2014 IEEE 2014 International Conference on Computer, Communication, and Control Technology (I4CT 2014), September 2 4, 2014 - Langkawi,

Kedah, Malaysia

Altitude Controller Design For Multi-Copter UAV

Zaki Mustapa, Shakir Saat, S. H. Husin, Norafizah Abas
Faculty of Electronic and Computer Engineering
Universiti Teknikal Malaysia Melaka, Malaysia

performance of automatic take off system indicates how fast

the quadcopter achieves its desired altitude in steady state. The
performance of the system also signifies how smooth it
maintains the height of quadcopter.
E. Abbasi (2012) was making simulation using fuzzy
controller for tuning the PID Gains in Hovering Mode[4]. The
efficiency of the PID controller is changed if the target height
is changed. Therefore, the kP, ki and kD also be adjusted to
response was very nice.
An automatic landing system requires landing the
quadcopter safely. Safe landing is much more critical which is
the uncontrolled decrease of motor speed which affects the
body drop drastically and leads to unsafe landing.
Automatic navigation system deals with trajectory program
which allows the quadcopter to have its own algorithms to
work autonomously[5]. A possible way to navigate the UAV to
land properly on the platform is provided using vision
feedback. To achieve this, single-down looking camera may be
located at UAV to navigate the copter autonomously and detect
an elevated landing platform as a surrogate for a rooftop. The
landing platform detection system uses a planar holographic
decomposition to detect landing targets and produce approach
waypoints for autonomous landing.
This paper will be presented a method to design
Quadcopter altitude control system. Additionally, detailed
model of the quadrotor is given. The controller designed will
be implemented in real time model. As in previous work, a
mathematic model of quadcopter was explained and tested to
simulate the behavior of the quadcopter. The result parameters
of the work were used in designing the altitude controller.
In previous work, a mathematic model of quadcopter was
explained and tested to simulate the behavior of the
quadcopter. The result parameters of the work were used in this
paper for designing the altitude controller[6].

Abstract This paper discusses on attitude control of a

quadcopter unmanned aerial vehicle (UAV) in real time
application. The paper describes the controller design method for
the hovering control of UAV automatic vertical take-off system.
In order to take-off the quadcopter and stable the altitude, PID
controller has been designed. The scope of study is to develop an
altitude controller of the vertical take-off as realistic as possible.
The quadcopter flight system has nonlinear characteristics. A
simulation is conducted to test and analyze the control
performance of the quadcopter model. The simulation was
conducted by using Mat-lab Simulink. On the other hand, for the
real time application, the PCI-1711 data acquisition card is used
as an interface for controller design which routes from Simulink
to hardware. This study showed the controller designs are
implemented and tuned to the real system using Real Time
Windows Target approach by Mat-Lab Simulink.
Keywords Quadcopter modelling, Altitude
Helicopter Test, PID Controller, UAV Multi Rotor.

Control,

INTRODUCTION
The Unman Aerial Vehicle (UAV) robot has gained
demand on the worldwide markets. There are two types of
UAV which is fixed wing and rotary wings. Nowadays, one of
the most popular rotary wings UAV is Quad copter. It has
more advantages in term of their capability to take off and
landing in vertical condition [1]. In addition, it has good
mobility, simple mechanics, and ability on load capacity.
The quad copter system is nonlinear system, which is the
speed of the motor are varies the torque and force onto the rigid
body of quad copter. The force and torque that produced by the
motor propeller will determine the behavior of the quad copter
[2].
Development of autonomous UAV has been growing fast
as it is more reliable in application. The system can be
categorized into three types, i.e. automatically take off system,
automatic landing system, and the automatic navigation
system. This project aims to develop a less-human interaction
UAV system that is capable of performing vertical take-off and
landing autonomously.
The automatic take-off system required to enable the UAV
begin to fly from the ground. The system should ensure the
height of quadcopter in a consistent condition. In these cases,
each four motor propeller must produce a force against the
gravitational force and able to lift the quadcopter[3]. The better

978-1-4799-4555-9/14/$31.00 2014 IEEE

sin

sin

cos

sin cos

(1)

382

Equation 1 is about the mathematical model that used to

represent the quadcopter system. The first 3 equations are
contributing to linear vector and the next 3 contribute to
angular velocity vector for the quadcopter rigid body system.

