Beruflich Dokumente
Kultur Dokumente
JID: MECH
Mechatronics
journal homepage: www.elsevier.com/locate/mechatronics
Technical note
a r t i c l e
i n f o
Article history:
Received 30 April 2016
Revised 10 September 2016
Accepted 18 October 2016
Available online xxx
Keywords:
Airborne seeker
Cascade loop control
Formation ight
Hardware in the loop
Leader-follower
Tracking loop
a b s t r a c t
This paper proposes a hardware in the loop simulation testbed by using airborne seekers (an onboard
visual tracking system) as a relative measurement sensor in the leader-follower formation ight. To the
best of the authors knowledge, this is the rst study on experimental applications of airborne seekers
in unmanned aircraft formation ight. The proposed structure gives an effective method to study the effects of uncertainties such as the camera process delay and the seeker measurement noises on formation
keeping. Compensation of errors in visual measurements is considered as well.
1. Introduction
A signicant amount of research efforts have been focused on
the formation control of unmanned vehicles. Excellent surveys of
formation control are found in [1,2] and [3] which review the existing results and the related control approaches. Among various
architectures for formation control, the leader-follower structure is
more popular in which the leader moves along a predened trajectory while another vehicles (followers) keep desired relative distances and orientations from the leader. According to sensing capabilities, formation control can be categorized in three classes:
communication-based, vision-aided and vision-based control [4]. In
the vision-based formation strategy, the follower only is equipped
with a relative measurement sensor and does not know its global
position and orientation.
Hardware-in-the-loop (HIL) simulation is an effective technique
that is used for the development and testing of control systems
while some of the control loop components are simulated in a
proper environment and the other components are real hardware.
Today, HIL techniques are frequently used to reduce time and cost
of development and prototyping of engineering systems. Advan-
R
This paper was recommended for publication by Associate Editor Radhakant
Padhi.
Corresponding author.
E-mail addresses: m.a.dehghani@aut.ac.ir (M.A. Dehghani), menhaj@aut.ac.ir
(M.B. Menhaj), hadighadery@yahoo.com (H. Ghaderi).
http://dx.doi.org/10.1016/j.mechatronics.2016.10.015
0957-4158/ 2016 Elsevier Ltd. All rights reserved.
Please cite this article as: M.A. Dehghani et al., A hardware in the loop simulation testbed for vision-based leader-follower formation
ight, Mechatronics (2016), http://dx.doi.org/10.1016/j.mechatronics.2016.10.015
ARTICLE IN PRESS
JID: MECH
2
the leader-follower formation ight of unmanned aircraft is proposed in [23] and [4]. However, this work, is the rst attempt for
experimental study of the mentioned topic. Indeed, HIL is a proper
facility to investigate the effect of optic sensor delay and seeker
mechanism dynamics on the formation keeping. For this purpose,
a three dimensional simulation for the leader-follower formation
control is presented and then a seeker hardware is used in respect
to the simulation using a HIL facility. In the proposed HIL setup, we
xed the seeker on a stand and used a monitor in front of it. The
monitor shows a bright spot as the leader image with respect to
the follower. The seeker camera locks on this spot and tracks it to
regulate the desired relative angles in the formation. The real time
HIL simulation results are compared with those of the fully simulated system which just contain a kinematic model of the seeker.
Therefore, effects of the electro-optic seeker dynamics on the formation can be studied carefully.
The structure of the paper is as follows. Section 2 describes
the leader-follower system equations in which a point mass model
for the follower is assumed and the relative kinematics is expressed. Section 3 introduces the controller structure which is used
for formation ight. This is followed by Section 4 which explains
the experimental setup of the hardware in the loop implementation for the airborne seeker in relation with the simulation of the
leader-follower system. Section 5 presents the results of the realtime hardware-in-the-loop testbed for formation control of two
unmanned aircraft. Finally Section 6 summarizes the results and
provides concluding remarks.
