lin
ear
mo
d
e
l
descri
b
in
g the
lon
g
itud
in
al
eq
uatio
ns of
moti
on i
n
st
rick fe
e
dback f
o
rm is
deriv
ed. Backs
teppi
ng
is uti
l
i
z
ed
for the constru
c
tion of a gl
oba
lly stabi
li
z
i
n
g
c
ontrol
l
er w
i
th a nu
mb
er of
free para
m
eters. It i
s
imp
l
e
m
e
n
ted
a
control
l
er w
i
th
an
intern
al
lo
o
p
co
ntrols
inv
o
lving
the
p
i
tch
rate of th
e
airc
raft and
a
n
ext
e
rna
l
l
o
o
p
w
h
i
c
h
inclu
des
a
ngl
e
of attack, p
a
th
ang
le
an
d
pitc
h a
ngl
e. F
i
n
a
ll
y, non
lin
ear
si
mu
lati
on r
e
sult
s for a
l
ong
itud
in
a
l
mo
de
l of a tran
sportatio
n
aircr
a
ft are displ
a
ye
d and d
i
scuss
e
d
.
Ke
y
w
ords
:
ba
cksteppi
ng, air
c
raft control, nonli
n
e
a
r contro
l, Lyapu
nov sta
b
ility
Copy
right
©
2015 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
The de
sig
n
o
f
flight control
system
s is
a
typical
nonli
near co
ntrol probl
em,
due
dire
ctly
to the
ch
ang
es i
n
ai
rcraft
dynami
c
s
wi
th flight
cond
itions
and
ai
rcraft
co
nfiguration. Fo
r thi
s
rea
s
on, a dy
namic m
odel
that is stable and
ade
qu
ately dampe
d in one flight condition
may
become unst
able
or at least in
adequately damped i
n
another one.
In transportation ai
rcraft, a
lightly damp
ed o
scill
atory
mode
may
cau
s
e
disco
m
fort to pa
ssen
gers
or
make
difficult
the
control of the aircraft for the pilot. For a
combat
aircraft, this conditi
on may lead
to more
criti
c
al
situation once the
aircraft is
i
nherently unstabl
e due the
maneuverability requirem
ents and
capability of attack.
In this pa
pe
r
we i
n
tro
d
u
c
e
an
altern
at
ive for
co
ntrol
se
paration; a
ba
ckstepping
controlle
r i
s
u
s
ed
to a
c
hi
eve glo
bal
stabi
lity with an
int
e
rnal
loop
controls involving the
pitch rat
e
of the ai
rcraft
and
extern
al
loop
which i
n
clu
d
e
s
a
ngl
e of atta
ck,
p
a
th an
gle
an
d pitch a
ngle
.
A
backsteppi
ng
controll
er is
prop
osed to solve t
he pat
h angle sta
b
il
ization p
r
obl
e
m
of the aircraft
transpo
rtation
.
Firstly, the full nonlin
ear l
ongit
udi
nal d
y
namics are descri
bed u
n
der
con
s
ide
r
a
t
ion
of external di
sturb
a
n
c
e
s
a
c
ting on p
a
th
angle. The
n
, block ba
ckste
pping
controll
er.
Backstep
ping
is a recursiv
e pro
c
ed
ure that
interla
c
e
s
the choi
ce of
a Lyapunov functio
n
with the de
si
gn of the fee
dba
ck
cont
rol
.
The advant
age of this te
chni
que i
s
that it can from
the
stabili
zing
no
nlinea
r te
rms rath
er than
eliminating
th
em. Backste
pping
ha
s
be
en a
pplie
d to
a
numbe
r of different de
sig
n
taske
s
[7].
The goal of this wo
rk is to design a
con
t
rol law able t
o
deal with th
e aircraft long
itudinal
dynamics
,
for all the normal opera
ting regimes
of the airc
raft, wit
h
minimal inf
o
rmation of the
aero
d
ynami
c
model. T
he
controlle
r m
u
st
be
able
to
m
a
ke
the
syste
m
se
ek the
re
feren
c
e
s
in
th
e
aero
d
ynami
c
velocity and
flight path
an
gle, usi
ng as actuato
r
s
the
elevator defl
e
ction
s
a
nd
t
he
thrust level.
The main ai
m of this pap
er is to asse
ss the respe
c
tive perfo
rm
ances of the
nonlin
ear
backsteppi
ng
techniq
ue ap
plied to the f
light-path a
ngl
e tracking
co
ntrol problem.
