Professional Documents
Culture Documents
HELICOPTER
A. Bousbaine*,M. H. Wu*, G. T. Poyi*
*University of Derby, School of Technology, Derby, UK, A.Bousbaine@derby.ac.uk
*University of Derby, School of Technology, Derby, UK, M.H.Wu@derby.ac.uk
*University of Derby, School of Technology, Derby, UK, G.T.Poyi@derby.ac.uk
Keywords: Unmanned Aerial Vehicle, Quad-rotor, Dynamic Inertial Measurement Units (IMUs) and low cost GPS have
Modelling, Simulink, Helicopter. given them roles of enabling technologies for new
autonomous vehicle applications [5].
Abstract
A quad-rotor, one such UAV with four fixed pitch rotors is
Small size quad-rotor helicopters are often used due to the highly manoeuvrable. It has the ability to hover and to take
simplicity of their construction and maintenance, their ability off, fly and land in small areas and also has a simple control
to hover and also to take-off and land vertically. The first step mechanism. However, it is a complex unstable vehicle and
in control development is an adequate dynamic system can be difficult to fly with the absence of modern embedded
modelling, which should involve a faithful mathematical control systems.
representation of the mechanical system. This paper presents
a detailed dynamic analytical model of the quad-rotor 2 The quad-rotor model and system
helicopter using the linear Taylor series approximation
method. The developed analytical model was simulated in the Design and analysis of control systems are usually started by
MatLab/Simulink environment and the dynamic behaviour of carefully considering mathematical models of physical
the quad-rotor assessed due to voltage changes. The model is systems. The model is very important because it gives a
further calibrated and linearized for use on any quad-rotor description of how the system responds to the inputs given to
helicopter. it. An adequate dynamic system modelling should involve a
faithful mathematical representation of the whole system.
1 Introduction
The dynamic behaviour of the quad-rotor is usually described
Helicopters are generally known to be dynamically unstable by ordinary differential equations. Such equations can be
vehicles; hence the need to achieve stability using suitable linearized and the Laplace transform can be used to simplify
control methods. The changing helicopter parameters and the method of solution to reduce computation time. This
complex weather conditions make the vehicle unstable, section describes the derivation of the model of the quad-rotor
although the unstable dynamics are good for providing it with helicopter from [2, 3]. It is possible to define and predict the
the required agility. positions that the helicopter will reach by investigating the
speeds of the four motors, using the equations obtained.
In the recent past, Unmanned Aerial Vehicles (UAVs) have
captured enormous commercial potential by attracting the 2.1 Quad-rotor basic concepts
attention of many people including laymen, potential appliers,
vehicle professionals and researchers. Research and The quad-rotor is a helicopter with four lift-generating
development in this field has gained increasing significance, propellers mounted on four motors. It is very well modelled
due to the emergence of a large number of potential civil with four rotors in a cross configuration style. In most cases,
applications. UAVs are important when it comes to each propeller is directly connected to a brushless DC motor,
performing a desired task in a dangerous and/or inaccessible with all the propellers having fixed and parallel axes of
environment. In fact, several industries including automotive, rotation. Furthermore, it has fixed-pitch blades and the air
medical, manufacturing, aerospace, require robots to replace flow from each of the blades points downwards to generate an
men in dangerous, boring or onerous situations. A wide area upward lift. These considerations show that the structure is
of this research is dedicated to aerial platform [3]. rigid and control of the vehicle can be achieved by varying
the propeller speeds by very small amounts because of the
Primarily enabled by advancements in Artificial Intelligence, high sensitivity of the vehicle to rotor speed changes.
Computer Science, Automatic Control, Robotics,
Communications and Sensor Technologies, UAVs are Two of the motors (front and rear) rotate counter-clockwise,
expected to become a major part of the aviation industry over while the other two (left and right) rotate clockwise. This
the next few decades. Furthermore, the ever-increasing configuration of opposite pairs directions completely
performance of Micro Electro-Mechanical Systems (MEMS) eliminates the need for a tail rotor, which is needed for
stability in the conventional helicopter structure. Figure 1 movements, the total vertical thrust is still maintained as in
shows the model in stable hover, where all the motors rotate hovering; hence this command leads only to a yaw angular
at the same speed, so that all the propellers generate equal lift acceleration.
and all the tilt angles are zero.
2.2 Dynamics of the quad-rotor system
There are two coordinate systems to be considered, figure 2:
The earth inertial frame (E-frame)
The body-fixed frame of the vehicle (B-frame)
These are related through three successive rotations:
Roll: Rotation of around the x-axis;
Pitch: Rotation of around the y-axis;
Yaw: Rotation of around the z-axis.
Figure 1: Simplified quad-rotor vehicle in a stable hover. The following assumptions have been made in this approach:
The origin of the body-fixed frame coincides
There are four basic movements, which allow the helicopter
with the centre of mass (COM) of the body of
to reach a certain altitude and attitude and are;
the vehicle.
