CHAPTER TWO

Q-COPTER DYNAMICS AND CONTROL

 CHAPTER TWO: Q-COPTER DYNAMICS AND CONTROL

2.1 Introduction
The Quadcopter is simple in mechanical design compared to helicopters. The Quadrotor, an
aircraft with four engines, features the electronic board in the centre and the engines at its
four extremities. The space motion of the rigid body Q-copter can be divided into two parts:
the movement of the barycenter and the movement around the barycenter. Six degrees of
freedom are required in describing any time-space motion. There are three barycenter
movements and three angular motions, namely, three translations and three rotations along
three axes. The motions include forward and backwards, lateral, vertical, roll, pitch, and yaw.

2.2 Q-copter Coordination System

Before describing the mathematical model of a quadrotor, it is necessary to introduce the

reference coordinates in which the structure and the position are described. For the quadrotor,
it is possible to use two reference systems.

The first is fixed and the second is mobile. The fixed coordinate system, also known as an
inertial system, is a system where Newton’s first law is considered valid. As a fixed
coordinate system, the North-East-Down (NED) system will be used. As can be observed
from Figure 2.1, its vectors are directed to the North, East, and the centre of the Earth.

Figure 2.1: ONED fixed reference system

The mobile reference system is united with the quadrotor's barycenter. It is called the OABC
system, where ABC stands for Aircraft Body Centre. Figure 2.2 illustrates the two coordinate
systems.
 Figure 2.2: Fixed reference and mobile reference systems

2.3 Q-copter Mechanism

The attitude and position of the quadrotor can be controlled as desired by changing the speeds
of the four motors. The gyroscopic effect is only apparent in lightweight construction
quadrotors. The propellers are divided into two groups. In each group, two motors rotate in
diametrically opposite directions. Namely, as shown in Figure 2.3, front and rear propellers
(numbers 2 and 4) rotate counterclockwise; right and left propellers (numbers 1 and 3) rotate
clockwise.

Figure 2.3: Direction of propeller’s rotations

The rotational moments of the quadrotor can be realised by a reactive torque produced by the
rotor. The size of the reactive torque is relative to the rotor speed. The four rotor speeds are
not the same; therefore, the reactive torques will not be balanced, and the quadrotor will start
to rotate. The direction of propeller rotations is conventional for the forces and moments
applied to the quadrotor.

1. Thrust

When the four rotor speeds are the same, the thrust caused by the rotors' rotation, and the
reactive torques will balance each other, and the quadrotor will not rotate, as shown in Figure
2.4.
 Figure 2.4: Thrust

2. Roll

The rolling moment is caused when propeller 1 (left) rotates faster than propeller 3 (right).
Propeller 4 (rear) rotates equally to propeller 2 (front), but slower than propeller 1. The
reactive torques will be created around the front-rear axis (x-axis), causing the quadrotor to
rotate and produce a roll, as shown in Figure 2.5.

Figure 2.5: Roll

3. Pitch

The pitching moment is caused when propeller 4 (rear) rotates faster than propeller 2 (front).
Propeller 1 (left) rotates equally to propeller 3 (right), but slower than propeller 4. The
reactive torques will be created around the left-right axis (y-axis), causing the quadrotor to
rotate and produce a pitch, as shown in Figure 2.6.

Figure 2.6: Pitch

4. Yaw

The yawing moment is caused by the unbalanced rotational speeds of the four rotors. The
yawing moment can be cancelled out when two rotors rotate in the opposite direction. The
yawing moment is caused when propeller 4 (rear) rotates similarly to propeller 2 (front).
Propeller 1 (left) rotates equally to propeller 3 (right), but faster than propellers 4 and 2. The
reactive torques will be generated around the vertical axis (z-axis), causing the quadrotor to
rotate and produce yaw, as shown in Figure 2.7.

Figure 2.7: Yaw

2.4. Q-copter Control

This section presents the fundamental dynamics of the Quadcopter, as well as the control
concept. The basic idea of the quadcopter's movement is illustrated in Figure 2.8. The
rotational speeds of different motors can be adjusted to control six degrees of freedom
motions. Movement in the horizontal frame is achieved by tilting the platform, whereas
vertical movement is achieved by changing the total thrust of the engines.