The p
ape
r i
s
stru
ctured a
s
follows: first, i
n
se
ction
2 i
s
pre
s
e
n
ted th
e longitu
dinal
aircraft
model eq
uati
on. This ai
rcraft repre
s
e
n
ts a
tran
sp
ort
a
tion aircraft like A320/A3
XX and Boeing
737-200/3
00
this n
online
a
r co
ntrol th
eory will control
the ai
rcraft ri
gi
d flight dyna
mics to a
c
hi
e
v
e
global stability.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Flight-Path T
r
acking
Control of an Aircraft Us
ing Ba
ckste
ppin
g
Co
ntrolle
r (La
b
a
ne Ch
rif)
271
In the sectio
n 3, backste
pping
control
desig
n pro
c
edure is sho
w
n tra
c
king f
our
state
variable
s
. Fin
a
lly, in the section 4 a num
erical sim
u
lati
on is do
ne to demon
strate
2. Nonlinear
Aircraft
Model
The lon
g
itudi
nal motion
of an aircraft is well de
scri
b
ed by the foll
owin
g stan
da
rd set of
dif
f
erential eq
uation
s
,
the state
vari
able
s
,
(
a
irs
p
ee
d)
,
(p
ath an
gl
e),
(angl
e of
attack),
(pitch a
ngle
)
and
(pitch rat
e
), are d
epi
cted in Figu
re 1
.
(1)
(2)
(3)
(4)
(5)
is the
flight
path angle, to which we shall return lat
e
r
.
is the en
gine thru
st
force a
nd
the elevator an
gle. Finally
,
,
et
are the aerodynamics
forces
lift, drag and
pitchin
g
mom
ent and
is the inertial mom
ent about
axis in body axe
s
.
Figure 1. Defi
nition of force
s
, moment
s a
nd angl
es
As usu
a
l in a
e
rodyn
a
mi
c forces a
nd mo
ments a
r
e co
mputed throu
gh their no
n-
dimens
ional coeffic
i
ent, as
follows
:
,
,
̅
(6)
Whe
r
e
is the air density
,
is the referen
c
e wing surfa
c
e,
̅
is the mean cho
r
d and
,
and
are the lift, drag and pitching mom
ent coef
fici
ents. Moreover
,
we consider the following
model
s for th
e drag a
nd m
o
ment co
ef
ficients [1, 2].
(7)
(8)
Whe
r
e
,
,
,
,
,
an
d
are ai
rcraft
aerodynam
i
c
coef
fi
cients,
and
is
the elevator angle. in this work
,
,
,
,
and
are con
s
idere
d
to be
unkno
wn
para
m
eters,
while is
k
n
own.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 15, No. 2, August 2015 : 270 –
276
272
3. Bac
ksteppin
g
Con
s
id
er a d
y
namic sy
ste
m
:
(9)
Whe
r
e
∈
,
∈
are st
ate variable
s
and
∈
is the co
ntrol input.
The control o
b
jective [x] to this dynami
c
syst
em i
s
to
desi
gn a
cont
rol la
w such that the
state
can b
e
stabilized
at 0, a supp
ose
d
global
asymptoticall
y
stable equi
librium with
null
input
0
0
, furthermore it can o
b
se
rved that
can b
e
re
ga
rde
r
as
a virtual (o
r medi
a
t
e)
control input
for the dyna
mics
,
the dynamics of
is controlled by
the real co
n
t
rol input
.
This i
s
an i
m
porta
nt feature to ma
ke
use of
for t
he followi
ng
control law
synthesi
s
. No
w,
sup
p
o
s
e that there exist
s
a
control la
w [5].
(10)
(11)
Whe
r
e
is positive definite functio
n
. Then
t
he whole dy
namics can b
e
expre
s
sed
as:
(12)
(13)
Whe
r
e,
(14)
(15)
With,
(16)
Then, the Lya
punov fun
c
tio
n
con
d
idate o
f
the full system is given by [9]:
,
(17)
The time deri
v
ative of
,
is given by:
,
(18)
Substituting the ineq
uality of Equation (9) and
Eq
uati
on (10
)
into Equation (14
)
yields:
,
(19)
By an edequ
ate choi
ce of
,
such a
s
:
(20)
Whe
r
e is p
o
sitive consta
nt, the full system is glob
all
y
asymptoticallt stable si
n
c
e it satisfie
s the
following condition:
,
(21)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Flight-Path T
r
acking
Control of an Aircraft Us
ing Ba
ckste
ppin
g
Co
ntrolle
r (La
b
a
ne Ch
rif)
273
Finally, sibstit
u
ting Equatio
n (7) a
nd Equ
a
tion (1
1
)
into
Equation (1
6
)
yields the ef
fective
control input:
(21)
4. Con
t
roller
Design
The main obj
ective of the
controlle
r is to track slo
w
states
,
,
and
.