= vF cos V cos V ,
= vF sin V cos V ,
= vF sin V ,
= axV ,
a
zV = yV ,
vF
a
yV = zV ,
x F
y F
z F
v F
(1)
vF
Please cite this article as: M.A. Dehghani et al., A hardware in the loop simulation testbed for vision-based leader-follower formation
ight, Mechatronics (2016), http://dx.doi.org/10.1016/j.mechatronics.2016.10.015
ARTICLE IN PRESS
JID: MECH
Fig. 5. The control structure for regulation of the relative angle in yaw channel.
Fig. 6. The control structure for regulation of the relative angle in pitch channel.
(4)
rL = rL
2
2
+ zL
+ axL ,
yL
ayL
r
xL yL ,
rL
rL
azL
r L
= 2 yL
+ xL zL ,
rL
rL
zL = 2 L zL +
yL
(2)
vxL
r L
xL
rL
r L
vyL = 0 + yL 0 = rL zL
vzL
0
zL
0
rL yL
where k11 , k12 , k21 , k22 , k31 and k32 are control gains that should be
selected as proper constants and, ( LVd , LVd ) and rLd are desired
values of the relative angles and the relative distance. By using this
control structure, the relative orientation and the relative range can
be regulated to the desired values and therefore results in the desired formation. The proposed cascade control structure for regulation of the relative distance, the relative angle in yaw channel,
and, the relative angle in pitch channel are depicted in Figs. 46,
respectively.
Cascade control is designed to allow the outer loop controller to
respond to slow changes in the relative distance and the relative
angles, while the inner loop controller controls disturbances that
happen quickly in angle rates or in the other words, speed loops.
4. Hardware in the loop testbed
where [axL ayL azL ] is the LOS acceleration with respect to the
frame L and [xL yL zL ] is the angular velocity components
in the frame L. Eqs. (1) and (2) describe the leader-follower kinematic model and are used for three dimensional simulation of the
leader-follower formation ight.
Now, by using the following Coriolis formula
(5)
(3)
in which vxL
vyL vzL is the LOS velocity with respect to the
frame L, the relative kinematic Eq. (2) can be restated as follows:
Please cite this article as: M.A. Dehghani et al., A hardware in the loop simulation testbed for vision-based leader-follower formation
ight, Mechatronics (2016), http://dx.doi.org/10.1016/j.mechatronics.2016.10.015
JID: MECH
4
ARTICLE IN PRESS
Fig. 9. Simulated model for the leader-follower formation control loop in the Simulink environment.
Please cite this article as: M.A. Dehghani et al., A hardware in the loop simulation testbed for vision-based leader-follower formation
ight, Mechatronics (2016), http://dx.doi.org/10.1016/j.mechatronics.2016.10.015
JID: MECH
ARTICLE IN PRESS
ss = os os ,
Please cite this article as: M.A. Dehghani et al., A hardware in the loop simulation testbed for vision-based leader-follower formation
ight, Mechatronics (2016), http://dx.doi.org/10.1016/j.mechatronics.2016.10.015
ARTICLE IN PRESS
JID: MECH
6
Fig. 13. The HIL result of yaw channel LOS rate in comparison with the pure simulation.
Fig. 14. The HIL result of pitch channel LOS rate in comparison with the pure simulation.
and
S
V
os = os = [
0] ,
ss = VS C S os V os
cos V S cos V S
sin V S
cos V S sin V S
=
sin V S cos V S
cos V S
sin V S sin V S
sin V S
0
cos V S
0 0
0
0
xss
yss ,
zss
s T = V sT V ss = [xs T
V
ys T
zs T ]T ,
st = V ss + V s t =
xss
yss
zss
d xss
ys T
+ xs T (d xss )
z
sT
(d xss )
xs T
d
ys T (d xss ) + yss
.
= xs T
zs T
(d xss ) + zss
xs T
Therefore to compensate the mentioned error, 2nd and 3rd arrays of this vector are calculated and considered in transmitting
the leader location to the image generator.
Please cite this article as: M.A. Dehghani et al., A hardware in the loop simulation testbed for vision-based leader-follower formation
ight, Mechatronics (2016), http://dx.doi.org/10.1016/j.mechatronics.2016.10.015
ARTICLE IN PRESS
JID: MECH
Fig. 15. The HIL result of yaw channel relative angle in comparison with the pure simulation.