Quadcopters present specific challenges in control design due to the four inputs and six
outputs of a quadrotor; such a quadrotor is considered an underactuated, nonlinear, and
complex system (an improper system). To control it, certain assumptions are made during the
quadrotor modelling process: the quadrotor is regarded as a rigid body; the structure is
assumed to be symmetric; and the ground effect is neglected.
 Figure 2.8: Q-copter Control Concept

2.2 The System Kinematics

The kinematics of a system refers to the fundamental mechanics that focus on motion. In
quadcopters, position, orientation, linear and angular velocities play critical roles in
influencing the system's flight characteristics.
2.2.1 Position
Position determination requires establishing two frames of reference: the inertial frame
represented as E I and the body frame represented as E B.
The inertial frame provides a fixed point of reference, describing the system’s position and
orientation as

[]
XG
E I= Y G
ZG

And the body frame is directly fixed to the centre of mass represented as

[]
XB
E B= Y B
ZB

2.2.2 Position vector

A vector is used to represent the position of the system as an element of the inertial frame.

[]
x
ξ= y ∈ E I
z
If 𝜉 is the position vector of the system, being an element of the inertial frame, 𝑥, 𝑦, 𝑧
represent the longitudinal, lateral, and vertical position of the system, respectively.
2.2.3 Orientation vector
A vector is used to represent the orientation of the system using Euler angles as an element of
the inertial frame

[]
ɸ
η= θ ∈ E I
ψ
𝜂 – the orientation vector of the system is an element of the inertial frame; 𝜙, 𝜃, 𝜓 represent
the system’s roll, pitch, and yaw angles respectively.

The rotation matrix (𝑅) from the body frame to the inertial frame using Euler angles is
represented in

𝑅 = 𝑅𝑧(𝜓)×𝑅𝑦(𝜃)×𝑅𝑥(𝜙)
If (𝜙), (𝜃), and (𝜓) are rotation matrices around the 𝑥, 𝑦, and 𝑧 axes with angles of 𝜙, 𝜃,
and 𝜓, respectively.

2.2.4 Linear velocity

A vector is used to represent the linear velocity of the system as an element of the body frame

[]
u
υ= v ∈ E B
w
If 𝜐 is the linear velocity of the system, being an element of the body frame; u, 𝑣, 𝑤 are the
linear velocity along the 𝑥, 𝑦, and 𝑧 axes, respectively.

2.2.5 Angular velocity

A vector is used to represent the angular velocity of the system as an element of the body
frame

[]
p
ω= q ∈ EB
r
If 𝜔 is the angular velocity of the system, being an element of the body frame; 𝑝, 𝑞, and 𝑟
are the angular velocities around the 𝑥, 𝑦, and 𝑧 axes, respectively.

2.2.6 Kinematic relationships

The kinematic relationships between the derivative of position and linear velocity, and the
derivative of orientation and angular velocity, are represented in

ξ̇=R . v ∈ E I
η̇=J .ω ∈ E I
where ξ̇ is the derivative of the position vector (𝜉) representing the linear velocity of the
system in the 𝑥, 𝑦, and 𝑧 axes.
η̇ is the derivative of the orientation vector (𝜂) representing the angular velocity of the system
in the 𝜙, 𝜃, and 𝜓 angles.
The system's rotation matrix utilises Euler angles to describe the orientation of the body
frame relative to the inertial frame. It displays the system's conversion from body coordinates
to inertial coordinates. The representation of the system's rotation matrix is

R=
[ cos ψ cos θ sinϕ sin θ cos ψ −cos sin ¿ cos ϕ cos ψ sinθ +sinψ sinϕ
cos θ sin ψ
¿ ϕ ψ

¿
cos ϕ sinθ−sinϕ cos ψ ¿ sin θ sin ϕ ¿ cos θ sin ϕ ¿ cos ϕ cos θ¿
]
The transformation matrix of the system links the orientation vector (for roll and pitch angles)
to the transformation matrix, showing how alterations in the roll and pitch angles impact the
transformation between the orientation and transformation matrices for the system. Thus, the
representation of the system's transformation matrix as

[ ]
1 0 −sinθ
J= 0 cos ϕ cosθ . sinϕ
0 −sinθ cos θ .cos ϕ

2.3 The System Dynamics

2.3.1 Thrust
A rotor's force is proportional to the square of its angular velocity based on aerodynamic
principles and is typically expressed using a mathematical equation as,

2
F thrust =K a . ωi

where F thrust is the total thrust generated by four rotors of the quadcopter; 𝑘𝑎 is the
aerodynamic force constant; ω i is the rotational speed generated by each propeller (the rotator
number i =1, 2, 3 and 4).