But the problem
is divide
d in t
w
o p
a
rt
s. The
first on
e is th
e co
ntrol
of
,
a
nd
, this pro
b
l
e
m can b
e
viewe
d
like
a two-scaleti
ne app
roa
c
h
bec
au
se the fast states
is use
d
as co
ntrol input as is propo
se
d in
Lee et al. (2001) [8]. the seco
nd part is the control of
, acco
rdingl
y to
achieve this obje
c
tive is
prop
osed a first orde
r ba
cksteppi
ng u
s
in
g the throttle setting inp
u
ts [3, 4].
4.1.
Control
of the Flight Path Angle
The
ba
ckstep
ping
pro
c
e
d
u
r
e to
be
ap
pl
ied
can
be
vi
ewe
d
a
s
t
w
o
-
timescal
e a
p
p
roa
c
h
becau
se the fast state
s
are
used a
s
cont
rol input
s for
the slo
w
stat
es
,
and
intermediately.
Ho
wever, thi
s
method
olo
g
y consi
d
e
r
s the transient
resp
on
se
s of the fast
states and, therefor
doe
s not re
quire the ti
mescale
sep
a
ration a
s
su
mption. First, it is nece
s
sary repla
c
e
the
aero
d
ynami
c
force
s
and
m
o
ments into
the
state
e
qua
tion. The
dyn
a
mics
of fligh
t
-path
angle
are
written as
[11, 12]:
(22)
The first step
of a backste
pping a
p
p
r
oa
ch
con
s
ist
s
i
n
difining the
output outp
u
t error
whi
c
h is give
n here by:
(23)
Whe
r
e
is the desire
d
flight-path angl
e. Then the
error dynami
cs
of flight-path angle is give
n
by:
(24)
A Lyapunov functio
n
for
is given by:
(25)
Its time-de
r
ivative is then given by:
(26)
With:
0
(27)
(28)
Since:
(29)
At the seco
nd
step the ne
w erro
r given b
y
:
(30)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 15, No. 2, August 2015 : 270 –
276
274
Its time-de
r
ivative is then given by:
(31)
With:
(32)
Such that:
(33)
1
(34)
Afin d’élimine
r
cette e
r
reur, la fon
c
ti
on de
Lya
pun
ov
augm
entée
d’un autre
te
rme qui
contie
nt une
nouvelle e
rre
ur
(35)
Its time-de
r
ivative is then given by:
(36)
(37)
With:
0
(38)
(39)
Finally, the effective cont
ro
l
of flight path angle given
by:
,
,
,
,
,
(40)
It appears th
at the backsteppin
g
app
ro
ach i
s
far fro
m
being
strai
gh forward a
nd that
su
ch an inttricate co
nst
r
u
c
tion sh
ould
be tested th
rou
gh fully to provide confiden
ce in
its
perfo
rman
ce
s.
5. Simulation
Results
I
n
t
h
is s
e
ct
i
on,
simul
a
t
i
o
n
re
sult
s
of
the co
ntroll
ers
develo
p
ed are sho
w
n a
nd
demon
strate the perfo
rma
n
c
e of this con
t
rol law.
Figure 2. Flight path angle
Figure 3. Angle of attack
0
1
2
3
4
5
6
7
8
9
10
-0
.
2
0
0.
2
0.
4
0.
6
0.
8
1
1.
2
1.
4
1.
6
ti
m
e
(
s
)
g
a
m
m
a
(
deg)
F
l
i
g
ht
pat
h
angl
e
0
1
2
3
4
5
6
7
8
9
10
-0
.
4
-0
.
3
-0
.
2
-0
.
1
0
0.
1
0.
2
0.
3
0.
