Quadcopter Control using Different Controllers

Rohit Khandait, Vasant Kumar, Vivek Bhurse, Vivek Tiwari, Swapnil Khubalkar
Center of Excellence in Smart City and Intelligent Transportation, Department of Electrical Engineering,
G H Raisoni College of Engineering
2022 International Conference on Intelligent Controller and Computing for Smart Power (ICICCSP) | 978-1-6654-7258-6/22/$31.00 ©2022 IEEE | DOI: 10.1109/ICICCSP53532.2022.9862416

Nagpur, India
khandait [email protected], [email protected]

Abstract—This paper discusses implementation of the PD, PID, coter, three control technique such as PD control, PID control
Fuzzy-PID controllers on quadcopter. Quadcopter have variety of and Fuzzy-PID control have been developed [16] and tuning of
applications in real. Generally, quadcopter are unstable system. each controller has been done [17] [18]. Complete Simulation
To avoid any form of damage to the quadcopter, a quadcopter
mathematical model should be built first, and then various is implemented on MATLAB/Simulation and Finally, three
control strategies can be employed. MATLAB/Simulink is used control methods are analyzed and compared [19]- [22].
for modelling and control of the quadcopter. The primary goal In the section II, modelling of the quadcopter system is
of the study is to obtain intended result from that of intended explained. In section III, control system architectures ex-
input. Finally, simulation outcomes were compared on the basis of plained. Section IV and section V conclude several results
settling time, overshoots and undershoots to see which controller
is the most effective for the quadcopter. The effectiveness of the and simulation and upcoming Works.
best controller is verified through MATLAB/Simulink. II. M ODELLING OF Q UADCOPTER
Index Terms—Quadcopter, Dynamics, Roll, Pitch, Yaw, PD
Controller, PID Controller, Fuzzy-PID Controller A. Dynamic model
The knowing of the mathematical model of the quadcopter
I. I NTRODUCTION is an important to control and stabilize its motion. The mod-
A quadcopter is a form of unmanned aerial vehicle (UAV) elling consists of the structure of the quadcopter, modelling
with four propellers operated by four BLDC motors. These its kinematics, motors, forces, torques and its equation of
motors regulate the quadcopter’s movement in space by offer- motion. Quadrotor flies with the assistance of four motors. Two
ing varied angular speeds [1]. Because of its numerous applica- opposed motors revolve in a clockwise manner for vertical
tions in today’s environment, quadcopters are commonly used flight, while the other two revolve in an anticlockwise fashion.
by researchers for their research. Quadcopters have numerous The x-axis is stabilised by couple of opposed motors revolving
applications in today’s scenario e.g., Photography, shipping in same manner, and the y-axis is stabilised by another pair of
and delivery, entertainment, weather forecast, search and res- opposite motors. A quadcopter has a body with four propellers
cue operation [2], precision agriculture [3], military operation, and six degrees of freedom (DOF), so 6 variables i.e., x, y,
spy vision, and daily life applications. Quadcopters control is z, ϕ, θ and ψ and they demonstrate spatial orientation of the
very important For their successful implementation. Several quadcopter.
academics have suggested alternative control approaches for • x, y and z: it indicates the position and altitude of

quadcopter control. Fuzzy logic control, PID control, PD quadcopter

control, LQG control, LQR control, and fuzzy-PID control • ϕ: It indicates the x-axis angle i.e., roll angle

are some of the control techniques used for the control of • θ: It indicates the y-axis angle i.e., pitch angle

quadcopter. [4] - [6]. • ψ: It indicates the z-axis angle i.e., yaw angle

PID controller is being used in various industries due to their A quadcopter is supposed to be solid and symmetrical, with
practicality [7] [10]. There are various methods of tuning are the propellers’ gyroscopic impact being minimal. Equations
available and this is the one of reason why they are preferred of motion for quadrotors are used to create its mathematical
in the real time applications [11] - [13]. This controller model which are on the basis of Newton and Euler equations.
also gives the efficient and easy technology to control the Equation (1)-(6) show the equation of motions for quadcopter.
quadcopter. However, It has a drawback that the primary term
1
gains are proportional, integral, derivative gains have to tuned ẍ = (cos ϕ sin θ cos ψ + sin ϕ sin ψ) U1 (1)
manually. For PID controller self-tuning, a fuzzy logic system m
is employed. [14]. 1
ÿ = (cos ϕ sin θ sin ψ + sin ϕ cos ψ) U1 (2)
The main purpose is to introduce the fundamental of quad- m
copter modelling and to develop control strategies for quad- 1
z̈ = −g + (cos ϕ cos θ) U1 (3)
copters roll, pitch, yaw and altitude [15] control, For quad- m
Iyy − Ixx Jr L
978-1-6654-7258-6/22/$31.00 ©2022 IEEE ϕ̈ = θ̇ψ̇( )−( )θ̇Ω + U2 (4)
Ixx Ixx Ixx

Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on June 02,2025 at 19:17:19 UTC from IEEE Xplore. Restrictions apply.
 Fig. 1: Quadcopter Structure
Fig. 2: Quacopter simulink model

Izz − Ixx Jr L
θ̈ = θ̇ϕ̇( )−( )ϕ̇Ω + U3 (5)
Iyy Iyy Iyy According to the equations described in the previous part, this
simulink model is separated into three different subsystems.
Ixx − I − yy 1
ψ̈ = θ̇ϕ̇( )− U4 (6) • Subsystem for propeller speed (Ω)
Izz Izz
• Subsystem for calculating angles (roll, pitch, yaw)
In the above equation, quadcopter’s weight is depicted by • Subsystem for calculating altitude (z)
‘m,’ with acceleration due gravity which is depicted by ‘g’,
whereas The quadcopter’s moments of inertia along the x-axis, III. C ONTROLLING S TRATEGY
y-axis, and z-axis are represented by the letters ‘Ixx ’, ‘Iyy ’,
For the quadcopter model shown in above section, indepen-
and ‘Izz ,’ accordingly, ‘Jr’ indicates the angular momentum
dent controller is created for each input. This paper presents
whereas ‘Ω’ indicates the propeller speed. The input or trans-
three different controller design techniques.
lation vector factors are U1 , U2 , U3 , and U4 , where U1 is the
• PD Controller
force exerted on the body of quadcopter, roll angle is causes
• PID Controller
by the input signal U2 , pitch angle is causes the control input
• Fuzzy-PID Controller
signal U3 , and U4 causes yaw angle. The propeller’s rotation
and speed can be expressed using the equations below. Details of controllers are presented and following that compar-
ison of all three controllers is made to identify which controller
U1 = B(Ω1 2 + Ω2 2 + Ω3 2 + Ω4 2 ) (7) has great results for the developed quadcopter model.

U2 = LB(−Ω2 2 + Ω4 2 ) (8) A. PD Controller

This controller provides combined action of both propor-
U3 = LB(−Ω1 2 + Ω3 2 ) (9) tional and derivative controller. Combining the derivative con-
U4 = D(−Ω1 2 + Ω2 2 − Ω3 2 + Ω4 2 ) (10) troller with the proportional controller as a result can minimize
the overshoots and settling time and response can be much
Ω = (−Ω1 2 + Ω2 2 − Ω3 2 + Ω4 2 ) (11) faster. Figure 3 shows the PD controller’s implementation.

In above equation, L describe the distance between propeller

to centre, B describe the thrust factor and D describe the drag
factor. Ω1 , Ω2 , Ω3 and Ω4 are the velocity of front, back, right
and left motors propeller.

B. Simulink Modelling
Considering U1 , U2 , U3 , and U4 as inputs and roll angle,
pitch angle, yaw angle, and altitude as outputs, a quadcopter
plant model is constructed. Plant model of the quadcopter will Fig. 3: PD Controller
contain multiple subsystems that will indicate various factors
based on their equations. The aim is to develop a control The PD controller’s output is given by
scheme for the quadcopter model after creating a model that de
includes all of the quadcopter’s dynamic equations. u(t) = Kp e + Kd (12)
dt

Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on June 02,2025 at 19:17:19 UTC from IEEE Xplore. Restrictions apply.
Kp denotes the proportional gain, Kd stands for derivative
gain, where ‘e’ symbolised as error signal, which can calcu-
lated as variation among the measured and original value. The
manual tuning method tuning is used to get desired output

B. PID Controller
The term PID refers to the proportional integral derivative
and it is employed in the management of various system. The
PID controller receives an error value representing the change
between measured and desired signal, which is multiplied by
Fig. 5: Fuzzy-PID Controller
the gains to obtain plant control. Proportional compares the
expected and actual values and multiplying the error value
by a constant. The steady state error is removed using the using ”If-Then” logic Rule.
integral term, which involves integrating the error signal with The Fuzzy-PID controller’s output can be written as
time until it reaches 0. By multiplying the rate of change of Z t
error with time, the derivative term makes an expert estimate de
uF P (t) = KpF e + KiF edt + KdF (14)
about the future error. Figure 4 depicts the construction of 0 dt
the PID controller. Where, KpF , KiF and KdF are proportional, integral and
derivative gain Fuzzy-PID controller.

