ARTICLE
International Journal of Advanced Robotic Systems
Information Fusion-Based Optimal
Attitude Control for an Alterable Thrust
Direction Unmanned Aerial Vehicle
Regular Paper
Ziyang Zhen1,*, Ju Jiang1, Xinhua Wang1 and Daobo Wang1
1 College of Automation Engineering of Nanjing University of Aeronautics and Astronautics, China
* Corresponding author E-mail: [email protected]
Received 13 Jun 2012; Accepted 6 Nov 2012
DOI: 10.5772/54886
© 2013 Zhen et al.; licensee InTech. This is an open access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,
distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract Attitude control is the inner‐loop and the 1. Introduction
most important part of the automatic flight control
system of an unmanned aerial vehicle (UAV). The In recent years, unmanned aerial vehicles (UAVs) have
information fusion‐based optimal control method is been widely used for military applications and also for
applied in a UAV flight control system in this work. civil use [1]. An advanced flight control strategy is
Firstly, a nonlinear model of alterable thrust direction indispensable for the autonomous flight of UAVs.
UAV (ATD‐UAV) is established and linearized for Reference [2] applies the LQG method that requires rate
controller design. The longitudinal controller and feedback and a H∞ controller only by properly choosing
lateral controller are respectively designed based on the weighting functions, which meet the performance
information fusion‐based optimal control, and then requirements for a vertical short take‐off and landing
the information fusion flight control system is built aircraft. A method of optimal flight control for a rigid
up. Finally, the simulation of a nonlinear model body model has been proposed in [3]. All parameters of
described as ATD‐UAV is carried out, the results of the point mass model need to be determined by
which show the superiority of the information fusion‐ comparison with the flight simulations. In the attitude
based control strategy when compared to the single‐ control loop, four control signals should be properly
loop design method. We also show that the ATD produced. In the case of the moderate and the short‐time
technique improves the anti‐disturbance capacity of flight manoeuvres, both flight trajectories also respond
the UAV. approximately. Neural networks can be used in flight
control systems to improve the adaptive performance, by
Keywords Unmanned Aerial Vehicle, Information Fusion, compensating the system uncertainties and thus adapting
Attitude Control, Automatic Flight Control, Optimal to the variations of flight conditions, while
Control accommodating the control system failures [4]. A neural
network‐based direct‐adaptive controller is proposed for
www.intechopen.com Ziyang Zhen, Ju Jiang, Xinhua Wang and DaobInot WJ Aadnvg:RIonbfortmic aStyio, n20F1u3si,oVno-Bl.a1s0ed, 4O3p:2ti0m1a3l 1
Attitude Control for an Alterable Thrust Direction Unmanned Aerial Vehicle
an unstable UAV. The control law is developed to track information fusion‐based optimal controllers are
the pitch rate command. A neural network with linear designed in section 4 and the simulations are carried out
filters and a back propagation learning algorithm is used in section 5. Finally, conclusion is drawn in section 6.
to approximate the control variable. The bounded signal
requirement is circumvented in the design of the neural 2. Nonlinear Model of the ATD‐UAV
controller, providing stability and tracking performances.
An on‐line learning scheme is adopted to compensate for In order to express the nonlinear model of the ATD‐UAV,
uncertainties due to variations in aerodynamic
coefficients, control surface failures and centre of gravity the input‐output relationship and the inner constitute
position. The performance of the above control method is
validated in [5] under different flight conditions. should be established. A nonlinear model with total
Furthermore, an intelligent control strategy based on a variable for an ATD‐UAV consists of dynamic equations
brain emotional learning (BEL) algorithm which is based
on the emotional learning process in the amygdale‐ and kinematic equations. The former includes force
orbitofrontal system of the mammalian brain is
investigated in [6] for application in UAV attitude flight equations and moment equations, while the latter
control. The UAV is in flat flight with wind disturbance,
and a BEL‐based intelligent controller (BELBIC) is includes navigation equations and motion equations. The
applied to improve the attitude stability control
performance of the UAV. In [7], fuzzy logic modules are variables’ definitions are given in Table 1.
developed in the autonomous controller for altitude
control, speed control and heading control, through
which the global positions of the aircraft are well
controlled. In [8], a genetic algorithm (GA) is used to Variables Notations
meet the requirements of the frequency domain handling
qualities for the longitudinal plane, in which the e , a , r , Elevator, aileron, rudder and throttle
parameters are implemented as fitness functions related T opening angles
to the desired magnitude of bandwidth and time‐delay.