4
Ti
m
e
(
s
)
al
pha
deg
)
A
n
gl
e of
A
t
t
a
c
k
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Flight-Path T
r
acking
Control of an Aircraft Us
ing Ba
ckste
ppin
g
Co
ntrolle
r (La
b
a
ne Ch
rif)
275
Figure 4. The
pitch angl
e
Figure 5. Pitch angle rate
Figure 6. Con
t
rol inputs:
el
evator defle
ction
Figure 7.Co
ntrol input
s: ele
v
ator throttle
s
e
tting
Figure 8. The
lift
force L
Figure 9. The
drag force
D
6. Conclu
sion
In this pa
per
nonlin
ear
ba
ckste
ppin
g
techniqu
e wa
s u
s
ed to
de
sign
law to be
ap
plied in
a nonli
nea
r a
i
rcraft mod
e
l. A cont
rolle
r
wa
s p
r
op
ose
d
to tra
c
k
,
and
u
s
ing ang
ular rate a
s
intermediate,
thus, it i
s
possible
control a
sl
o
w
dy
namics
usi
n
g
the fast
dyn
a
mics. With
this
0
1
2
3
4
5
6
7
8
9
10
-0
.
2
0
0.
2
0.
4
0.
6
0.
8
1
1.
2
1.
4
1.
6
1.
8
Ti
m
e
(
s
)
t
e
t
ha (
deg)
P
i
t
c
h angl
e
0
1
2
3
4
5
6
7
8
9
10
-1
-0
.
8
-0
.
6
-0
.
4
-0
.
2
0
0.
2
0.
4
0.
6
0.
8
1
Ti
m
e
(
s
)
q (
r
a
d/
s
)
Pi
tch
r
a
t
e
0
1
2
3
4
5
6
7
8
9
10
-7
-6
-5
-4
-3
-2
-1
0
1
Ti
m
e
(
s
)
de
l
t
e
(
d
eg
)
E
l
ev
a
t
or
de
f
l
ec
t
i
o
n
0
1
2
3
4
5
6
7
8
9
10
0
1
2
3
4
5
6
x 1
0
5
Ti
m
e
(
s
)
Th
r
u
s
t
(
N
)
Th
r
u
s
t
f
o
r
c
e
0
1
2
3
4
5
6
7
8
9
10
-5
-4
-3
-2
-1
0
1
x 1
0
4
Ti
m
e
(
s
)
L
L
i
ft
c
o
e
f
fi
c
i
e
n
t
0
1
2
3
4
5
6
7
8
9
10
0
2
4
6
8
10
12
x 1
0
4
Ti
m
e
(
s
)
Dr
a
g
D
r
ag f
o
r
c
e
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 15, No. 2, August 2015 : 270 –
276
276
controlle
r ap
plied the
erro
r expon
entiall
y conve
r
ge
s, then sy
stem
can
be gl
ob
aly stabili
zed
or
reach a new
equilibrium st
ate.
We have
addressed the
path
angl
e control probl
em for a transportation ai
rcraft. A
nonlin
ear m
o
del de
scribi
n
g
the longitu
dinal mo
tio
n
in stri
ck fe
edba
ck form
at wa
s de
rived,
carefully sele
cting the sta
t
es inclu
d
e
s
and maki
ng
prope
r app
roximations.
A backstep
p
i
n
g
control la
w was
de
sign
ed
recursively in
thre
e
step
s,
resulting i
n
a
nonli
nea
r
co
ntrolle
r
with fi
ve
free pa
ram
e
ters. T
he
cl
ose
d
loop
stability of
the error
stat
es a
nd the
para
m
eters of
backsteppi
ng
techniq
ue a
r
e examine
d
by the Lyapu
nov theory, a
nd it is
shown that the error
expontially co
nverge to a compa
c
t set whose si
ze
is a
j
ustabl
e by the desi
gn pa
ra
meters. Finall
y
,
a nonline
a
r
si
mulation of a
n
aircraft ma
neuver i
s
pe
rformed to de
monst
r
ate th
e perfo
rman
ce o
f
the prop
osed
control laws.
Referen
ces
[1]
V
an Orrt Eduard
Ric
hard.
Ad
a
p
tive Bac
kstep
pin
g
C
ontr
o
l
a
nd
Safety Analysis
for Modern
Air
c
r
a
ft.
PhD
thesis.
T
e
chnische Uni
v
ersiteit Delft.
201
1.
[2]
AL
da Silv
a,
P
P
agli
o
n
e
, Y
o
n
e
yam
a
T
.
