drones

Article
Design of the System for Measuring UAV Parameters
Jozef Novotňák 1, * , Martin Fil’ko 1 , Pavol Lipovský 1 and Miroslav Šmelko 2

1 Department of Aviation Technical Studies, Faculty of Aeronautics, Technical University of Košice,

Rampová, 7, 040 01 Košice, Slovakia
2 EDIS vvd., Rampová 7, 040 01 Košice, Slovakia
* Correspondence: [email protected]

Abstract: This article deals with the design and creation of a tensometric measuring system to
measure the parameters of an unmanned aerial vehicle (UAV) of the quadcopter type. The system
was designed to measure the total UAV thrust and the thrust of its individual motors. The distribution
of forces from the UAV motors and their transmission to the sensors was ensured by a specially
designed construction, for which the mechanical stresses were simulated and analysed for different
modes of the UAV flight. The thrust measurement was performed by four pairs of strain gauges. A
measurement system designed in this way and the measured parameters of the UAV can be used for
tuning the flight control algorithms applied in the autopilot.

Keywords: measurement system; unmanned aerial vehicle (UAV); tensometric measurement system;
motor thrust measurement; quadcopter; strain gauge sensors

1. Introduction
Unmanned aerial vehicles (UAVs), especially quadcopter drones, are currently widely
Citation: Novotňák, J.; Fil’ko, M.; utilized in many sectors. The reason for the high usability of UAVs is mainly their economy
Lipovský, P.; Šmelko, M. Design of and that they provide relatively simple equipment to deploy different types of sensors and
the System for Measuring UAV cameras. The UAVs are widely used nowadays. It is possible to use them for wildlife moni-
Parameters. Drones 2022, 6, 213. toring [1–3], rescue operations [4,5], property inspection [6], monitoring and mapping [7–9],
https://doi.org/10.3390/ magnetic field mapping [10,11], monitoring in public spaces [12,13] or, at present, the
drones6080213
import of medicines [14,15], disinfecting of public spaces or even monitoring people’s
Academic Editor: Abdessattar temperatures [16]. However, in such uses, it is necessary to consider the increased weight
Abdelkefi of these UAVs when they are equipped with various devices [17–21], or they may be sup-
plemented by different systems in the future [22]. It is also necessary to carefully consider
Received: 22 July 2022
the use of drones at airports [23], where their precise management to ensure safety is the
Accepted: 17 August 2022
first priority. One possible solution that has been proposed for use at airports is the use of
Published: 18 August 2022
tethered UAVs with external power via a tether [24], which would allow more power to be
Publisher’s Note: MDPI stays neutral supplied to the drone’s motors than is available using an on-board battery. Drones are also
with regard to jurisdictional claims in often operated in non-standard conditions, for example at extremely low temperatures,
published maps and institutional affil- which can change the traction properties of the drone and its controllability. Even in this
iations. case, it is necessary to ensure thorough and precise control of these UAVs.
Platforms of various autopilots can be placed onboard UAVs. These autopilot plat-
forms provide partially or fully autonomous flight depending on the selected mode [25].
The basic algorithm of the autopilot consists of the proportional-integral-derivative (PID)
Copyright: © 2022 by the authors.
controllers, which are used for flight control and stabilization [26,27]. The values of the
Licensee MDPI, Basel, Switzerland.
individual controllers can be set with the help of software to improve the controllability of
This article is an open access article
distributed under the terms and
the UAV. This is important, for example, in solving the problem of precise landing of UAVs
conditions of the Creative Commons
on various moving platforms, such as ship decks and the like [28]. In such manoeuvres,
Attribution (CC BY) license (https:// it is often necessary to consider factors that show a high degree of variability in time,
creativecommons.org/licenses/by/ for example weather conditions or ground effect and time-varying payload during UAV
4.0/). landing, when adaptive altitude control of the quadcopter is needed [29]. Tuning of the