T , T Thrust deflection angles
The thrust vectoring (TV) technique has been proved to V Airspeed
be able to improve the mobility and agility of a manned
jet‐propelled aircraft [9‐12]. The application of TV , Angles of attack and sideslip
technology in UAVs is being increasingly studied, and Bank and azimuth angles of flight path
has been successfully demonstrated on several aircraft to ,
provide tactical manoeuvring advantages at slow speed,
high attack angle, and prevent the aircraft from loss of x,y,z General axial, lateral and normal
control due to aerodynamic surface saturation. However, positions
TV technology may have a potentially significant pay‐off , ,
in some critical areas, such as vehicle complexity, p,q,r Pitch angle, roll angle and yaw angle
maintenance, and total cost of ownership.
Roll rate, pitch rate and yaw rate
With reference to the TV technique of jet‐propelled
aircraft, an alterable thrust direction (ATD) technique is L,M,N Roll, pitch and yaw moments
presented for screw propeller aircraft, which essentially
vectorizes the thrust of the screw propeller engine. The LT , MT , NT Roll moment, pitch moment and yaw
objective of this work is to employ a new modern optimal moment due to thrust
control named information fusion control for an ATD‐
UAV. This information fusion control is an optimal Table 1. Variables definition
control based on information fusion estimation.
Compared to traditional UAVs, the mathematical model
The remainder of this paper is organized as follows. of this ATD‐UAV has several unique features. There are
Section 2 derives the nonlinear dynamic model of the thrust components on the yb and zb axes of the body axis
ATD‐UAV. In section 3, linear models of the ATD‐UAV system, while there are none of these components for
are given, in which the thrust deflection angles are traditional UAVs. Thus, the total force and moment of the
treated as the additional control variables. The ATD‐UAV are different to a traditional UAV. Therefore,
the dynamics of the ATD‐UAV is different to that of a
traditional UAV. The corresponding kinematics
properties are also different.
For this ATD‐UAV, the three components of thrust on the
body axis system are given by
T
Tx 1 tan2 T tan2 T
Ty Tx tan T
T tan T (1)
tan2 T
1 tan2 T
Tz Tx tanT 1 T tan2 T tanT
tan2 T
For a traditional UAV, T T 0 , Tx T , Ty Tz 0 .
The three components of thrust on the body axis system
of the ATD‐UAV are given by
2 Int J Adv Robotic Sy, 2013, Vol. 10, 43:2013 www.intechopen.com
LT Tylz Tzly Tylz kinematic equations, when the UAV is in level and no‐
MT Txlz Tzlx sideslip flight, the kinematic equations can be divided
Txlz Tzlx (2) into longitudinal and lateral kinematic equations which
are decoupled. The linearization of the nonlinear
NT Txly Tylx Tylx equation is always conducted in an equilibrium state. For
the ATD‐UAV, after the linearization process, we get the
linear longitudinal motion equations
Thus, the force equation, the moment equation, the
navigation equations and motion equations of the ATD‐ X lon AlonXlon BlonUlon (3)
UAV can be easily derived according to the dynamic
model of the traditional UAV.