The total thrust F thrust generated by all rotors can be expressed as,
F thrust =K a . ( ω 21+ ω22 +ω 23+ ω24 )
If ω 1=ω 2=ω3 =ω 4 for hovering the total thrust is then
2 2 2 2
F thrust =4 K a ω1=4 K a ω 2=4 K a ω 3=4 K a ω 4

2.3.2 Moment
Propellers create moments (torque) that are proportional to the square of their angular
velocity, giving a relationship expressed as,

2
M total=K m . ωi

where M total is the total moment generated by four propellers of the Q-copter and K mis the
moment constant.

The total moment can be expressed as,

2 2 2 2
M total=K m .(ω1 +ω2 +ω 3 +ω 4)

If ω 1=ω 2=ω3 =ω 4 for the total moment then

2 2 2 2
M total=4 K m ω1=4 K m ω 2=4 K m ω3=4 K m ω 4
2.3.3 Torques
The torque generated by the propellers of the Q-copter are expressed for rolling by,

τ ɸ=L . K a [ ( ω 21+ ω23 ) −( ω22 +ω 24 ) ]

If ω 2=ω 4 and ω 1> ω3 or ω 3> ω1 for rolling then

[ 2 2
]
τ ɸ=L . K a ( ω1 +ω3 ) −2 ω 2 =L . K a [ ( ω1 +ω 3 )−2ω 4 ]
2 2 2 2

The torque generated by the propellers of the Q-copter is expressed for pitching by,

τ θ =L. K a [ ( ω22 +ω 24 ) −( ω 21+ ω23 ) ]

,If ω 1=ω3 and ω 2> ω4 or ω 4 >ω 2 for pitching then

[ 2 2 2
]
τ θ =L. K a ( ω2 +ω 4 ) −2 ω3 =L . K a ( ω2 +ω 4 ) −2 ω1 [ 2 2 2
]
The torque generated by the propellers of the Q-copter is expressed for yawing by,

[
τ Ψ =K m ( ω1 +ω3 ) −( ω2 +ω 4 )
2 2 2 2
]
,,If ω 1=ω3 and ω 2=ω 4 and ω 1> ω2 or ω 2> ω1 for yawing then

τ Ψ =2 K m [ ω21−ω22 ]=2 K m [ ω 23−ω 24 ]

where τ ɸ∧¿, τ θ , τ Ψ are the torques generated on roll, pitch, and yaw, respectively

𝐿 is the distance between the centre of the system and the Quadcopter’s motor.

2.3.4 Rotational motion

The rotational dynamics of a rigid body, such as a quadcopter, can be modelled using the
Newton-Euler method. This approach considers the change in angular momentum over time,
which is equal to the sum of the external torques applied to the system. The general rotational
motion equation is expressed as:

τφ
∂(Iη̇ )
=I η̈+ İ η̇=⌊ τθ ⌋
∂t
τψ

where
I is the moment of inertia matrix

[]
φ
η= θ represents the angular position vector
ψ
η̇ is the angular velocity
η̈ is the angular acceleration.
The term İη̇ accounts for the variation in the inertia matrix over time, which is typically
negligible for a rigid body (İ = 0), the equation simplifies to:

τφ
Iη̈ =⌈ τθ ⌉
τψ

This simplified expression is commonly used in quadcopter modelling, where the inertia
matrix is assumed to be constant, and the torques along each axis directly determine the
angular accelerations.

2.3.5 Rotational acceleration

The rotational accelerations affecting roll, pitch, and yaw can be obtained by

τ ɸ I yy I zz
ϕ̈= + θ̇ Ψ̇ − θ̇ Ψ̇
I xx I xx I xx
 τ θ I zz I xx
θ̈= + ɸ̇ Ψ̇ − ɸ̇ Ψ̇
I yy I yy I yy

τ Ψ I xx I yy
Ψ̈ = + ɸ̇ θ̇− ɸ̇ θ̇
I zz I zz I zz

2.3.6 Translational acceleration

The dynamics of the system have a direct relationship with the acceleration of the rotational
and translational motions, meaning that the dynamic equations must be in acceleration form.
Therefore, Newton’s second law of motion, which refers to the acceleration of the system,
can be used to derive the translational motion equation of the system.

Fthrust
ẍ= ×( cosϕ cos ψ sin θ +sin ψ sinϕ )
ms
F thrust
ÿ= ×(cos ϕ sinψ sin θ−cos ψ sin ϕ )
ms
Fthrust
z̈= ×(cosϕ cos θ )
ms
where ms is the mass of the system

2.4 Dynamic Model

The drone will be simulated using the preceding dynamic model based on linear derivative
equations in rotational and translational forms. The numerical solutions of those equations
will be obtained iteratively over flight simulation time using the simple Euler method. The
Euler method is preferred over the Runge-Kutta method due to the complexity of the
formulas required for the intended simulation. However, the accuracy and validity of the
simulation favour the Runge-Kutta method when it is used. Nonetheless, the objective is to
run a preliminary simulation to understand the Q-copter flight behaviour, and the Euler
method is adequate for small time steps.