Where:m=mass of quad copter body, (kg)

g=acceleration due to gravity, (m/s2)
T=period of swing in the 1 cycle (s) from IMU data.
b=radius between the rope point (m)
L=the rope length (m)

(2)

The Quad copter design is symmetrical. The inertia for rolling

is the same as that for pitching movement. Table 1 shows the
value of each constant parameter in bifilar test.

Equation 2 represents all the input variables in producing

the movement of the system. Each vector represents the thrust,
rolling, pitching, and yawing rotation respectively.

TABLE 1: CONSTANT PARAMETER

Parameter

PHYSICAL PARAMETER DETERMINATION

After deriving quadcopter model equations, its practical
parameters have to be identified. The parameters to be
identified are constants value in the model equations. The
parameter are, the acceleration of gravity (g), mass (m) of
quadcopter body, and inertial (I) parameters; aerodynamic
parameters of constant thrust factor (b), drag factor (d), and the
distance between the motor and the quad Copter centers [8].
There are three inertia parameters of movements i.e. rolling
due to x-axis, pitching due to y-axis and yawing due to z-axis.

Value

Unit

0.80

kg

9.81

m/s2

0.20

1.50

Table 2 shows the result for the moment of inertia. It

consists of three rotation movements i.e. rolling, pitching and
yawing rotation. They will be used in control system design.
TABLE 2: MOMENT OF INERTIA VALUE

Moment of inertia
There are 2 methods to determine inertia parameter which
use moment of inertia calculation and by bifilar pendulum
experiment. This paper focused of the bifilar pendulum
experimental method. Figure 1 shows the bifilar pendulum
system for measuring the value of inertia.

Moment
of inertia

Period per
cycle (s)

Value
(kgm2)

Ix

1.40

0.0398

Iy

1.40

0.0398

Iz

1.85

0.0421

Thrust force factor

Thrust force factor can be determined using different 2
methods, one is by blade element theoretical calculation, and
another is by force lift test experiment. In this paper, force lift
test method is chosen, because the blade element analysis
requires a lot of environmental and blade specification
variables, and the experiment deals with real situation, hence
the parameter can be much more reliable [10]. Figure 2 shows
the method to measure the relationship of force produced by
the speed of propeller.
Tachometer

Figure 1: Bifilar pendulum test[9]

In bifilar experiment, two ropes are used to string up the
quadcopter body, next the quad copter body is held at a certain
degree by hand, and then it is released. Due to the tension of
the rope, the system is swung about the vertical axis of the
system. The inertia can then be calculated with bifilar equation
[9].
(3)

383

Servo tester
(PWM generator)

ESC

Scope

12Vdc

Figure 2: Lift test method for Brushless motor

The constants of quad copter system which include the mass

of the quad copter (m) is 0.8kgg, the thrust factor (b) is 1.50
and the gravity acceleration(g)) is 9.81. In this ideal case, roll
and pitch angles are approximaated to be zero.

Electronic speed controller is used and trriggered by PWM

generator. Propeller speed will be determ
mined with micro
tachometer. The thrust force is recorded by changing the
propeller speed.

Result
Figure 5.1 shows the positivee changing of position for the
speed of propeller at 400rad/s.

p
Figure 3: Force VS speed of propeller
ween thrust forces
Figure 3 shows the relationship betw
produced by propeller speed. Theoretical aerodynamic has
already proved that thrust force varies prroportionally with
squaring velocity. To define the constannt values of the
relationship, equation below is applied.
1.50
0

Thus;

10

Figure 5.1: The position change when =400rad/s

Figure 5.2 shows the negative changing
c
of position for the
propeller speed at 390rad/s.