3. Linear Models of the ATD‐UAV Xlon [V , , , q]T
Ulon [T , e , T ]T
In order to use the modern control theory, the nonlinear
model is usually linearized by using the small
perturbation method. According to the nonlinear
TV cos * DV T* sin * D mg cos * g cos *
m
m 0
T cos * L mg sin *
TV sin * LV L mV * mg sin * Lq mV *
L mV 0 L mV * L mV *
* 0 1
0
Alon MV M Mq
Iy Iy Iy
M M M
M mg sin *
Iy(L mV * )
TV sin * LV T* cos * L mg sin * Lq mV *
Iy (L mV * ) Iy(L mV * ) Iy (L mV * )
TT cos * D e T* sin *
m
m m
L e
TT sin * L mV * T* cos *
L mV * 0 L mV *
0 0
Blon
M T T sin * M e M L e MT MT* cos *
I y ( L mV * ) Iy Iy(L mV * ) Iy
I y ( L mV * )
The lateral motion equation is given by
X lat AlatXlat BlatUlat (4)
Y a Y r T*
Xlat [ , , p, r]T mV * mV * mV *
0
0 0
IzL r IzxN r
Ulat [a , r , T ]T Blat Iz L a IzxN a IxIz Izx2 IzLT IzxNT
IxIz I 2 IzxL r IxN r IzIx Izx2
zx IxIz Izx2
Y g cos * sin * Yp I zx L a IxN a I zx L T IxNT
mV * V* mV *
0 Yr cos * IxIz I 2 IzIx Izx2
1 mV * zx
0
Iz Lp IzxNp
0 0 IxIz Izx2 tan * Given the above linear models of the ATD‐UAV, the
natural properties can be analysed. Furthermore, the
Alat I z L I zx N IzxLp IxNp Iz Lr IzxNr information fusion‐based attitudinal controllers can also
IxIz Izx2 be designed, as shown in the following section.
2 IxIz Izx2
IxIz I zx
IzxL IxN IzxLr IxNr
IxIz I 2 IxIz Izx2
zx
www.intechopen.com Ziyang Zhen, Ju Jiang, Xinhua Wang and Daobo Wang: Information Fusion-Based Optimal 3
Attitude Control for an Alterable Thrust Direction Unmanned Aerial Vehicle
4. Information Fusion Estimation‐Based Optimal Control Hence, a linear discrete time system of longitudinal
For the ATD‐UAV channel of ATD‐UAV can be derived from (3) ‐ shown
as
4.1 Information Fusion Estimation Theory
Xlon(k 1) AlonXlon(k) Blon,1Ulon,1(k) Blon,2Ulon,2 (9)
Li Xiao Rong puts forward the optimal fusion rules based
on the linear estimation of minimum variance [13]. From where Ulon,2 is the throttle opening, Ulon,2 T ,
the perspective of information fusion, Zhi Sheng Wang Ulon,1 [e , T ]T .
and Ziyang Zhen put forward information fusion control
methods, which take the desired tracking information,
system dynamic information and ideal control strategy The design of the longitudinal controller is such that we
information as measure information of the control
variables [14‐16]. obtain the optimal control sequence to make the UAV’s
Assume that a measurement zi mi is a linear function longitudinal states keep the equilibrium states and make
of the state x n , expressed by
the following LQ performance index function minimum.
J Xlon* (N) Xlon ( N ) 2
Qlon
(10)
N 1 2 2
zi = Hix + vi , i 1 ~ n (5) Xlon* (k) Xlon ( k ) Qlon U lon ,1( k ) Rlon ]
[
where Hi mi n is the mapping matrix, vi mi is the
measurement error which is a random vector with zero k0
mean value and co‐variance of Ri . The optimal
estimation problem of the state x can be described by where Xlon*(k) 0 is the desired state vector, Qlon and
Rlon are positive definite matrices, and N is the terminal
n 2 time.
Ri ‐1
argmin
x i=1 The solving steps based on the sequential information
xˆ = zi ‐ Hix . (6)
fusion method are as follows.
The information weight of information on itself and its
Step 1: From performance index function, we get an
co‐variance are generally reciprocals of each other. Thus
information expression as
we get the following definitions.
Ulon,1* (k) Ulon,1(k) n(k) (11)
I[zi zi ] Ri1 : information weight of zi about itself; I[Ulon,1* (k)] Rlon (12)
I[zi x] HiT Ri1Hi : information weight of zi about x. where Ulon,1* (k) 0 , E[n(k)] 0 , var[n(k)] Rlon1 .