Drones 2022, 6, 213. https://doi.org/10.3390/drones6080213 https://www.mdpi.com/journal/drones

Drones 2022, 6, 213 2 of 17

controller constants can be solved in several ways [30–34]. Many operators set the values
of individual controller constants based on their experience during a test flight of the
UAV or based on quick calculations [35]. However, such adjustments carry the risks of an
unexpected UAV incident and are also time-consuming, as a test flight must be performed
after each adjustment. Where it is impossible to perform a test flight, it is possible to set
corrections based on mathematical modelling of the physical properties of the UAV. There
are many different models of UAVs available today [36–43]. Tuning of the controllers based
on the existing quadrotor models could lead to poor quadrotor performance due to various
inaccuracies in the underlying model. Therefore, it would be necessary to create an accurate
model. This procedure is significantly time-consuming and requires the involvement of
highly qualified experts in the process of the mathematical modelling. Due to the difficulty
of first creating it and the subsequent complexity of the mathematical model of the UAV,
in which it is necessary to model each of its parts, it is desirable to design and apply a
simplified mathematical model. However, such a simplified mathematical model, which
is often not experimentally confirmed, or for which a test flight has not been carried out,
can significantly differ from the actual flight characteristics of the modelled UAV. These
deviations are difficult to detect and quantify, as an experienced pilot compensates for
them subconsciously.
Despite the mentioned problems, it is possible to use a simplified mathematical model
of UAV dynamics [44–46], but it must be able to directly measure the UAV parameters,
especially the complete thrust characteristics of the motors, which are very important in
setting the control algorithms of the UAV, as already mentioned. Currently, there are several
methods that can measure UAV motor parameters, from measuring propeller parameters
to measuring motor thrust, but such measurements require removing the motor from the
UAV [47,48]. However, this could be again time-consuming. The main motivation for this
study, and the main contribution of it, was the design of a measurement system that can
measure UAV parameters, especially the thrusts of individual motors, without the need to
disassemble the motors and the individual structural parts of the UAV. Furthermore, the
aim was that this system could be used in future to create a mathematical model of the UAV,
which would allow the simulation of the position angles of the UAV and the position of
the UAV in space. Further motivation was that by creating a UAV model, this contribution
will reduce the time required to tune PID controllers, especially by limiting the number
of test flights that are currently required after each retuning. By designing a system for
measuring UAV parameters, its possible future use will indirectly increase safety and limit
UAV incidents, which can occur, for example, even in the phase of test flights after retuning
the controllers.

2. Materials and Methods

The design of the measurement system consisted of mechanical and electrical parts.
When designing the electrical parts, it was necessary to select sensors that would be suitable
for measuring the force acting from the UAV motors. Strain gauges were deemed to be a
suitable element, due to their wide use in industrial applications, where they are often used
for universal and accurate measurement of deformation, force, pressure, and moments
acting on solids. The most widely used type are metal resistance strain gauges. Sensing
takes place by changing the ohmic resistance of the sensing element when changing the
mechanical stress acting on the sensor changes. Strain gauges are commonly used in
the detection and analysis of mechanical stresses on structures and individual structural
elements. A brief description of the resistance strain gauge is shown in Figure 1.
Drones 2022, 6, x FOR PEER REVIEW 3 of 17
Drones 2022, 6, 213 3 of 17

Figure1.1. Description
Figure Descriptionof
ofresistance
resistancestrain
straingauge.
gauge.

In
Inour
ourcase,
case,an
an“S”
“S”type
typestrain
straingauge,
gauge,the
theMIK-LCS
MIK-LCS11was wasused.
used. As
As these
these strain
strain gauges
gauges
are not very widespread, their parameters are summarized in Table 1. There were
are not very widespread, their parameters are summarized in Table 1. There were a total a total of
eight strain gauges used in the system. The thrust was measured by two sensors
of eight strain gauges used in the system. The thrust was measured by two sensors for for each
motor. The system
each motor. therefore
The system contained
therefore four pairs
contained of sensors,
four pairs one pair
of sensors, onefor each
pair for motor.
each motor.