Fig. 16. The HIL result of pitch channel relative angle in comparison with the pure simulation.
5. Experimental results
[V IV ] =
In this section to evaluate the performance of the seeker in the
formation ight of two unmanned aircraft, the HIL results are compared with the results of the pure simulation. In the pure simulation, just kinematics of the seeker measurements are included and,
processing delay of the camera and dynamic of the seeker mechanism are ignored. To model seeker measurements using inertial
position and velocity of the leader and the follower, at rst, based
on the quaternion equations
P =
V
1 IV
2
V TIV
IV
0
P,
zV
yV
xV
yV
xV ,
0
where
zV
0
RL = VI C I RT I RF ,
and by considering
P = [ p1
p2
p3
p4 ]T ,
RL = VL C L RL
Please cite this article as: M.A. Dehghani et al., A hardware in the loop simulation testbed for vision-based leader-follower formation
ight, Mechatronics (2016), http://dx.doi.org/10.1016/j.mechatronics.2016.10.015
ARTICLE IN PRESS
JID: MECH
8
=
=
we have
cos LV cos LV
sin LV cos LV
sin LV
sin LV
cos LV
0
rL cos LV cos LV
rL sin LV cos LV
rL sin LV
LV = atan
cos LV sin LV
sin LV sin LV
cos LV
rL
0
0
A1
A2 ,
A3
A3
A21 + A22
A
LV = atan2 2 ,
A1
rL =
I C VC
V
L
Fig. 17. The HIL result of relative distance in comparison with the pure simulation.
r L = vxL ,
zL = vyL /rL ,
yL = vzL /rL .
In the simulation environment, parameters of the controller
(5) are considered as follows:
rLd = 100, LV d = 10 , LV d = 20 ,
k11 = 0.5, k12 = k22 = k32 = 1, k21 = k31 = 50.
A Gaussian white noise with the following specication is observed in the rate gyros measurements:
Fig. 18. The HIL result of 2D trajectory in comparison with the pure simulation.
Field of Regard = 60
Field of View = 5
Maximum Angular Rate = 150 s1
Tracking Ability = 50 s1
Tracking Bandwidth= 10 rad/s
Stability Bandwidth= 100 rad/s
In Fig. 13, the HIL results in comparison with the pure simulation for the LOS rate in the yaw channel are depicted. Since, the
employed hardware require 3 s time for preparing and locking on
the leader spot, in second 3 a little jump is occurred in all the
signals. However, the behaviors of the signals are similar and the
error is not so much.
Note 1. Since in a typical ight scenario, just one time locking on
the leader aircraft is needed; therefore, the 3 s delay occurs in the
rst stage of the ight (before locking), and it is not related to the
ight path of the leader. Indeed, as the seeker locks on the leader,
locking will be kept independent of the leader trajectory.
Please cite this article as: M.A. Dehghani et al., A hardware in the loop simulation testbed for vision-based leader-follower formation
ight, Mechatronics (2016), http://dx.doi.org/10.1016/j.mechatronics.2016.10.015
JID: MECH
ARTICLE IN PRESS
[11] Kendalla IR, Jonesb RP. An investigation into the use of hardware-in-the-loop
simulation testing for automotive electronic control systems. Control Eng Pract
1999;7(11):134356.
[12] Ren W, Steurer M, Baldwin TL. Improve the stability and the accuracy of power
hardware-in-the-loop simulation by selecting appropriate interface algorithms.
IEEE Trans Ind Appl 2008;44(4):128694.
[13] Li H, Steurer M, Shi KL, Woodruff S, Zhang D. Development of a unied design, test, and research platform for wind energy systems based
on hardware-in-the-loop real-time simulation. IEEE Trans Ind Electron
2006;53(4):114451.
[14] Steurer M, Edrington CS, Sloderbeck M, Ren W, Langston J. A megawatt-scale
power hardware-in-the-loop simulation setup for motor drives. IEEE Trans Ind
Electron 2010;57(4):125460.