Step 2: Assume the co‐state optimal estimation Xˆ lon( j 1)
Then the information weight of xˆ about x is and its information weight Plon1( j 1) are obtained,
j k 1, k 2, , N , then we get an information
nn expression as
I[xˆ x] = I[zi x] = HiT Ri‐1Hi (7) Xˆ lon( j 1) Xlon( j 1) w( j 1) (13)
i=1 i=1 where E[w( j 1)] 0, var[w( j 1)] Plon( j 1) . Submitting
the system dynamic equation (9) and the control sequence
which shows that the information weight of xˆ about x is soft constraint information expression (11) into (13), we get
equal to the sum of the information weight of zi about x.
If I[xˆ x] is non‐singular, then the optimal estimate of the Xˆ lon( j 1) AlonXlon( j) Blon,1Ulon,1( j) Blon,2Ulon,2 w( j 1) (14)
state x is expressed by [14]
xˆ I[xˆ x] 1 n HiT Ri1zi (8) I[Xˆ lon ( j 1) Xlon ( j)] AlonT [Plon ( j 1) Blon,1Rlon1Blon,1T ]1 Alon (15)
i1 Step 3: Information expression can be derived from the
performance index function, described as
Equation (5~8) is a uniform linear model of information
fusion estimation. Equation (5) is also called an Xlon*( j) Xlon( j) m( j) (16)
information expression in information fusion fields. I[Xlon*( j)] Qlon (17)
4.2 Longitudinal Information Fusion Controller
For the studied ATD‐UAV, the throttle opening angle
is always set to be constant in cruise flight to save fuel.
4 Int J Adv Robotic Sy, 2013, Vol. 10, 43:2013 www.intechopen.com
where Xlon* ( j) 0 , E[m( j)] 0, var[m( j)] Qlon1 . it has
Step 4: By fusing the above information, we get the
lim Xˆ lon( j) Xˆ lon (28)
optimal fusion filtering of the co‐state sequence and
N
information weights as
Xˆ lon( j) Plon( j){AlonT[Plon( j 1) Blon,1Rlon1Blon,1T ]1 (18) Therefore, the approximate optimal estimation of co‐state
[Xˆ lon( j 1) Blon,2Ulon,2 ]} and optimal control sequence are respectively rewritten
as
Plon1( j) Qlon AlonT[Plon( j 1) BlonRlon1BlonT ]1 Alon (19) Xˆ lon [I PlonAlonT (Plon Blon,1Rlon1Blon,1T )1]1 (29)
PlonAlonT (Plon Blon,1Rlon1Blon,1T )1 Blon,2Ulon,2
where Xˆ lon(N 1) 0 , Plon1(N 1) 0 . By iterative
calculation of (18~19), Xˆ lon(k 1) and Plon1(k 1) can be Uˆ lon,1(k) (Rlon Blon,1T Plon1Blon,1)1 BlTon,1Plon1
obtained. Then we get the following information [[I PlonAlonT (Plon Blon,1Rlon1Blon,1T )1]1 (30)
expressions. PlonAlonT (Plon Blon,1Rlon1Blon,1T )1 Blon,2Ulon,2
Xˆ lon ( k 1) AlonXlon ( k) Blon,1Ulon,1(k) Blon,2Ulon,2 w( k 1) (20) AlonXlon(k) Blon,2Ulon,2 ]
I[Xˆ lon(k 1) Ulon,1(k)] Blon,1T Plon1(k 1)Blon,1 (21) 4.3 Lateral Information Fusion Controller
Step 5: By fusing the information of Xˆ lon(k 1) and The linear discrete time system of lateral channel of ATD‐
Ulon,1* (k) on control sequence, we get the optimal fusion UAV can be derived from (4), given by
estimation of control sequence as
Xlat (k 1) AlatXlat (k) BlatUlat (k) (31)
Uˆ lon,1(k) [Rlon Blon T Plon1( k 1)Blon,1 ]1{BlTon,1Plon1(k 1) By solving the discrete state regulator problem that
,1 making the following linear quadratic performance
index minimum, the optimal control sequence.can be
[xˆlon(k 1) AlonXlon(k) Blon,2Ulon,2 ]} (22) obtained.
where Plon(k) is a symmetrical nonnegative definite J XlatT (N)QlatXlat(N)
matrix, satisfying the following Riccati difference
equation N1 (32)
Plon1( AlonT [Plon(k ,1Rlon1Blon T ]1 [XlatT (k)QlatXlat(k) UlatT (k)RlatUlat(k)]
,1 k0
k) Qlon 1) Blon Alon
(23)
Plon1(N) Qlon here Qlat(k) QlatT (k) 0 , Rlat(k) RlatT (k) 0 .