The axes of the B-frame coincide with the body
a) Throttle
principal axes of inertia.
The drag torque is proportional to the propeller
This command is provided by simultaneously increasing (or
speed.
decreasing) all propeller speeds by the same amount and at
the same rate. This generates a collective vertical force from
the four propellers, with respect to the body-fixed frame. In
consequence, the quad-rotor is raised or lowered to a certain
altitude.
b) Roll
1 4
.
The dynamics of the motor is described by the following
x Fi [sin sin cos cos sin ]
m 1
equation.
1 4
y 1 Fi [sin sin cos cos sin ] J m d (6)
m
1 4
z 1 Fi [cos cos ] g Where J is the total motor moment of inertia, is the
m
motor angular acceleration, m is the motor torque and d is
(2)
the load torque.
Where Fi is the thrust force generated by motor i, l is the
2.2.2 Voltage and angular velocity of propeller
length of the quad-rotor arm, is the torque produced by
Since the voltage inputs V to motors affect the rotational
each motor, I i s are the moments of inertia with respect to speed of propellers.
the axes and m the mass of the helicopter. V e iZ (7)
The dynamics of the motor is described by the following
2.2.1 Motor dynamics equation
m d J (8)
The basic electrical circuit is shown in figure 3.
Equation (8) states that when the motor torque m and the
drag torque d are not equal there is an acceleration (or
deceleration). This variation of speed depends also on the
rotor inertia J the smaller the value of J , the higher the
acceleration. The back electromotive-force voltage is
proportional to motor speed
e K e (9)
Figure 3: Electrical equivalent circuit of the motor The motor torque is proportional to the field current
m K qi (10)
By applying Kirchhoffs voltage law, it follows that. On substituting, we get
v vZ vL e (3)
Z ZJ (11)
Where v Z is the voltage across the resistor Z and v L is the V d Ke
Kq Kq
voltage across the inductor L. The equation above can be
As stated earlier, the drag torque is proportional to the square
rewritten as in the next equation.
of propellers speed
i
v Zi L K e d D 2 (12)
t (4) The relationship between angular velocity and voltage as
i is the motor current, Ke is called the motor constant and found in [51] and [5] can thus be obtained as
is the motor angular speed. The first term has been changed ZD 2 ZJ
V K e (13)
using Ohms law v Z= Zi , while the second one using the Kq Kq
inductor differential equation v L= i . The contribution of
L
t 2.2.3 Voltage and thrust
the inductor part is important to determine the characteristic Voltage is the input of the quad-rotor plant and each rotor
of the DC-motor driver. However it is often neglected in the produces a thrust force as it turns. The motor torque is known
mechanics computation because of the reasons that follow [2, to be proportional to the field current
3]: m K qi (14)
A greater percentage of the motors used in robotics
show very small inductance because of design m (15)
i
optimization. Kq
The response time of the electrical part is always The electrical power according to Joules law is
much faster than the mechanical part, hence the m
P IV V (16)
speed of the overall system will be determined by Kq
the slowest contribution.
It is much easier to solve a first order differential And the mechanical power output is given as
equation rather than a second order one.
m (17)
Pm P V
Kq b) Pitch Moment
with as the motor efficiency.
The pitch motion is about the y-axis and the pitching moment
The propellers figure of merit f is defined as the ratio of caused by the actions of motors 1 and 3 can be defined as:
the induced power in air Ph to the mechanical power Pm [2].
(31)
I xx ( F1 F2 F3 F4 )l
Ph (18) Just like that of the roll motion, the final equation for pitch
f
Pm motion is obtained as
2
Where Ph is given by 2 Al fK t
2 2 (32)
(V3 V1 )
Ph f m V (19) I xx K q
Kq
The ideal power is the product of the thrust force and the c) Yaw moment
speed at which it is applied. At hover Ph is
Ph Fvh (20) The yaw moment is caused by a counter-torque imbalance.
