2012 Brazilian Robotics Symposium and Latin American Robotics Symposium
Simulation Platform for Quadricopter: Using Matlab/Simulink and X-Plane
Helosman V. Figueiredo Osamu Saotome
Aeronautical Mechanics Engineering Division Electronics Engineering Division
Instituto Tecnológico de Aeronáutica - ITA Instituto Tecnológico de Aeronáutica - ITA
São José dos Campos, Brasil São José dos Campos, Brasil
[email protected] [email protected] Abstract— The unmanned aerial vehicles (UAVs) has grown present high risk of dangerous accidents. Therefore, before
in military and civilian areas of application. Several being embedded, autopilot systems must be thoroughly
industries (automotive, military, factories, space, etc.) use tested in the laboratory [5].
robots for dangerous and repetitive tasks. This paper is Several platforms have been developed for UAV
dedicated to a special type of aerial platforms, the simulation, but they are focused on complex vehicles and
quadricopter. This platform has been highlighted by having integration of their systems. Great effort and development
unique characteristics in flight and construction, for time are required to build a custom configuration for
example, hovering flight, vertical takeoff and landing, high specific integration of sensors and normally complex
maneuverability, low speed flight, and simple mechanics. software is required to operate. Many of them do not
This work proposes a tool for simulation and visualization of support modifications of hardware or software, and to
an aerial air-type quadricopter robots, using the flight update or to modify it, one spends almost the same time
simulator X-Plane 9, Matlab and Simulink. The aircraft necessary to develop a new UAV. Besides the simulation
under study is control the so-called ITA-001, developed by
platform related problems, it is also difficult to know the
the quadricopter study group the Aeronautical Institute of
Technology – ITA.
actual meteorological conditions along the route [6].
Some authors have addressed the topic of simulation
Keywords- Quadricopter, Simulation, X-Plane, Matlab platform for UAVs. The reference [6] treats the problems
of small fixed wing UAVs, analyzing the control and
navigation algorithms. The reference [7] treats the problem
I. INTRODUCTION of the ducted type fan UAVs, and proposes a concept of
Currently, the use of unmanned aerial vehicles (UAVs) simulation platform based on the commercial flight
has increased both in military and civilian areas, especially simulator X-Plane 9 and Matlab \ Simulink. The reference
in missions where human operation is unnecessary, [8] is a simulation platform for multi-UAV helicopters,
repetitive or dangerous. These are employed in diverse and proposes a simulator for test and evaluation of control
areas such as surveillance, photography, traffic monitoring, algorithms for stabilization of the aircraft and formation of
identification of pests in agriculture and educational the swarm. According to our knowledge, none of the
platforms [1]. recent works presents a simulation platform for
UAVs can be divided into two groups: fixed-wing and quadricopter control and navigation and prepared for
rotorcraft. Fixed wing flight vehicles are suitable for swarm research, at the point of view of its embedded
outdoor use and can cover large area. The rotary-wing system.
vehicles have features such as: increased maneuverability, For the reasons mentioned above and for not having
vertical takeoff and landing and hovering flight, enabling found a platform to quadricopters simulations, with
flight at low altitudes and indoor application [2]. For the support for autopilot tests, ability to customize features of
group of rotorcrafts, a configuration that is gaining the aircraft and to simulate real conditions (such as winds,
prominence is the quadricopter, which consists of four turbulences, etc.), we propose in this article a platform for
independent blades arranged in a cross shape. simulation and visualization of quadricopter type aircraft
The concept of quadricopter appears in 1907 with using the commercial simulator X-Plane 9 and Matlab
Breguet brother’s and professor Richet work [3]. Their Simulink. The main objective of this platform is to test and
aircraft concept was large and heavy, and it was not to help the development process of autopilot systems for
possible to make a flight with large payloads or large quadricopter type aircraft. The quadricopter ITA-001 will
distance flights. Since then, rotorcraft research had been be used as an object under test of this study.
developed, but the concept of quadricopter was forgotten. This article is organized as follows: section 2 describes
In recent years, the development of lightweight building the quadricopter ITA-001, section 3 presents the
materials, the improvement of the relation between power simulation platform, section 4 describes the data interface
and weight, and the miniaturization of the engine and for the simulation, section 5 describes the implementation
control systems, has been factor that turned the platform, section 6 is about platform application and
quadricopter feasible. Since then, several civil and military section 7 contains conclusions and suggestions for future
institutions are researching and developing this type of works.
aircraft [4].
The autopilot is a major component of a UAV. The II. QUADRICOPTER ITA-001
design of automatic pilot systems requires many The quadricopter is an aircraft with four independent
simulations and tests on real aircraft. The real aircraft tests rotors, fixed at the end of each axle as shown in Figure 1
978-0-7695-4906-4/12 $26.00 © 2012 IEEE 51
DOI 10.1109/SBR-LARS.2012.15
[4]. In order to cancel the moment generated by the
rotating blades, a pair of propellers rotates clockwise and
the second pair in a counterclockwise direction.