For above Riccati difference equation, given any semi‐ The solving steps of optimal control sequence based on
positive matrix P01 , when N , based on optimal the sequential information fusion method are as
follows.
control theory, we get
Step 1: An information expression can be obtained from
lim Plon1( j) Plon1 (24) the performance index function
N
Here j [k 1, km] and km is big enough. Hence, let Ulat*(k) Ulat(k) n(k) (33)
Plon1 replace Plon1( j) , then co‐state filter (18~19)
becomes I[Ulat*(k)] Rlat (34)
Xˆ lon( j) Plon{AlonT [Plon Blon,1Rlon1Blon,1T ]1 (25) where Ulat* (k) 0 , E[n(k)] 0 , var[n(k)] Rlat1 .
[Xˆ lon( j 1) Blon,2Ulon,2 ]} Step 2: Assume the co‐states optimal estimations
{ Xˆ lat( j 1) , Plat1( j 1) , j k 1, k 2, , N } are
Plon1 Qlon AlonT (Plon Blon,1Rlon1Blon,1T )1 Alon (26) obtained, then we get an information expression as
According to the convergence theorem of iteration
method in numerical analysis theory, when Xˆ lat( j 1) Xˆ lat( j 1) w( j 1) (35)
{Plon{AlonT[Plon Blon,1Rlon1Blon,1T ]1} 1 (27) where E[w( j 1)] 0, var[w( j 1)] Plat( j 1) . Submitting
(31) and (33) into (35), we get
www.intechopen.com Ziyang Zhen, Ju Jiang, Xinhua Wang and Daobo Wang: Information Fusion-Based Optimal 5
Attitude Control for an Alterable Thrust Direction Unmanned Aerial Vehicle
Xˆ lat ( j 1) AlatXlat ( j) Blatn( j) w( j 1) (36) Based on the above fusion control laws of the attitudes of
I[Xˆ lat ( j 1) Xlat ( j)] AlatT [Plat ( j 1) BlatRlat1BlatT ]1 Alat (37) the ATD‐UAV, a flight control system scheme is designed
as shown in Figure 1. Variables with “*” denote the flight
states under equilibrium conditions.
Step 3: The information expression can be derived from
the performance index function, described as
Xlat*( j) Xlat( j) m( j) (38) T V
I[Xlat*( j)] Qlat (39)
where Xlat*( j) 0 , E[m( j)] 0, var[m( j)] Qlat1 . V* e
*
Step 4: By fusing the above information, we get the * T q
optimal fusion filter of the co‐state sequence as
q*
* a
*
p*
Xˆ lat ( j) Plat ( j)AlatT [Plat ( j 1) BlatRlat1BTlat ]1 Xˆ lat ( j 1) (40) r* r
T p
Plat1( j) Qlat AlatT[Plat( j 1) BlatRlat1BlatT ]1 Alat (41)
r
where Xˆ lat (N 1) 0 , Plat1(N 1) 0 . Thus we get
Xˆ lat (N) 0 , Plat1(N) Qlat , and Xˆ lat (k 1) Xˆ lat (k 2)
Xˆ lat (N) 0 .
Figure 1. Attitude stable fusion control system
Step 5: After { Xˆ lat (k 1) , Plat1(k 1) } are obtained, an
information expression can be obtained as 5. Simulation Study
Xˆ lat(k 1) Xlat (k 1) w(k 1) (42) 5.1 Traditional Control Responses
where E[w(k 1)] 0 , var[w(k 1)] Plat(k 1) . This section provides the simulation results conducted
with the ATD‐UAV. Suppose the ATD‐UAV is in the
Submitting (31) into (42), the soft constraint information straight‐and‐level flight with a flight speed of Vd =50m/s ,
of Xˆ lat (k 1) on Ulat (k) can be expressed as anaedro t=h4a5t it. isT shued dcoennltyro dlliesdtu robbejde cbt y ias wrienpdr:e sVeanetreod= 20bmy /sa,
nonlinear model, in which the aerodynamic parameters
Xˆ lat (k 1) AlatXlat (k) BlatUlat (k) w(k 1) (43) are obtained by wind tunnel tests. The controller is
designed by the traditional single‐loop method.