Torque about z-axis is generally defined as
Where vh is the air velocity
I zz ( m1 m 2 m 3 m 4 ) (33)
(21)
f m V Fvh
Kq From Equation (8) each motor supplies machine torque m
By using the momentum theory which is balanced by the drag torque so the net torque on
F (22) propeller is:
vh
2 A m J d (34)
Then The yaw torque along z-axis given by:
F (23)
f m V F zz I zz m1 m 3 m 2 m 4 (35)
Kq 2 A
The torque is proportional to the trust with constant I zz J ( 1 3 2 4 ) ( d1 d 3 d 2 d 4 ) (36)
ratio K t depends on blade geometry. Since it has been assumed in equation (12) that the drag
torque is proportional to the square of propellers speed,
m Kt F (24)
2 2 2 2
I zz J (1 3 2 4 ) D (1 3 2 4 ) (37)
Then
J D
(1 3 2 4 ) (38)
2 2 2 2
Kt F F (25) (1 3 2 4 )
f V F I zz I zz
Kq 2 A
Then the relationship between thrust and voltage thus
d) Forces along z-axis
established as
2
f K t 2 (26) From equation (2), the net force at the centre of mass of the
F 2 A V
K q vehicle is
1 4
z 1 Fi [cos cos ] g
a) Roll moment m
On substituting equation (26) for F , we have
The roll motion is about x-axis and the rolling moment caused
by the actions of motors 2 and 4 can be defined as: 2
2 A fKt 2
z
m Kq
2 2
2
V1 V2 V3 V4 cos cos g
I yy ( F1 F2 F3 F4 )l (27) (39)
Since F1 and F3 remain unchanged,
e) Forces along x-axis
I yy ( F2 F4 )l (28)
By actuators action, we have the force acting along the x-axis
Substituting equation () for F , we have as
2 2 On substituting and simplifying further, we have
fK t 2 fK t 2 (29) 2
I yy
( 2 A V2 ) (2 A V4 )l 2 a A fK t 2 (40)
K q
K q x
m K q
2 2
2
V1 V2 V3 V4 [sin sin cos cos sin ]
The final equation for roll motion is obtained as
2
f) Forces along y-axis
2 Al fK t 2 2 (30)
(V 2 V4 )
I yy K q
Similar to that of the x-axis, actuators action produces the simulated using the Matlab/Simulink software, figures 4 and
following force, acting along the y-axis 5. This is very essential because it will help to verify the
2 a A fK t
2
(41) correctness of the helicopter dynamic model and to test the
y
m
2 2 2
2
V1 V 2 V3 V4 [sin sin cos cos sin ] control algorithms performance.
K q
With the input voltages, V1, V2, V3 and V4 figure 4, the
3 Simulation of quad-rotor dynamics dynamic behaviour of the quad-rotor due to change in the
excitation voltage is analysed, figure 6.
The set of differential equations describing the dynamics of
quad-rotor that were obtained earlier have been modelled and
8
0
7
-0.2
6
p o s itio n (m )
p o s itio n (m )
-0.4
5
4 -0.6
3
-0.8
-1
1
0 -1.2
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70
time (s) time (s)
Pitch Angle vs Time
Z-Position vs Time 16
2.5
14
2 12
10
a n g le ( d e g )
1.5
p o s itio n (m )
1 6
0.5
2
0
0 0 10 20 30 40 50 60 70
0 10 20 30 40 50 60 70
time (s)
time (s)
9
3
8
2.5 7
a n g le (d e g )
a n g le (d e g )
1.5
4
3
1
2
0.5
1
0 0
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70
time (s) time (s)
Figure 6: The dynamic behaviour of the quad-rotor due to change in the excitation voltage
References
[1] E. Altug, J. P. Ostrowski, R. Mahony. Control of a Quad-
rotor Helicopter Using Visual Feedback, International
Conference on Robotics and Automation, pp 72-77,
(2002).
[2] M. Y. Amir, V. Abbas. Modeling and Neural Control of
Quad rotor Helicopter, Yanbu Journal of Engineering
and Science, volume 2, pp 35-49, (2011).
[3] T. Bresciani. Modelling, Identification and Control of a
Quad rotor Helicopter, MScThesis, Lund University,
(2008).
Figure 7: 3-D view of the Quad-rotor motion [4] M. Claudia, C. T. Luminita, K. K. Simon. Modelling and
Control of Autonomous Quad-rotor, MSc Group Project,
The model as shown in figure 5 has a total of 10 sub-systems, University of Aalborg, Denmark, (2010).
with four of them simulating the rotor dynamics, another [5] R. Czyba. Attitude Stabilization of an Indoor Quad-
three simulating the angular accelerations and the last three rotor, (2009), http://www.emav09.org/EMAV-final-
simulating the linear accelerations. papers/paper_64.pdf.
[6] A. S. Sanca, P. J. Alsina, J. F. Cerqueira. Dynamic
Modelling of a Quad-rotor Aerial Vehicle with Nonlinear
5 Conclusions Inputs, Robotic Symposium, 2008. LARS '08. IEEE Latin
Flying a quad-rotor is a difficult task because of its inherent America, pp 143-148, (2008).
sensitivity to changes in rotor speeds. With its aerodynamics
being quite complex, the need for accurate modelling in order
to achieve precise control cannot be over-emphasised. A good
understanding of the helicopters responses to speed changes
of the four rotors can be very helpful in the development of a
suitable controller. This paper presents a detailed dynamic
analytical model of the quad-rotor helicopter using Matlab/
Simulink. The dynamic behaviour of the quad-rotor was
assessed and its responses to voltage changes were critically
analysed. The model is further calibrated and linearized using
Taylors Series for use on any quad-rotor helicopter. This will
be the subject for future publications.