2.4.1 Linear model

For the sake of simplicity, a linear dynamic model would be helpful to have a cheap flight
simulation to investigate drone performance. By neglecting any forms of the coupling
rotations, i.e., θ̇ Ψ̇ =0 , ɸ̇ Ψ̇ =0 ,∧ɸ̇ θ̇ = 0, the linear model of the rotational accelerations
affecting roll, pitch, and yaw can be expressed by

τɸ
ϕ̈=
I xx

τθ
θ̈=
I yy

τΨ
Ψ̈ =
I zz

By assuming small roll and yaw compared with pitch, i.e., and ψ∧ϕ ≪ θ , the linear model of
the translational accelerations affecting forward, lateral, and normal can be expressed by

Fthrust
ẍ= sinθ
ms

F thrust
ÿ= sinϕ
ms

Fthrust
z̈= cosθ −g
ms

Note that the normal acceleration is taken account for gravity acceleration ( g=9.81 m/ sec 2).

2.4.2 Euler method

Euler's Method is one of the simplest numerical techniques for solving first-order ordinary
differential equations. This approach approximates solutions using finite difference
principles, where state variables are updated step-by-step based on their rate of change
(derivative) and a small-time step (Δt ). The general form of the Euler solution can be given
for roll ∅ as below

d ∅˙
ϕ̈=
dt

d∅
p=
dt

pi = p 0 +
( )
τɸ
I xx
Δt
 ϕ i=ϕ 0+ p Δt

where p is the roll rate. i represents the number of iteration steps required to reach the
steady, converged responses. Quantities with subscript zero are the initial values of
simulations. They may be set to zero and then replaced by the new values at each iteration.
Similarly, the pitch θ and yaw ψ can be obtained by applying the above steps based on the
pitch rate q (pitch derivative) and yaw rate r (yaw derivative), respectively. These will be
applied in Chapter Three to obtain simulations of the quadcopter's roll, pitch, and yaw flight.

2.5 Time Response and Performance

2.5.1 Time domain criteria

The simulation (response) can be scoped using time response to assess the convergence as a
measure of stability and divergence as a measure of instability. Among many time-domain
criteria considered are settling time and peak or overshoot. These paramount objectives are
widely sought to be met to investigate system performance. Settling time is defined as the
time required for the system to settle within a specific percentage of the input amplitude. The
settling time is the Time for which the response remains within 2% of the final value.
4
T s=
ζ ωn
And the percent overshoot is given by
PO=100 e−ζπ /√ 1−ζ
2

where ω n is the natural frequency and ζ is the damping ratio.

2.5.2 System performance

A quantitative measure of a system's performance is necessary for modern control systems

and for designing optimal systems. The performance index is a quantitative measure of a
system's performance, chosen to emphasise critical system specifications. A system is
considered an optimum control system when the system parameters are adjusted so that the
index reaches a minimum value. The general form of the performance integral is

Ts

I =∫ f (e ( t ) .r ( t ) . y ( t ) . t) dt
0
 where f is a function of the error e ( t ) , input r ( t ), output y ( t ) , and time performance indices,
which are helpful for the design of control systems. Of those performance indices, the
approaches are as follows.

1. The integral of the square of the error, ISE, which is defined as

Ts

ISE=∫ e 2 ( t ) dt
0

The squared error is mathematically convenient for analytical and computational purposes.

2. The integral of time multiplied by the squared error, ITSE, which is written as

Ts

ITSE=∫ t e 2 (t)dt
0

3. The integral of the absolute magnitude of the error, IAE, which is written as

Ts

IAE=∫ |e ( t )|dt
0

This index is handy for computer simulation studies.

4. The integral of time multiplied by the absolute magnitude of the error, ITAE, which is
written as

Ts

ITAE=∫ t |e ( t )|dt
0

It is proposed to reduce the initial error and minimise errors that occur later in the response.

2.5.3 Optimum control design

A control system is optimum when the selected performance index is minimised. The general
closed-loop transfer function of control system is

Y ( s) b0
= n
R ( s) s + bn−1 s +…+ b1 s +b0
n−1

where b’s are control parameters that have to found based on optimization.