I[Xˆ lat (k 1) Ulat(k)] BlatT Plat1(k 1)Blat (44)
Figure 2 shows the responses of the pitch angle, pitch
Step 5: By fusing the information of Xˆ lat (k 1) and angle rate, roll angle, roll rate, yaw angle, yaw rate,
Ulat*(k) , we get the optimal fusion estimation of the elevator output, aileron output, rudder output and thrust
control sequence as angle. As can be seen from Figure 2, we get that ATD
control can decrease the magnitude of change caused by
Uˆ lat (k) [Rlat BlatT Plat1(k 1)Blat ]1 the wind disturbance. Turbulence magnitudes on the
attitudes of the hybrid control are smaller than those of
BlatT Plat1(k 1)AlatXlat(k) (45) the pneumatic rudder control, which means that the ATD
technique is useful in enhancing the flight attitude
where Plat(k) is symmetrical nonnegative definite matrix, maintaining performance in wind turbulence. Moreover,
satisfying the following Riccati difference equation with the pneumatic rudder control, the rudders tend to
saturate under strong wind disturbance. However, the
Plat 1(k) Qlat AlatT[Plat (k 1) Blat Rlat 1BlatT ]1 Alat (46) ATD technique overcomes this problem, and
Plat 1(N) Qlat compensates the effectiveness of the pneumatic rudder
control.
6 Int J Adv Robotic Sy, 2013, Vol. 10, 43:2013 www.intechopen.com
Pitch angle (rad) 00 Roll angle(rad)
-0.1 -0.05
Rudders control Rudders control
Rudders and ATD control
-0.2 Rudders and ATD control
0 2 4 6 8 10 -0.10 2 4 6 8 10
Time (s) Time (s)
0.2 0.4
Pitch rate (rad/s) 0 Roll rate (rad/s) -0.2
-0.2 Rudders control Rudders control
Rudders and ATD control Rudders and ATD control
-0.80 2 4 6 8 10
0 2 4 6 8 10 Time (s)
Time (s)
(a) Pitch angle and pitch rate responses (b) Roll angle and roll rate responses
0.1 0.02
Rudders control
Yaw angle (rad) 0.01
0 Rudders and ATD control
0
-0.1
Elevator angle (rad) -0.01
-0.2
-0.02
0 2 4 6 8 10
Yaw rate (rad/s) Time (s) -0.03
0.5 -0.04
Rudders control
Rudders and ATD control -0.05 Rudders control
0 Rudders and ATD control
-0.50 2 4 6 8 10 -0.060 2 4 6 8 10
Time (s) Time (s)
(c) Yaw angle and yaw rate responses (d) Elevator responses
0.1 Thrust longitudinal angle (rad) 0.3
0.2
Aileron angle (rad) 0
0.1
-0.1 Rudders control Thrust lateral angle (rad) 0 10
Rudders and ATD control
-0.10 2 4 6 8 10
-0.20 2 4 6 8 10 Time (s)
TIme (s)
0.4
0.01 0.2
Rudder angle (rad) 0 0
-0.2
-0.01 Rudders control -0.40 2 4 6 8
Rudders and ATD control Time (s)
-0.020 2 4 6 8 10
TIme (s)
(e) Aileron and rudder responses (f) Thrust angle responses
Figure 2. Traditional flight control responses
5.2 Information Fusion Control Responses investigated. Beforehand, the weight matrices of the
information fusion control are determined. Simulation
Under the same simulation conditions using the results are shown in Figure 3.