Table1.1.Parameters
Table Parametersof
ofthe
theMIK-LCS
MIK-LCS11strain
straingauge.
gauge.
Parameter
Parameter Value
Value
Measuring
Measuring range
range 0–50–5kgkg
Sensitivity
Sensitivity 2.0 ±
2.0 ± 0.05 mV/V
0.05 mV/V
Nonlinearity ≤±0.03% F.S.
Nonlinearity ≤±0.03% F.S.
Hysteresis ≤±0.03% F.S.
Hysteresis
Repeatability ≤±0.03%
≤±0.03% F.S.
F.S.
Repeatability
Zero output ≤±0.03% ±1F.S.
◦
Zero Zero
temperature
outputcoefficient ≤±0.03% ±1 F.S./10 C
Temperature sensitivity coefficient ≤± 0.03% F.S./10 ◦ C
Zero temperature coefficient ≤±0.03% F.S./10 °C
Input resistance 350 ± 20 Ω
Temperature sensitivity
Output resistancecoefficient ≤±0.03% 350F.S./10
± 5 Ω °C
Input resistance
Safe overload 350 ±
≤150%Ω
20 F.S.
Insulance
Output resistance 350 ±5Ω
≥5000 MΩ
Recommended excitation voltage
Safe overload 10 V~15
≤150% F.S.V
Working temperature −20~80 ◦ C
Insulance ≥5000 MΩ
Recommended excitation voltage 10 V~15 V
The mechanical part of the system consisted of a rod construction,
Working temperature –20~80 °Cbase, and mounting
frame. The dimensions of the base were 52 × 52 cm to allow the placement of the rod
construction together with
The mechanical part the sensors
of the systemandconsisted
the mounting frame
of a rod through which
construction, base,a and
quadcopter
mount-
type UAV could be attached to the measurement system. The height of the base could
ing frame. The dimensions of the base were 52 × 52 cm to allow the placement of the rod
be adjusted to the requirements. The height of the measuring system from the base to
construction together with the sensors and the mounting frame through which a quad-
the mounting frame was 65 cm. The dimension of the mounting frame was 74 cm on the
copter type UAV could be attached to the measurement system. The height of the base
diagonal. A sensor was placed on each rod in the lower part of the rod construction. The
could be adjusted to the requirements. The height of the measuring system from the base
sensors were used to sense the forces that are transmitted by the rod construction from
to the mounting frame was 65 cm. The dimension of the mounting frame was 74 cm on
the mounting frame to the base. The rod construction was firmly connected to the base
the diagonal. A sensor was placed on each rod in the lower part of the rod construction.
behind each sensor in each of its corners. In the upper part, the rod construction was firmly
The sensors were used to sense the forces that are transmitted by the rod construction
connected to the mounting frame at its end points. The length of each rod including the
from the mounting frame to the base. The rod construction was firmly connected to the
sensor was 83 cm. To design the mechanical part, the PTC Creo Parametric CAD program
base behind each sensor in each of its corners. In the upper part, the rod construction was
was used to model the individual parts of the measurement system. The essential idea of
firmly connected to the mounting frame at its end points. The length of each rod including
the whole solution was the transmission of forces from UAV motors to strain gauges. The
the sensor
design was
of the 83 cm. To design
measurement theismechanical
system part, the
shown in Figure 2. PTC
When Creo Parametric
designing CAD pro-
the structure,
an emphasis was put on functionality of the individual elements and that they essential
gram was used to model the individual parts of the measurement system. The did not
idea of
place the whole demands
unnecessary solution was theproduction
on the transmission of forces
process and thefrom UAV motors
availability to strain
of materials.
gauges.
To The design
maintain of the measurement
a low weight system
of the measuring is shown
system, in Figurewas
aluminium 2. When
useddesigning the
as the main
Drones 2022, 6, x FOR PEER REVIEW 4 of 17
Drones 2022, 6, x FOR PEER REVIEW 4 of