Fig. 4: PID Controller

PID controller’s output can be given by Fig. 6: Fuzzy logic Controller Basic Structure
Z t
de
u(t) = Kp e + Ki edt + Kd (13)
0 dt
In MATLAB/Simulink, a PID controller can be developed
using simple gains, and also there is a in-built PID block
available in the MATLAB/Simulation. The PID controller
regulates the real values retrieved from the plant model based
on the desired input values. For altitude, roll, pitch, and yaw,
we have four different PID controller.

C. Fuzzy-PID Controller
Fuzzy-PID Control methods is the addition of two regula-
Fig. 7: Membership function of error and derivative of error
tion methods that is Fuzzy with PID. It takes PID controller as
core for fuzzy thinking and regulate the PID gains. The Fuzzy-
PID controller generates control rules based on theoretical and
practical study. As a result, by modifying the other controlling
parameters and factors, it may fine tune the gains Kp , Ki ,
and Kd . This increases total control precision, resulting in
improved performance. Figure 5. depicts the implementation
of a Fuzzy-PID controller.
The fuzzy controller rules are used to modify PID gains
in a fuzzy-PID controller that gets error “e” and derivative
of error “de/dt” as inputs. As an inference engine, the
”mamdani-type” inference is used. The position error value
on output in the form of quadcopter motion can be reduced Fig. 8: Kp and Kd membership functions

Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on June 02,2025 at 19:17:19 UTC from IEEE Xplore. Restrictions apply.
 so to transform this fuzzy output in the form of a number
which can reflect fuzzy outputs, a defuzzification method is
required. The control signal should be continuous and should
not cause significant changes in output with respect to input
signal. Furthermore, the defuzzification must have large num-
ber of membership values and correlate to approximate centre
of function. for that purpose, we use “centroid defuzzification”
method for the defuzzification of the output. And it is more
recommended for nonlinear system

Fig. 9: Ki membership function

TABLE I: Fuzzy logic Rules(Kp)

e/de NB NM NS ZO PS PM PB
NB Big Big Big Big Big Big Big
NM Small Big Big Big Big Big Small
NS Small Small Big Big Big Small Small
ZO Small Small Small Big Small Small Small
PS Small Small Big Big Big Small Small
PM Small Big Big B Big Big Small
PB Big Big Big Big Big Big Big
Fig. 10: Block diagram of the quadcopter with controller

TABLE II: Fuzzy logic Rules(Kd) IV. R ESULTS AND SIMULATIONS

e/de NB NM NS ZO PS PM PB Complete mathematical model for the quadcopter and three
NB Small Small Small Small Small Small Small controllers for stabilizing quadcopter simulations were per-
NM Big Big Small Small Small Big Big
NS Big Big Big Small Big Big Big
formed in MATLAB/Simulink software. The main objective
ZO Big Big Big Big Big Big Big of the study is to see which control strategy can control
PS Big Big Big Small Big Big Big quadcopter successfully. The input signal to the system is the
PM Big Big Small Small Small Big Big step signal and accordingly got output for roll, pitch, yaw and
PB Small Small Small Small Small Small Small
altitude. For all three controllers for the quadcopter model and
comparative performance are shown. PD controller is tuned
TABLE III: Fuzzy logic Rules(Ki) with the help of manual tuning method of tuning and for gains
of PID controller we used auto tuning methodology. And in the
e/de NB NM NS ZO PS PM PB
NB 1 1 1 1 1 1 1 scenario of Fuzzy-PID controller, the gains of PID controller
NM 2 2 1 1 1 2 2 are selected using a fuzzy logic method. Among the all three
NS 3 2 2 1 2 2 3 controllers, the Fuzzy-PID produces better results. Figure 11
ZO 4 3 2 2 2 3 4
PS 3 2 2 1 2 2 3
shows the altitude or Z response of the quadcopter.
PM 2 2 1 1 1 2 2
PB 1 1 1 1 1 1 1