traditional control, the information fusion control is
www.intechopen.com Ziyang Zhen, Ju Jiang, Xinhua Wang and Daobo Wang: Information Fusion-Based Optimal 7
Attitude Control for an Alterable Thrust Direction Unmanned Aerial Vehicle
0 0.02
Pitch angle (rad) -0.05 Roll angle (rad) 0
-0.1
-0.02 Rudders control
-0.15 Rudders control Rudders and ATD control
Rudders and ATD control -0.040 2 4 6 8 10
-0.20 2 4 6 8 10 Time (s)
Time (s)
0.2 0.1
Pitch rate (rad/s) 0 Roll rate (rad/s) 0
-0.2 Rudders control -0.1 Rudders control
Rudders and ATD control Rudders and ATD control
-0.20 2 4 6 8 10
-0.40 2 4 6 8 10
Time (s) Time (s)
(a) Pitch angle and pitch rate responses (b) Roll angle and roll rate responses
Yaw angle (rad) 0.1 Elevator angle (rad) 0.02
Rudders control Rudders control
Yaw rate (rad/s)
0 Rudders and ATD control 0.01 Rudders and ATD control
-0.1
0
-0.2
-0.01
0 2 4 6 8 10
Time (s) -0.02
0.5 -0.03
Rudders control
Rudders and ATD control -0.04
0 -0.05
-0.50 2 4 6 8 10 -0.060 2 4 6 8 10
Time (s) Time (s)
(c) Yaw angle and yaw rate responses (d) Elevator responses
0.1 Thrust longitudinal angle (rad) 0.4
0.2
Aileron angle (rad) 0
0
-0.1 Rudders control Thrust lateral angle (rad) 10
Rudders and ATD control -0.2
10
-0.20 2 4 6 8 10 -0.40 2 4 6 8
Time (s) Time (s)
0.01 0.4
Rudder angle (rad) 0 0.2
-0.01 Rudders control 0
Rudders and ATD control -0.2
-0.020 2 4 6 8 10 -0.40 2 4 6 8
Time (s) Time (s)
(e) Aileron and rudder responses (f) Thrust angle responses
Figure 3. Information fusion flight control responses
As can be seen from the above results, the ATD technique traditional control, the ATD technique in information
is able to improve the manipulation effectiveness of the fusion control also plays an important role in
aircraft and decrease the rudder angles, which can avoid compensating the rudders.
the adverse case of rudder saturation. Similar with the
8 Int J Adv Robotic Sy, 2013, Vol. 10, 43:2013 www.intechopen.com
Furthermore, compared with traditional control, [5] S. Suresh, N. Kannan, Direct adaptive neural flight
information fusion control has better responses in flight control system for an unstable unmanned aircraft [J].
attitude stability, which shows that as a multivariable Applied Soft Computing. 2008, 8: 937–948.
control method, it is better than the single‐loop design
method for aircraft flight control systems. [6] Guoyong Huang, Ziyang Zhen, Daobo Wang, Brain
emotional learning based intelligent controller for
6. Conclusion nonlinear system [C]. Second International
Symposium on Intelligent Information Technology
As is well‐known, the efficiency of the rudders is greatly Application, Shanghai, China, December 20‐22, 2008,
reduced when the UAV is in low speed flight. In this 2: 660‐663.
work, for the ATD‐UAV, the ATD technique is applied
supplementary to the rudders in controlling the UAV. A [7] Sefer Kurnaz, Omer Cetin, Okyay Kaynak, Fuzzy
novel optimal control method based on information logic based approach to design of flight control and
fusion is used to design the longitudinal controller and navigation tasks for autonomous unmanned aerial
the lateral controller. Simulation results verify that the vehicles [J]. Journal of Intelligent and Robotic
ATD technique compensates the effectiveness of the Systems, 2009, 54(1‐3): 229‐244.
rudders to improve the manipulation performance, also
showing that the proposed modern control method [8] Roberto Fantinutto, Giorgio Guglieri, Fulvia B.
improves the anti‐wind disturbance performance, which Quagliotti, Flight control system design and
is stronger than the traditional single‐loop method. optimization with a genetic algorithm [ J]. Aerospace
Science and Technology, 2005, 9: 73–80
However, in order to extend the thrust nozzle life, the
ATD technique is activated only when the conventional [9] H. Beh, R. van den Bunt and B. Fischer, High angle of
aerodynamic control surfaces are insufficient, i.e., they attack approach and landing control law design for
are unable to generate the necessary forces and moments the X‐31A [C]. AIAA Aerospace Sciences Meeting
required for the required manoeuvres. Therefore, the and Exhibit, 40th, Reno, NV, Jan. 14‐17, 2002.