structure, an emphasis wasan

structure, put on functionality
emphasis was put on offunctionality
the individual elements
of the and elements
individual that theyand that th
Drones 2022, 6, 213 4 of 17
did not place unnecessary demands on the production process and the availability
did not place unnecessary demands on the production process and the ofavailability
ma- of m
terials. To maintain a low
terials. To weight
maintain ofathe
lowmeasuring
weight of the system, aluminium
measuring was
system, used as the
aluminium was used as t
main construction material.
main The low
construction weightThe
material. of the
lowmeasuring system
weight of the structure
measuring wasstructure
system im- was im
construction
portant material.
especially
portant The low weight
for theespecially
sensors to
forbethe of the measuring
minimally
sensors toloaded system
by structure
the weight
be minimally loaded wasweight
of the
by important
structure.
of the structu
especially
Another for the
reason was sensors
the low
Another tomomentum
reasonbewas
minimally andloaded
the low influence
momentum by the
on weight
the
and of the
on structure.
measurement
influence Another
dynamics.
the measurement dynamics.
reason was the low momentum and influence on the measurement dynamics.

Figure 2. Design of measuring system.

Figure
Figure 2. 2. Design
Design of of measuring
measuring system.
system.
Load simulations were performed to give an overview of the behaviour of the mea
Load simulations
urement were
system performed
when ittowasto give an overview of the ofbehaviour of the mea- a simp
Load simulations were performed giveloaded with the
an overview offorce 1 N. Forofthe
the behaviour thesimulation,
meas-
surement system when
rod model it was loaded
of loaded
the system withwas the force of 1 N. For the simulation,
sufficient forasimple
a rod
urement system when it was with thedesigned
force of that
1 N.was
Forquite
the simulation, the evaluation
simple of t
model of the system
forces was
actingdesigned
on the that was
structure. quite sufficient for the evaluation of the forces
rod model of the system was designed that was quite sufficient for the evaluation of the
acting on the structure.First, simulation of the mechanical stress of the measuring system was perform
forces acting on the structure.
First, simulation
with the of quadcopter
the mechanical in thestress
thrustofmode
the measuring
(see Figuresystem was performed
3), in which with
First, simulation of the mechanical stress of the measuring system wasallperformed
four quadcopter moto
the quadcopter havein thethe
thrust mode (see Figure 3), in which all four quadcopter motors
same thrust. The thrust of the individual motors is shown in the figure as forc have
with the quadcopter in the thrust mode (see Figure 3), in which all four quadcopter motors
the same thrust.F1The for thrust
motor of 1, the individual
F2 for motor 2,motors is shown
F3 for motor in the
3, and F4 figure as forces
for motor 4. FromF1 the
for simulati
have the same thrust.
motor 1, F2 forresults, The
motor 2, thrust
F3 for of the
motor individual motors is shown in the figure as forces
it was clear that 3, andequal
with F4 for motor
thrust on4.allFrom
four the simulation
motors results, the m
of the quadcopter,
F1itfor
wasmotor
clear1,that
F2 with
for motor 2,thrust F3 foron motor 3, and F4 for
of motor 4. From the thesimulation
chanicalequalstructure was all four
evenly motors
stressed and the
the quadcopter,
individual rods mechanical
were loaded (stretche
results, it was
structure wasclear thatstressed
evenly
equally. with equal thrust
and the on all four
individual rodsmotors of the quadcopter,
were loaded the me-
(stretched) equally.
chanical structure was evenly stressed and the individual rods were loaded (stretched)
equally.

Figure
Figure 3. UAV thrust 3. UAV thrust
simulation simulation
(3D view on the(3D
leftview
side,on the view
front left side, front
on the viewside).
right on the right side).