(4)

According to aerodynamic theory, drag factor depends on

tan , where is
i attack angle of
lift factor by a ratio
propeller [11]. The value of equal to 15, and drag factor is
equal to
4.02 10 .
S
ALTITUDE ANALYSIS
Open loop simulation
After obtaining each physical parameter from experiments,
their values will be inserted into the modeels equation. The
system can be analyzed in simulation using Mat-lab. The
behavior of altitude can also be observed inn simulation. Base
on the mathematical modeling, figure 4 expplains the altitude
system which is used to analyze the behaviorr of quad copter at
certain propeller speed. The simulation setup is shown below.

c
when =390 rad/s
Figure 5.2: The position change
h
motor speed (H) of the
From analysis, minimum hovers
quad copter body could be deteermined. From basic movement
concept, the speed is used to be the reference speed of each
altitude control. The result shoows the minimum speed of each
motor to thrust the load of quaad copter body is
400
/
, for the payload of 0.8kg, andd the thrust factor 1.50 .
Close loop with PID controllerr simulation
The next step deals with deesigning the control system. The
control system design will be
b tested in simulation using
SimulinkThe parameters of the control system can be tested.
P controller setup for the test.
Figure 6 shows the close-loop PID

Figure 4: The altitude system simulates in Simulink.

Z

cos

cos

(5)
(6)

384

Run a simulation in non-real time to observe the

behavior of the controller in the model.
Create a real-time appplication by generating the code
using real-time workshop code generation software
from the Simulink model.
Run an application in real time by interfacing with the
hardware. Using the PCI1711 data acquisition card.
Analyze and visualizee signal data in the MATLAB
workspace.

Input section
For real time applications, proximity sensor required to
respond for the height in the control system. Proximity sensor
suitable for this experiment is infrared proximity sensors.
Infrared proximity sensor (Sharrp GP2Y0A02YK0F) is capable
to measure the distance betweeen 15cm to 150cm. The signal
produced is in analog form.. However, the analog signal
produced is proportional to non-linear
n
response. Therefore,
analog voltage signal from the sensor response must convert to
distance scale in meters. Figuree 8 shows the Simulink block set
up for the proximity infrared seensor.

Figure 6: The altitude system with PID conttroller simulation

The simulation does not take hardwarre limitation into
account. Saturation block can be used to declare
d
output and
input limitation in the system. Hence, simulaation result can be
more reliable to be implemented in real time application.

1-D T(u
u)
Analog
Input
Scope

Analog Input
Adv antech
PCI-1711 [auto]

lookup table for 20-150cm

2

Figure 8: The infrared proximitty sensor signal converter block

Figure 7: PID controller responnse.

Controller parameters for PID; proportionnal = 50, integral =
30, derivative = 10 and filter coefficient = 100. The response
shows the controller system takes 2 second to reach the target
height of 1 meter.
S
ALTITUDE ANALYSIS
Real time execution
Creating a real-time application and ussing the Simulink
External mode, real-time workshop code geeneration software,
real-time windows target software, and thhe open Watcom
C/C++ compiler, to produce an executable file
f the kernel can
run in real-time. This real-time application uses the initial
parameters available for Simulink model att the time of code
generation. Based on your selected sampple rate, real-time
windows target software uses interruptss to setup your
application in real time at the proper ratee. With each new
interrupt, the executable file computes the enntire block outputs
from your model. Development processs of a software
component is shown in the following:
AB environment.

Design a control system in MATLA

Create a Simulink model as a block graphical model.

b
data for infrared sensor.
Figure 9: The look up table block
The lookup table block is useed to convert the voltage signal
into a height variable in meeter. The configuration of the
lookup table is same to the innfrared proximity sensor signal
specification. Figure 9 shoows the lookup table block
configuration.
Output section
The output of the system is a brushless DC motor (BLDC).
Brushless DC motor speed is controlled using an electronic
speed controller (ESC). The ESC is functioning as motor
driver. ESC is used to produce 3 phase DC supply to brushless
DC motor. It is working with
w
PWM signal input as a

385

triggering the speed of the motor. To impleement the ESC in

Simulink, PWM signal generator block neeed to design. The
design is shown in figure 10. The frequenccy of the signal is
fixing to 50hz and the duty cycle varies wiith term of analog
number.