ATD is actually a discontinuous control variable and
normally inactive for the UAV with a low attack angle, [10] Reinhold Steinhauser, Gertjan Looye and Oliver
high speed and low altitude. Brieger, Design and evaluation of control laws for the
X‐31A with reduced vertical tail [C]. AIAA Guidance,
7. Acknowledgements Navigation, and Control Conference and Exhibit,
Providence, Rhode Island, Aug. 16‐19, 2004.
This study was supported by the NUAA Fundamental
Research Funds (no. NS2013029, NP2011012, NN2012101, [11] Li Fu, Meixiang Yu, Xinhe Xu, The strategy studying
NP2011049), Aeronautical Science Foundation of China of air combat about the unmanned combat air vehicles
(no.2010ZA52002), and Specialized Research Fund for the [C]. 20th Chinese Control and Decision Conference,
Doctoral Program of Higher Education (no. Yantai, Chinese, 2‐4 July, 2008, pp. 3775–3778.
20123218120015).
[12] Takano Hiroyuki, Wada Hidesuke, Yamasaki
8. References Takeshi, Baba Yoriaki, Optimal vertical maneuvers of
the aircraft with thrust vectoring in the rigid body
[1] John A. N. Farnsworth, John C. Vaccaro, Michael model [C]. International Conference on Control,
Amitay, Active flow control at low angles of attack: Automation and Systems, Seoul, South Korea, 2007,
stingray unmanned aerial vehicle [J]. AIAA Journal, pp. 2100‐2104.
2008, 46(10): 2530–2544.
[13] Xiaorong Li, Yunming Zhu, Jie Wang, Chongzhao
[2] Jafar Zarei, Allahyar Montazeri, Mohmmad Reza Han, Optimal linear estimation fusion – part I:
Jahed Motlagh, Javad Poshtan. Design and unified fusion rules [J].IEEE Transactions on
comparison of LQG/LTR and H∞ controllers for a Information Theory, 2003, 49(91): 2192‐2208.
VSTOL flight control system [J]. Journal of the
Franklin Institute, 2007, 344: 577‐594. [14] Zhisheng Wang, Daobo Wang, Ziyang Zhen, Primary
exploration of nonlinear information fusion control
[3] Takuya Kinoshita, Fumiaki Imado, A study on the theory [J]. Science in China Series F: Information
optimal flight control for an autonomous UAV [C]. Sciences, 2007, 50(5): 686‐696.
Proceedings of the 2006 IEEE International
Conference on Mechatronics and Automation. June [15] Zhisheng Wang, Weijie Liu, Ziyang Zhen, Design of
25 ‐ 28, 2006, Luoyang, China, pp. 996‐1001. optimal tracking controller for nonlinear discrete system
with input delay using information fusion estimation
[4] Aydogan Savran, Ramazan Tasaltin, Yasar Becerikli, method [C]. 2007 IEEE International Conference on
Intelligent adaptive nonlinear flight control for a Control and Automation, Guangzhou, Guangdong,
high performance aircraft with neural networks [J]. China, May 30‐June 1, 2007, pp. 2535‐2538
ISA Transactions. 2006, 45(2): 225‐247.
[16] Ziyang Zhen, Zhisheng Wang, Zhou Hu, Information
fusion based solving method for linear quadratic
optimal control problem [C]. 7th IEEE International
Conference on Control & Automation, Christchurch,
New Zealand, 9‐11 December, 2009: 727‐731.
www.intechopen.com Ziyang Zhen, Ju Jiang, Xinhua Wang and Daobo Wang: Information Fusion-Based Optimal 9
Attitude Control for an Alterable Thrust Direction Unmanned Aerial Vehicle