The basic structure of fuzzy controller in which two in-

put and three output membership function are present and
mamdani-type inference is used as shown in fig. 6. Figure
7. depicts the membership function for input as error with
derivative of error. It consist of seven composite linguistic
terms which are negative big, negative medium, negative small,
zero, positive small, positive medium and positive big and
represented as NB, NM, NS, ZO, PS, PM and PB. Fig. 8
represents the output’s membership function as for Kp and
Kd which consist of two composite linguistic terms and fig.
9. shows the output Ki Membership function which consist of Fig. 11: Altitude control of quadcopter
four linguistic variables. Total 49 rules are made for the given
quadcopter system. Table no. I, II and III shows rules for Kp , Altitude control is one of the challenging and difficult duty
Ki and Kd gain respectively. within the control of quadcopter. There were many overshoots
The fuzzy system’s output is fuzzy, and that we cannot directly and undershoots in the system when we employed the PD
supply this fuzzy output to the dynamic system for its control, controller for altitude control, as illustrated in the figure 10.

Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on June 02,2025 at 19:17:19 UTC from IEEE Xplore. Restrictions apply.
when we used the PID controller for altitude control some
overshoot and undershoots are present in the system. And
when we used Fuzzy-PID controller for altitude control almost
all part of the oscillation vanishes and the quadcopter system’s
settling time and amplitude have enhanced and are now nearly
equal to the input.

Fig. 14: Yaw angle control of Quadcopter

V. C ONCLUSION
In the paper, a quadcopter nonlinear mathematical model
is provided using MATLAB/Simulink. During modelling of
the quadcopter, all Rotor dynamics and aerodynamics are
considered and Euler-Newton equations were used to model
Fig. 12: Roll angle control of quadcopter the quadcopter. Here, three different control techniques are
implemented i.e. PD controller, PID ontroller, and Fuzzy-
PID controller to stabilize quadcopter. Various simulation in
MATLAB/Simulink were performed to control quadcopter
By looking at the roll angle control of the quadcopter
using different controllers and comparison between all three
in figure 12. It is conclude that Fuzzy-PID controller has
control techniques have been done. By looking at the graphs,
greater control than PD controller and PID controller. There is
it is concluded that the performance of Fuzzy-PID controller
almost zero overshoots and undershoot recorded in Fuzzy-PID
outperformed both the PD and PID controllers. Fuzzy-PID
controller.
controller is removing the overshoot and undershoots from sys-
tem and it has almost zero oscillations and smaller settling time
than the other two controllers. Desired output is achieved using
Fuzzy-PID controller.Hence, it can be concluded that Fuzzy-
PID controller is more perfect controller for quadcopter model
provided in the paper. Implementation of additional control
system methodologies like LQR, robust control technique and
SMC for the comparison of results with the Fuzzy-PID control
can be tried as future scope of the work.

A PPENDIX
Constant parameter which were used during the simulation
of quadcopter are showed in Table IV.

Fig. 13: Pitch angle control of quadcopter TABLE IV: Constant parameters
Symbol Description values
Moment of inertia (x-axis)
Ixx 0.0075
(Nmsˆ2)
By looking at the pitch angle control of the quadcopter Moment of inertia (y-axis)
in figure 13. we can say that using PD controller the issue Iyy 0.0075
(Nmsˆ2)
we were having was with the output’s settling time which Moment of inertia (z-axis)
Izz 0.0130
(Nmsˆ2)
is partially reduced by introducing PID controller and it is Total rotational moment
almost negligible in the case of Fuzzy-PID controller. In Jr 6.50*10ˆ-5
(Nmsˆ2)
addition, There are no oscillations noticed, and the system’s m
Mass of quadcopter
0.65
(Kg)
reaction is the same as the input. The comparative Yaw control
D Drag Factor 7.50*10ˆ-5
of quadcopter using all three controller is mentioned in the B Thrust Factor 3.21*10ˆ-5
figure 14. The Fuzzy-PID controller outperforms the other two L
Distance between propeller and
0.23
controllers in this scenario as well. center of quadcopter(m)

Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on June 02,2025 at 19:17:19 UTC from IEEE Xplore. Restrictions apply.
 ACKNOWLEDGMENT
We would like to thank to Electrical Engineering De-
partment, G H Raisoni College Of Engineering, Nagpur for
supporting us for this work.
R EFERENCES
[1] Praveen Ekka, ”A review paper on unmanned aerial vehicle,” Interna-
tional journal of engineering research and technology (IJERT), vol. 5,
April 2018.
[2] Piotr Kardasz and Jacek Doskocz, Drones and possibilities of their using,
Journal of civil and Environmental engineering, researchGate, January
2016 .
[3] UM Rao Mogili and B B V L Deepak, “Review on application of drone
in precision agriculture,” in Procedia computer science, science direct,
vol. 133, pp. 502–509, 2018.
[4] S. Bouabdallah, A. Noth and R. Siegwart, ”PID vs LQR control tech-
niques applied to an indoor micro quadrotor,” IEEE/RSJ International
conference intelligent robot and system, vol. 3, pp.2451-2456, 2004.
[5] Heba Talla Mohamed Nabil EIKholy, ”Dynamic modelling and control
of a quadrotor using linear and non linear approach,” Cairo, 2014.
[6] Pavel chmelar, ”Building and controling the Quadrocopter,” vol. VI,
2011.
[7] Ferry, Quadcopter plant model and control system development with
MATLAB/Simulink implementation, Rochester institute of technology,
2019.
[8] M Usman, ”Quadcopter Modelling and control with MATLAB/Simulink
implementation,” 2020.
[9] Semih YAGLI, Semih GOREN and Sena TEMEL, ”P, PD, PI, PID
controllers”, 2013.
[10] Antonio Visioli, ”Practical PID control, London, Springer,” 2006.
[11] Abdel-Razzak MERHEB and Hassn NOURA, ”Novel bioinspired
stochastic tuning of a Quadrotor PD controller,” Australian control
conference, Sydney, 2012.
[12] O’Dwyer, Aidan, ”Handbook of PI and PID CONTROLLER tuning
rule,” World Scientific, 2009.
[13] L. Reznik, ”Fuzzy Controllers Handbook,” 1997.
[14] R. Rinkeviciene S. Lisauskas, ”Fuzzy adaptive PID control design,”
Department of Automation Vilnius Gediminas Technical University,
Vilnius, Paper
[15] Fahmizal, Atikah Surriani, Ma’un Budiyanto and Muhammad Arrofiq.
”Altitude control of Quadrotor using Fuzzy Self Tuning PID controller,”
2017 5th International conference on Instrumentation, Control, and
Automation (ICA), 2017.
[16] M. Rosyadi, S. M. Muyeen, R. Takahashi, J. Tamura. ”Transient stability
enhancement of variable speed permanent magnet wind generator using
adaptive PI-Fuzzy controller,” 2011 IEEE Trondheim PowerTech, 2011
[17] Ehsan Abbasi and Mohammad Mahjoob, ”Controlling of Quadrotor
UAV Using a fuzzy System for Tuning the PID gains in Hovering mode,”
ResearchGate, 2013.
[18] Patil, M. D., Vadirajacharya, K., Khubalkar, S. W., ”Design and tuning
of digital fractional-order PID controller for permanent magnet DC
motor,” IETE Journal of Research, 1-11.,2021.
[19] Patil, M. D., Vadirajacharya, K., Khubalkar, ”Design of fractional
order controllers using constrained optimization and reference tracking
method,” International Journal of Power Electronics and Drive Systems,
11(1), 291, 2020.
[20] Kakde, A. A., Tarnekar, S. G., Daigavane, P. M., ”MatLab Simulation
of Reconfigured Ultracapacitors for Co-Working With Battery for Faster
Energy Exchange,”International Conference on Smart Electric Drives
and Power System (ICSEDPS) (pp. 1-6). IEEE, June 2018.
[21] Chaudhari, M., Babu, K., Khubalkar, S. W., Daigavane, P., ”Off-Grid
Hybrid Solar Power Conditioning Unit for Critical and Non-Critical
Loads,” International Conference on Intelligent Computing and Control
Systems (ICCS) (pp. 969-974). IEEE, 2019.
[22] Lanjewar, A., Khubalkar, S. W., Junghare, A. S. (2018, June). Compar-
ative analysis of two loop integer and fractional order PID controller for
inverted pendulum. In 2018 International Conference on Smart Electric
Drives and Power System (ICSEDPS) (pp. 380-383). IEEE.

Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on June 02,2025 at 19:17:19 UTC from IEEE Xplore. Restrictions apply.