Control
l
ab
ility an
alysis
and
stab
ili
z
a
ti
on
of
fle
x
i
b
le
aircrafts
.
XV
II
I
C
o
n
g
r
e
ss
o Br
asil
ei
r
o
d
e
A
u
tomática.
20
10.
[3]
C Laba
ne, K Z
e
malac
he Megu
en
ni.
Aircraft Control System Usin
g L
Q
R and LQG
control
l
er w
i
th
Optim
a
l Estimation-Kalm
an Filter Design
. Procedi
a Engi
ne
er
ing. 20
14; 80:
245-
257.
[
4
]
A
B
G
u
i
m
a
r
a
e
s
N
e
t
o
.
D
i
n
â
m
i
c
a
e
c
o
n
t
r
o
l
e
d
e
a
e
r
o
n
a
v
e
s
f
l
e
x
í
v
e
i
s
c
o
m
mod
e
lagem
a
e
r
o
d
i
n
â
m
i
c
a
p
e
l
o
método d
o
u
b
le
t
lattice. Maste
r
’
s
thesis.
T
e
chnol
ogic
a
l Instit
ute
of
Aeron
aut
ics
(IT
A
).
2008.
[5]
L
Son
nev
eldt,
Q Ch
u,
J
Muld
er
.
Nonlinear
flig
h
t
c
o
n
t
r
o
l
des
ign
u
s
i
n
g
c
o
n
s
t
r
a
i
n
e
d
a
dap
t
i
v
e
bac
k
step
pi
ng
.
J.
Guid.
Contr
.
D
y
nam
. 20
07; 3
0
(2):
3
2
2
-
336.
[6]
A
Morales, P
P
aglione, F
T
r
i
v
e
n
o
.
No
nli
n
ear
flig
h
t
con
t
r
ol
u
s
i
n
g
b
a
ck
steppi
ng
t
ec
hn
iq
u
e. 2
0
1
1:
555-
558.
[7] L
Sonn
ev
e
ld
t
.
A
d
aptive
Bac
k
s
te
p
pin
g
F
lig
ht Contr
o
l
for Mo
dern
F
i
g
h
te
r
A
ir
c
r
af
t.
P
h
D
.
Th
e
s
i
s
.
Delft
T
e
c
hnolog
yc
Un
i
v
e
rs
i
t
y
. 2
010.
[8]
T
Lee,
Y
Ki
m.
Nonli
n
e
a
r a
d
a
p
t
i
v
e
fli
gh
t
co
n
t
ro
l
us
ing
bac
kstep
ping
an
d
ne
ura
l
ne
t
w
or
ks
con
t
ro
l
l
er
.
20
0
1
;
24:
675-
68
2
.
[9]
ER
V
Oor
t
.
A
d
aptive
Bac
k
s
te
p
p
i
ng C
ontr
o
l
and
Safet
y
A
nal
ysis
f
o
r
Mo
de
r
n
F
i
ghter
A
i
rc
ra
f
t
.
PhD.
T
hesis. Delft
T
ec
h
no
lo
gy
c
Un
i
v
e
rs
i
t
y
.
20
1
1
.
[10]
AL
da
Silv
a, P
.
P
a
gli
o
n
e
,
Y
o
n
e
yam
a
T
.
Conceptua
l
fl
e
x
ible
aircr
a
ft
mod
e
l
for
mo
de
lin
g, ana
lysis an
d
control stud
ies
.
Submitted
for pr
esentation
in
AIAA.
2010.
[11]
Boua
di H, Mor
a
-Cami
no F
.
Ai
rcraft trajectory tracking
by no
nlin
ear s
patia
l
inversi
o
n
. AIA
A
Guidanc
e
,
Navig
a
tio
n
an
d
Control C
onfer
ence Mi
nne
ap
olis, Min
nesota
,
USA. 2012: 1
-
17.
[12]
Boua
di H, Mor
a
-Cami
no F
.
Space-
base
d
n
o
n
lin
ear
dyna
mi
c inversi
on c
o
ntrol for aircr
a
ft continu
o
u
s
desce
nt ap
pro
a
ch
. IEEE Evol
ving
and A
d
a
p
tive Intel
lig
ent
S
y
stems C
onfe
r
ence
.
Ma
drid,
Spai
n. 20
12
:
164-
169.
Evaluation Warning : The document was created with Spire.PDF for Python.