In the second step, simulation of the mechanical stress of the measuring system with
Figure 3. UAV thrust simulation (3D view on the left side, front view on the right side).
the quadcopter in pitch or roll mode was performed; this meant two motors having a
higher speed (F1 and F4) and two motors having a lower speed (F2 and F3). In this mode
one side of the structure will be stressed by tension and the opposite side will be stressed
by mechanical pressure (see Figure 4).
 InInthe
thesecond
secondstep,
step,simulation
simulationofofthe
themechanical
mechanicalstress
stressofofthe
themeasuring
measuringsystem
systemwiw
the quadcopter in pitch or roll mode was performed; this meant
the quadcopter in pitch or roll mode was performed; this meant two motors having two motors havin
higherspeed
higher speed(F1
(F1and
andF4)
F4)and
andtwo
twomotors
motorshaving
havinga alower
lowerspeed
speed(F2 (F2and
andF3).
F3).InInthis
thismod
mo
Drones 2022, 6, 213 one side of the structure will be stressed by tension and the opposite 5 of
side 17 be stress
will
one side of the structure will be stressed by tension and the opposite side will be stresse
bymechanical
by mechanicalpressure
pressure(see
(seeFigure
Figure4).
4).

Figure 4. UAV pitch and4.roll

Figure UAVsimulation
pitchand (3D
androll view on the(3D
rollsimulation
simulation left view
side, on
front view
theleft on the
leftside,
side, right
front side).
view
Figure 4. UAV pitch (3D view on the front view ononthe
theright
rightside).
side).

Finally, simulation of the

Finally, mechanical
simulation stress of the measurement system withsystem
the with
Finally, simulation ofofthe
the mechanical
mechanical stressofofthe
stress themeasurement
measurement system with th
quadcopter in yaw mode was
quadcopter in performed.
yaw mode was Inperformed.
this case, theIn twocase,
this opposite
the motors
two had motors
opposite a ha
quadcopter in yaw mode was performed. In this case, the two opposite motors had
higher speed (F1higher
and F3) and (F1
speed the and
otherF3)
twoandlower speedtwo
theother
other (F2lower
and F4),
speedwhich depends
(F2and
andF4), on the depends
F4),which
which
higher speed (F1 and F3) and the two lower speed (F2 depends o
UAV configuration.
the The
UAV simulation
configuration.results
The showed
simulation the stress
results of individual
showed the rods,
stress of while the rods,whi
individual
the UAV configuration. The simulation results showed the stress of individual rods, wh
rods marked by red
the are
rods stressed
marked byby tension
red are and
stressed the
by rods marked
tension and by
the blue
rods are stressed
marked by
the rods marked by red are stressed by tension and the rods marked by blue are stresse by
blue are stress
mechanical pressure
by (see Figure
bymechanical
mechanical 5). (see
pressure
pressure (seeFigure
Figure5).
5).

Figure 5. UAV yaw simulation

Figure (3D view on the (3D
left side, top view on thetop
right side).
Figure 5.5.UAV
UAVyaw
yawsimulation
simulation viewon
(3D view onthe
the leftside,
left side, view
top view ononthe
theright
rightside).
side).