Figure 3. Control System
III. SIMULATION PLATFORM
This platform has been implemented using the flight
simulator X-Plane 9 and Matlab. The chosen means of
communication between software is UDP (user datagram
protocol). The Matlab/Simulink runs the autopilot and
processes data flight for analysis. The X-Plane simulates
the dynamics of the aircraft and displays the 3D view of
the Simulation. The Simulation Platform concept is
presented in Figure 4.
Figure 1. Configuration quadricopter withdrawal [4]
A. ITA-001 Description
The model ITA-001 was based on an experimental
structure, assembled from parts of commercial RC
(Remote Controlled) airplanes, as shown in Figure 2. Its
structure is made of aluminum rods, together with pieces
of acrylic. It contains four brushless motors located on the Figure 4. Simulation Platform Concept
ends of the rods, four controllers engine and a controller
for stabilization. A. X-Plane flight Simulator
This simulator was chosen for the following
characteristics:
• Certification by Federal Aviation Administration
(FAA).
• Plane Maker: tool for drawing and modifying
aircraft.
• Has been used as the basis for designing and
testing control algorithms.
• Has a god interface for data injection and data
extraction.
• Accurate flight models.
The X-Plane uses the geometric shape modeled in
Plane Maker to see how the aircraft will fly. He uses an
Figure 2. ITA-001 engineering method called blade element theory, which
means to divide the aircraft into small elements and then
B. ITA-001 Control System find the forces acting on each element several times per
second [9].
The quadricopter ITA-001 features two control X-Plane native method of communication (import and
problems: glide control and path control. The first control export data) is the UDP protocol. This protocol is a non-
system is used to obtain a stable operation hover. The guaranteed protocol and gives no assurance that data
second control system is to perform missions that require packets arrived in order or didn’t arrive at all, so it may
to travel certain paths. Figure 3 shows the control system present a possible problem resulting from data corruption.
for a quadricopter. [8]
This simulator has configurations where it is possible
to view the various forces acting on the aircraft, the path
traveled by the aircraft and, also, allows the introduction of
in-flight failures.
B. Matlab/Simulink
Matlab/Simulink® is an environment for multi-domain
simulation and Model-Based Design for dynamic and
52
embedded systems. It provides an interactive graphical V. PLATFORM IMPLEMENTATION
environment and a customizable set of block libraries that This platform has been implemented using two desktop
let you design, simulate, implement, and test a variety of computers. The first computer runs the first flight
time-varying systems, including communications, controls, simulator X-Plane and the second one runs control system
signal processing, video processing, and image processing and processing, and also analysis of simulation data. The
[10]. computers communicate through Ethernet port via UDP
Simulink is integrated with MATLAB®, providing protocol.
immediate access to an extensive range of tools that let you For the implementation it was necessary to build the
develop algorithms, analyze and visualize simulations, quadricopter ITA-001 model for X-Plane 9. To this
create batch processing scripts, customize the modeling implementation the tool Plane-Maker was used. For the
environment, and define signal, parameter, and test data modeling process, we made use of the Plane-Maker
[10]. Manual [11], and Tutorial Plane-Maker [12].
In the simulation platform, the software Matlab / Making ingenious use of Plane-Maker special options,
Simulink is responsible for processing the UDP data, and also taking advantage of symmetry existing in the
control system for glide and post processing of data system, the modeling was performed successfully, as
simulation. illustrated in Figures 7 and 8.
IV. DATA INTERFACE
The UDP protocol is chosen because it is the standard
system for sending and receiving data from the X-plane
and, also, it is compatible with Matlab / Simulink.
The speed of communication is a point of great
importance, because the control commands must be
synchronized with the simulation in X-plane. The X-Plane
has capacity to send 99.9 packets per second, but most of
the inertial measurement systems operate at a rate of 40 to
60 times per second. Therefore, we chose to work with 40
packets per second.
The type and size of the UDP packet depends on the
amount and types of data to be exported. The X-Plane has Figure 7. ITA-001 X-Plane model
a interface to set the output and receiving data, which is
shown in Figure 5.
Figure 8. ITA-001 Cockpit Vision
Figure 5. UDP Configuration X-Plane 9
VI. APPLICATION
The data packet sent by the X-plane follows this To demonstrate the use of this as a development tool
pattern: the first four bytes represent the type of packet, the for autopilot systems, control system for hovering was
fifth byte is an internal policy, the next four bytes indicate designed and implemented in simulation platform, The
what type of parameter is being sent, the following four high level hovering control system block diagram is shown
bytes represent the value in single-precision float. An in Figure 9.
example of packet data is shown in Figure 6.
Figure 6. UDP Pack
Figure 9. High Level Hovering Control
53
The quadricopter dynamics is simulated at X-Plane,
and the autopilot blocks are inserted in the block control
system. Figure 10 presents a detailed vision of the control
system block.