Scope3

Scope1
<= 0

Digital
Output

Compare
C
T Zero4
To

Digital Output

t
Clock

Signal
Generator1

Adv antech
PCI-1711 [auto]

s
simout

reference

To Workspace
W

Scope3

<= 0

Saturation1
Signal
Generator1

Compare
To Zero4

S
Scope

Step1
for 1 meter2

>= 0.5
Step

simout1

for 1 meter1
Clock

Switch

T o Workspace1

1-D T(u)

Figure 10: PWM signal generator blockk in Simulink

Analog
Input

PID(s)

(u/(6*10^(-5))^(1/2)

Analog Input

PID Controller

Fcn

Adv antech lookup table for 20-150cm

PCI-1711 [auto]

height input

Figure 12: Altitude controols RTWT Simulink block.

CLUSION
CONC
The analysis of the altitudde quad Copter system can be
realized by discussing relevant model description. The result of
the analysis helps others to staart designing a controller that is
compatible with the quad copteer system. Each model needs to
be analyzed by undergoing simulation to understand the
behavior of a system. The systematic procedure for the
controller design of Quad-copteer design has been presented.

Figure 11: PWM signal

ESC has to calibrate first, to set up thhe minimum and
maximum PWM duty cycle to operate.

Acknowledgement
The authors would like to thank Universiti Teknikal
Malaysia Melaka (UTeM) for sponsoring this work under the
PJP grant (PJP/2013/FKEKK(17C)S1204).

Altitude control
The Altitude control system consists of three sections: the
first one is the input, second is output and thhe third is control.
PID control system parameters are taken froom the simulations
performed in the previous session. Figurre 13 shows the
combined block to perform the controlller in real time
application.
Figure 12 is about single output motor control. After the
PID block, function block is used to converrt the parameter of
force that produced by the PID controller into motor speed
u
to initiate the
variation in analog number. Switch block is used
ESC by providing the minimum signal vaalue for the 2s in
period.

Reference
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]

386

M. D. L. C. de Oliveira, Moodeling, Identification and Control of a

Quadrotor Aircraft, 2011.
S. Bouabdallah and R. Siegw
wart, Full control of a quadrotor, in
IEEE/RSJ International Confference on Intelligent Robots and
Systems, 2007, no. 1, pp. 1533158.
T. Bresciani, Modelling , Iddentification and Control of a Quadrotor
Helicopter, Department of Automatic
A
Control, Lund University,
2008.
A. Guclu and IN, Attitude and
a altitude control of an outdoor
quadrotor, Atlim Universityy, 2012.
B. Erginer and E. Altug, Moodeling and PD Control of a Quadrotor
VTOL Vehicle, in 2007 IEE
EE Intelligent Vehicles Symposium,
2007, pp. 894899.
a. Bousbaine, M. H. Wu, andd G. T. Poyi, Modelling and simulation
of a quad-rotor helicopter, 6th
6 IET Int. Conf. Power Electron.
Mach. Drives (PEMD 2012),, pp. C34C34, 2012.
H. Lin, T. Lin, H. Yae, D. Opptimization, and I. City, - 1:35, pp.
23222326.
G. Ingram, Wind Turbine Blade
B
Analysis using the Blade Element
Momentum Method . Versionn 1 . 1 List of Figures, 2011.

[9]
[10]
[11]

O. a. Bauchau, J. Rodriguez, and S.-Y. Chen, Modeling the Bifilar

Pendulum Using Nonlinear, Flexible Multibody Dynamics, J. Am.
Helicopter Soc., vol. 48, no. 1, p. 53, 2009.
A. H. Pedersen, Test and Modelling of Four Rotor Helicopter
Rotors, 2006.
I. Maybe, V. Nolte, T. Konzepte, R. Sanders, and E. Cowan, Lift
and Drag, no. March, pp. 19, 2011.

387