The coefficients that reduce the ITAE performance criterion for a step response have been
determined for the transfer function shown in Figure 2.9. Normalised time is time with
respect to the settling time. The normalised response is the amplitude with respect to the
steady-state value. Table 2.1 shows the ITAE optimum coefficients for a step input. The
characteristic equation of the closed-loop system transfer function will be matched to those
shown in Table 2.1 to find the optimal controller gains.

Figure 2.9 ITAE step responses

Table 2.1 The ITAE optimum coefficients for a step input

n Optimal characteristic equations

1 s+ω n
2 2
2 s +1.4 ω n s+ ωn
3 2 2 3
3 s +1.75 ω n s +2.15 ω n s+ω n
4 3 2 2 3 4
4 s + 2.1ω n s + 3.4 ω n s +2.7 ω n s+ω n

2.5.4 PID control

A PID (Proportional-Integral-Derivative) control system is a foundational closed-loop

feedback mechanism widely used in automation and engineering to maintain precise control
of physical systems. Its core purpose is to dynamically adjust a system's output—such as
speed, or position—to match a desired target value (setpoint) by continuously calculating the
error (the difference between the setpoint and actual measured value). Its versatility stems
from its ability to handle diverse, dynamic systems without requiring complex mathematical
models. However, effective operation relies on careful tuning of its parameters to avoid
oscillations, overshoot, or sluggishness.

The PID controller then applies three distinct corrective actions based on this error by
achieving a balance of responsiveness, accuracy, and stability.

 The Proportional (P) component responds immediately to the current error magnitude.

P=Kₚ e ( t )

where Kₚ is the proportional gain and e of t is the error at time. A higher proportional gain
leads to a faster response but may cause overshoot or steady-state error.

 The Integral (I) component sums up past errors over time, eliminating steady-state error
(offset).

I =K i∫ e ( t ) dt

where K i is integral gain and ∫ e ( t ) dt integral of error over time. An excessively high
integral gain can lead to overshoot and instability.

 The Derivative (D) component anticipates future behaviour by reacting to the error’s
rate of change, helping to dampen oscillations and improve stability.

d e (t)
D=K d
dt

d (t )
where K d derivative gain and rate of change of error. It can amplify noise in the system
dt
if the derivative gain is too high.

2.5.5 Control system design

The combined PID control output signal is calculated as

u ( t )=( P+ I + D ) e(t )

This output adjusts the system to minimize error, reduce overshoot, and achieve stable and
precise control. Thus, the PID transfer function can be given by
 U (s) Ki
G c ( s )= =K p + + K d s
E( s) s

The selection of the three coefficients of PID controllers is basically a search problem in a
three-dimensional space, a PID controller can often found by a trial-and-error basis.

A closed-loop negative unity feedback system is used to control a dynamic system, as shown
in Figure 2.10.

Figure 2.10 Closed-loop control system

where G p ( s ) and G ( s ) are prefilter transfer function and system or plant transfer function,
respectively.

The overall closed-loop transfer function of the system of Figure 2.10 is

Y ( s) Gc ( s ) G ( s )
=G p ( s )
R ( s) 1+ Gc ( s ) G ( s )

The PID design procedure consists of the following steps to tune controller gains optimally:

1. Select ζ of the closed-loop system by specifying PO.

2. Compute ω n of the closed-loop system by specifying Ts and using ζ .

3. Use the appropriate ITAE optimum equation to match with the characteristic equation
(1+Gc ( s ) G ( s ) ) to obtain Gc ( s ) by determining K p , K i and K d .

4. Determine G p ( s ) so that the closed-loop system transfer function does not have any
zeros by selecting the system numerator to be the denominator of G p ( s ).

2.6 Summary

This chapter discussed the dynamics and control of quadcopters. Their motion, defined by six
degrees of freedom, is achieved by adjusting rotor speeds to generate thrust, roll, pitch, and
yaw. Kinematic and dynamic models were developed using Newton–Euler equations, with
simplified simulations applied through the Euler method. System performance was evaluated
using time-domain criteria and performance indices, while PID control was introduced as an
effective method for stabilization. For a large proportion of control systems, the primary
objective is obtaining a high steady-state accuracy. Another goal is maintaining the transient
performance of these systems within reasonable limits. The relative stability and the transient
response performance of a closed-loop control system are directly related to the location of
the closed-loop roots of the characteristic equation. The movement of the characteristic roots
on the s-plane as key system parameters (such as controller gains) are varied to analyse the
initial design of a system and determine suitable alterations to the system structure and
parameter values. A computer is then commonly used to obtain the accurate controller
implementation for use in the final design and analysis.