The combination Theof combination

electrical and mechanical andparts createdpartsa system capable of mea-
The combination ofofelectrical
electricaland mechanical
mechanical created
parts created a asystem
system capableofofmea
capable me
suring the UAVuring
parameters.
the UAV However,
parameters. as the output
However, assignal
the from
output the sensors
signal from had
the a low had a l
sensors
uring the UAV parameters. However, as the output signal from the sensors had a lo
voltage level (ofvoltage
the orderlevelof several
(of theordermV),ofitseveral
order was necessary
mV),ititwas to amplify
was necessarythis signal tothis thesignal to
voltage level (of the of several mV), necessary totoamplify
amplify this signal to th
order of volts, which was solved
orderofofvolts,
volts,which by
whichwas a circuit
wassolved for
solvedby analog signal
bya acircuit
circuitfor processing.
foranalog
analogsignal The main
signalprocessing. part
processing.The of main p
The
order main pa
this circuit wereofthe operational
this wereamplifiers
circuitwere AD620,
theoperational
operational which were
amplifiers AD620, used for signal
which wereused amplifica-
used forsignal
signalamplif
ampl
of this circuit the amplifiers AD620, which were for
tion. The electronics
cation. were
The adapted
electronics in such
were a way
adapted that
in it
such allowed
a way different
that it levels
allowed of signal
different levels of sig
cation. The electronics were adapted in such a way that it allowed different levels of sign
amplification toamplification
be selected using to be jumpers,
selected and thus
using it could
jumpers, and be
thususedit withbe
could different
used types
with different ty
amplification to be selected using jumpers, and thus it could be used with different typ
of sensors if necessary.
of sensorsBased
if on the designed
necessary. Based circuit
on the solution, circuit
designed it was solution,
possible to it choose
was the to cho
possible
of sensors if necessary. Based on the designed circuit solution, it was possible to choo
signal amplification
the
levelamplification
thesignal
signal of 10, 100 or 1000,
amplification leveland
level
it 100
ofof10,
10, was
100ornecessary
or1000,
1000,and
andtoitbe
itwaschosen
was appropriately,
necessary
necessary totobebechosen
chosenappr
app
so that at the maximum
priately, measured
so that at thevalue,
maximumwe did not
measuredexceed the
value, maximum
we did
priately, so that at the maximum measured value, we did not exceed the maximum inpunot input
exceed voltage
the maximum inp
level of the analog to
voltage digital
level converter
of the analog(ADC).
to After
digital amplification,
converter (ADC). the signal
voltage level of the analog to digital converter (ADC). After amplification, the signal ww
After was
amplification, further the signal
processed by a microcontroller, which provides measurement with sampling frequency
of 100 Hz and digital signal processing (DSP), which was ensured by its own software
solution for calculating the thrusts of individual UAV motors and the total thrust of the
UAV. The electrical part of the measurement system can be represented by a simplified
block diagram shown in Figure 6.
 further processed by a microcontroller, which provides measurement with sampling fre-
further
quency processed
of 100 Hz by
anda digital
microcontroller, which provides
signal processing measurement
(DSP), which with by
was ensured sampling
its ownfre-
soft-
quency of 100 Hz and digital signal processing (DSP), which was ensured by its own soft-
ware solution for calculating the thrusts of individual UAV motors and the total thrust of
ware solution for calculating the thrusts of individual UAV motors and the total thrust of
Drones 2022, 6, 213 the UAV. The electrical part of the measurement system can be represented by a simpli-6 of 17
the UAV. The electrical part of the measurement system can be represented by a simpli-
fied block diagram shown in Figure 6.
fied block diagram shown in Figure 6.

Figure 6. Block diagram of the electrical part of the measuring system.

diagram of
Figure 6. Block diagram of the
the electrical
electricalpart
partof
ofthe
themeasuring
measuringsystem.
system.