[15] Martin A, Emami MR. An architecture for robotic hardware-in-the-loop simulation. In: Proceedings of the 2006 IEEE international conference on mechatronics and automation; 2006. p. 21627.
[16] Chhabra R, Emami MR. A holistic concurrent design approach to robotics using
hardware-in-the-loop simulation. Mechatronics 2013;23(3):33545.
[17] Korpela C, Orsag M, Oh P. Hardware-in-the-loop verication for mobile manipulating unmanned aerial vehicles. J Intell Robot Syst 2014;73(1):72536.
[18] Cai G, Lee BM, Chen TH, Dong M. Design and implementation of a hardware-in-the-loop simulation system for small-scale uav helicopters. Mechatronics 2009;19(7):105766.
[19] Gans NR, Dixon WE, Lind R, Kurdila A. A hardware in the loop simulation platform for vision-based control of unmanned air vehicles. Mechatronics
2009;19(7):104356.
[20] Prabowo YA, Trilaksono BR, Triputra FR. Hardware in-the-loop simulation for
visual servoing of xed wing uav. In: The 5th international conference on electrical engineering and informatics; 2015. p. 24752.
[21] Abdo MM, Vali AR, Toloei AR, Arvan MR. Stabilization loop of a two
axes gimbal system using self-tuning pid type fuzzy controller. ISA Trans
2014;53(2):591602.
[22] Zhan ST, Yan WX, Fu Z, Pan G, Zhao YZ. Robust control of a yaw-pitch gimballed seeker. Aircraft Eng Aerosp Technol 2015;87(1):8391.
[23] Dehghani MA, Menhaj MB. Takagi-sugeno system for supervisory formation control of seeker mounted unmanned aerial vehicles. Assembly Autom
2016;36(2):11119.
[24] Dehghani MA, Menhaj MB. Integral sliding mode formation control of
xed-wing unmanned aircraft using seeker as a relative measurement system.
Aerosp Sci Technol 2016;58:31827.
[25] Evans DJ. On the representatation of orientation space. Mol Phys: Int J Interface
Between Chem Phys 1977;34(2):31725.
[26] Abdo MM, Vali AR, Toloei AR, Arvan MR. Improving two axes gimbal seeker
performance using cascade control approach. Proc Inst Mech Eng Part G J
Aerosp Eng, 2014. doi: 01177/0954410014525130.
Please cite this article as: M.A. Dehghani et al., A hardware in the loop simulation testbed for vision-based leader-follower formation
ight, Mechatronics (2016), http://dx.doi.org/10.1016/j.mechatronics.2016.10.015
JID: MECH
10
ARTICLE IN PRESS
Mohammad Ali Dehghani received his M.Sc. degree in electrical engineering, control eld from K.N.Toosi University of Technology in 2010. At the present he is a Ph.D.
candidate in electrical engineering at Amirkabir University of Technology. His research interests include control of nonlinear systems, ight guidance, Automation and Instrumentation.
Mohammad Bagher Menhaj received his Ph.D. degree from the School of Electrical and Computer Engineering at Ohio State University (OSU) in 1992. After completing one
year with OSU as a post-doctoral fellow in 1993, he joined Amirkabir University of Technology, Tehran, Iran, where he is currently a Full Professor. From December 20 0 0 to
August 2003, he was with School of Electrical and Computer Engineering and Department of Computer Science at OSU as a visiting faculty member and research scholar.
His research interests include complex networks, multi-agent systems, nonlinear adaptive control, articial neural networks, and computational intelligence.
Hadi Ghaderi received his M.Sc. degree in electrical engineering, control eld from K.N.Toosi University of Technology in 2012. His research interests include control of
nonlinear systems, ight guidance and estimation theory.
Please cite this article as: M.A. Dehghani et al., A hardware in the loop simulation testbed for vision-based leader-follower formation
ight, Mechatronics (2016), http://dx.doi.org/10.1016/j.mechatronics.2016.10.015