Figure 12. Altitude response
For this simulation, the quadricopter starts from the
ground, and then put the reference signals to hover in 470
meters. The control loops gain were tuned empirically. In
order to ease this work, we designed a graphical user
interface shown in Figure 13.
Figure 10. Detailed view of the control system
The reference signal applied to the system and the
response to this stimulus is presented in Figure 11. For
hovering, the reference signals of the angular positions
were set to zero, the reference altitude is set to 476 meters.
Altitude response is shown in Figure 12.
Figure 13. Adjust interface
A method to obtain better answers is to do more
simulations for making fine adjustments in the controller.
An example of the utilization of this platform as a
design tool for an autopilot system in context of an
embedded autonomous system is the analysis of the
frequency control loop. This analysis is important because
for small aircrafts, such is the case of quadricopters, power
is supplied by batteries and payload are very restricted,
making critical the choice of sensors and microcontrollers
that will be embedded. The frequency control loop can be
limited by the speed of data acquisition for feedback or the
Figure 11. Angular Positions response processing capacity of the microcontroller where the
control system was embedded.
As a design tool we could use this platform and test the
control system in various frequencies. As an example we
use the hovering control system previously proposed, and
the response of the system at various frequencies,
generated by the computer. Figure 14 presents the system
response to altitude.
54
[5] L. R. Ribeiro and N. M. F. Oliveira, “UAV autopilot controllers
test platform using Matlab/Simulink and X-Plane,” 2010 IEEE
Frontiers in Education Conference (FIE), p. S2H-1-S2H-6, Oct.
2010.
[6] S. Santos, C. Nascimento, S. Givigi, A. Bittar, and N. Oliveira,
“EXPERIMENTAL FRAMEWORK FOR EVALUATION OF
GUIDANCE,” COBEM, 2011.
[7] T. Indriyanto and Y. I. Jenie, “Modeling and Simulation of a
Ducted Fan Unmanned Aerial Vehicle ( UAV ) Using X-Plane
Simulation Software,” 2010.
[8] R. Garcia and L. Barnes, “Multi-UAV Simulator Utilizing X-
Plane,” Journal of Intelligent and Robotic Systems, vol. 57, no. 1–
4, pp. 393-406, 2009.
[9] L. Research, “X-Plane Operation Manual,” 2011. [Online].
Available: http://www.x-plane.com/files/manuals/X-
Plane_Desktop_manual.pdf. Accessed 10/02/2012.
[10] Mathworks, “Simulink Overview,” 2012. [Online]. Available:
http://www.mathworks.com/products/simulink/. Accessed
15/05/2012.
Figure 14. Altitude Response [11] L. Research, “Plane Maker Manual,” 2011. [Online]. Available:
http://www.x-plane.com/files/manuals/Plane_Maker_manual.pdf.
Analyzing the resultant curves of Figure 14, we find Accessed 10/02/2012.
that for low frequencies the system shows a larger [12] A. Bittar, “Tutorial do Plane-Maker,” São José dos Campos, 2011.
overshoot and longer accommodation. Increasing the
frequency of operation, we can observe that the maximum
peak and the settling time are both lowering with the
increase of frequency. This experiment is an example of
how this simulation platform can generate the minimum
requirements for the design of a prototype, given
appropriate conditions.
This platform is also a good tool for design and
analysis of algorithms for navigation, because it allows the
test of these algorithms without the utilization of real
aircraft, thus reducing the development time and system
costs.
VII. CONCLUSION
This test platform is a great tool for design and study
systems for quadricopter type UAV autopilot. Also it is
possible to monitor the response of the aircraft in a variety
of flight conditions.
This simulation tool is useful also to try easily various
techniques of control and navigation algorithms for
hardware-in-the-loop tests. This architecture also enables
simulating quadricopter swarms and interactions with
vehicles on the ground.
To continue this work, the simulation results will be
compared with the non-linear model of quadricopter, then
with the real quadricopter, thus increasing the reliability of
the results obtained in the simulation.
REFERENCES
[1] S. Bouabdallah and R. Siegwart, “Full control of a quadrotor,”
2007 IEEERSJ International Conference on Intelligent Robots and
Systems, no. 1, pp. 153-158, 2007.
[2] P. McKerrow, “Modelling the Draganflyer four-rotor helicopter,”
IEEE International Conference on Robotics and Automation, 2004.
Proceedings. ICRA ’04. 2004, no. May, pp. 3596-3601 Vol.4,
2004.
[3] J. G. Leishman, “A History of Helicopter Flight,” 2000. [Online].
Available:
http://terpconnect.umd.edu/~leishman/Aero/history.html.
[4] R. Goel and S. Shah, “Modeling, Simulation and Flight Testing of
an Autonomous Quadrotor,” of the IISc, 2009.
55