For the
For the purposes
purposes of of measurement,
measurement, it was was first
first necessary
necessary to to identify
identify the the measurement
measurement
For the purposes of measurement, ititwas first necessary to identify the measurement
system, to determine
system, determine its its resolution,
resolution, transfer
transfer function,
function, and and the
the sensitivity
sensitivity of of the individual
the individual
system, to to determine its resolution, transfer function, and the sensitivity of the individual
sensors.
sensors. Next, it was necessary to introduce the required
sensors. Next, it was necessary to introduce the required software gain corrections toto
Next, it was necessary to introduce the required software
software gain
gain corrections
corrections to
make
make
make thethe recalculation
the recalculation
recalculation of of the
of the thrust
the thrust during
thrust during
duringthe the measurement
themeasurement
measurementas as accurate
asaccurate as possible.
accurateasaspossible.
possible.An An
An
ATMega2560microcontroller
ATMega2560
ATMega2560 microcontrollerwas was used
wasusedused tototo
determine
determine
determine thethe
the parameters
parameters
parameters ofofthe
the measuring
ofmeasuring
the measuring sys-
sys-
tem. It
system.
tem. It was
was an 8-bit
It was
an 8-bit
an AVR-based
8-bit AVR-based
AVR-based microcontroller
microcontroller
microcontroller with
with aa10-bit
with 10-bit
a 10-bitADC.
ADC. ADC.
The
The identification
identification ofof
thethe response
response of of
thethe measuring
measuring
The identification of the response of the measuring system was system
system was was realized
realized
realized using using
using pre-
precisely
pre-
cisely determined
determined weights. weights.
The The
system system
was was
testedtested
by by pushing
pushing and
cisely determined weights. The system was tested by pushing and pulling during the ini- and
pullingpulling during
during the the ini-
initial
tial experiments,
experiments, and and
the the results
results showed
showed the the
samesame values.
values. In
tial experiments, and the results showed the same values. In the beginning, the transferIn the
the beginning,
beginning, the
the transfer
transfer
function of
function
function of the
of the strain
the strain gauge
strain gaugeMIK-LCS
gauge MIK-LCS111was
MIK-LCS wasdetermined.
was determined.During
determined. Duringthe
During themeasurements,
the measurements,
measurements, the
the
the
system was
system
system was gradually
was graduallyloaded
gradually loadedwith
loaded withdifferent
with differentweights.
different weights.The
weights. Thechange
The changeinin
change inthe
thesensor
the sensor
sensor output
output
output was
was
was
expressed
expressed by
expressed the
by the
by output
the output value
outputvalue
valueof of the
ofthe ADC,
theADC,
ADC,which which
whichisis given
giveninin
isgiven bits.
inbits. The
bits.The transfer
Thetransfer function
transferfunction
function ofof
of
the
the MIK-LCS
the MIK-LCS111strain
MIK-LCS straingauge
strain gaugeisisisshown
gauge shown
shown ininin
Figure
Figure
Figure 7,7,where
7,where
whereinin
the
inthegraph,
the graph,
graph, each point
each
each represents
point
point repre-
repre-
asents
different
sents weight.
aa different
different weight.
weight.

Figure 7.
7. Transfer
Figure 7.
Figure Transfer function
Transfer functionof
function ofthe
of theMIK-LCS
the MIK-LCS111strain
MIK-LCS straingauge.
strain gauge.
gauge.

The measurement
measurementshows
The measurement showsthat
shows thatthe
that thetransmission
the transmissioncharacteristic
transmission characteristicofof
characteristic ofthe
theused
the usedsensor
used sensor
sensorwaswas
was
linear, which means that
linear, which means that the the transfer
the transfer function
transfer function
function ofof the
of the measuring
the measuring system
measuring system (with
system (with eight
(with eight strain
eight strain
strain
gauges)
gauges)willwill be
willbe also
bealso linear.
alsolinear. The
linear.The same
Thesame procedure
sameprocedure
procedure was
was
was used
used totodetermine
used todetermine
determine the
the transfer
transfer
the func-
function
transfer func-
tion
of theof
tion of the
entire
the entire
entire measuring
measuring
measuring system,
system, but inbut
system, thisinin
but this
case
this case
the the applied
applied
case the weights
applied weights were
were evenly
weights were evenly dis-
distributed
evenly dis-
Drones 2022, 6, x FOR PEER REVIEW 7 of 17
tributed
among among
eight eight
strain strain
gauges gauges
(see (see
Figure Figure
8). The 8). The
transfer transfer
function function
of the
tributed among eight strain gauges (see Figure 8). The transfer function of the measuring of the
measuringmeasuring
system
system
was alsowas
system also
linear
was inlinear
also in
inthis
this case.
linear thiscase.
case.

Figure 8.
Figure 8. Transfer
Transfer function
function of
of the
the measuring
measuring system.
system.

The resolution of the measuring system could be determined from the transfer func-
tions measured this way. The resolution of the sensor (Rs) can be determined as:
∑ 𝑤
𝑅𝑠 (1)
Drones 2022, 6, 213 7 of 17

The resolution of the measuring system could be determined from the transfer func-
tions measured this way. The resolution of the sensor (Rs) can be determined as:

∑7n=1 w
Rs = (1)
∑7n=1 out

where w is the weight used and out is the output at a certain weight. The resolution of the
sensor per least significant bit (LSB) determined using the Equation (1) was 3.70961 g/LSB,
which is 0.03637 newtons/LSB. The next step was to determine the time constant of the
system. This constant was determined from the normalized graph of load with a certain
weight at an output value of 0.63 on the y-axis. Such a method of determining the time
constant is widely used [49]. The value of the time constant determined in this way was
0.300456 s. Based on the previous results, the transfer function Gx(s) of the measuring
system for individual sensors was determined as:

3.709
Gx (s) = (2)
0.300456s + 1
In addition to these parameters, it was also necessary to determine the gain correction
constants for individual sensors. This issue and the calculation of the accuracy (which
depends on the individual sensors, and is at least 99%) of the model was described in
previous work [50]. Based on these previous results, the resulting transfer function of the
sensor was determined as:

3.709
Gx (s) = · GCX (3)
0.300456s + 1
where GCX is gain correction for the individual sensors (x is the sensor number). The
individual parameters of the measuring system and gain corrections are summarized in
Table 2.

Table 2. Parameters of the measuring system.

Parameter Value Unit

System resolution (sensor) 3.70961 grams/LSB 1
0.03637 newton/LSB 1
Transfer function time constant 0.300456 seconds
Gain correction for sensor 1 0.9919 -
Gain correction for sensor 2 1.0007 -
Gain correction for sensor 3 1.0014 -
Gain correction for sensor 4 1.0112 -
Gain correction for sensor 5 0.9901 -
Gain correction for sensor 6 1.0114 -
Gain correction for sensor 7 1.0031 -
Gain correction for sensor 8 0.9901 -
1 Least significant bit (LSB).

The mathematical model of the measuring system is shown in Figure 9.

For the measurement purposes, a customized quadcopter type UAV with a 3D printed
frame was used. Parameters of the quadcopter are listed in Table 3.
The quadcopter together with the measurement system can be seen in Figure 10.
 Gain correction for sensor 8 0.9901 -
1 Least significant bit (LSB).
Drones 2022, 6, 213 8 of 17
The mathematical model of the measuring system is shown in Figure 9.

Figure 9.
Figure 9. Mathematical
Mathematical model of the measuring
measuring system.
system.

Table For the measurement

3. Parameters purposes,
of the customized a customized quadcopter type UAV with a 3D
quadcopter.
printed frame was used. Parameters of the quadcopter are listed in Table 3.
Parameter Value
Table 3. Parameters
Frameofdimensions
the customized quadcopter. 45.3 cm diagonal
Motors RAY G3 C2830-1300 BLDC
Parameter
Regulators Value
RAY G2 30A
Frame dimensions
Flight battery 45.3 cm diagonal
4500 mAh 3S LiPo
Propellers type
Motors RAY G3 three-bladed
C2830-1300 9BLDC× 5 inch
Autopilot
Regulators control unit Pixhawk
RAY G2 30A 4 mini
Quadcopter weight 1.4 kg
Drones 2022, 6, x FOR PEER REVIEW Flight battery thrust
Maximum 4500 mAh353Snewtons
LiPo 9 of 17
Propellers type
Flight time three-bladed 915× min
5 inch
Autopilot control unit Pixhawk 4 mini
Quadcopter weight 1.4 kg
Maximum thrust 35 newtons
Flight time 15 min

The quadcopter together with the measurement system can be seen in Figure 10.

(a) (b)
Figure 10. Measuring
Figure set (a)
10. Measuring setMeasurement system
(a) Measurement withwith
system fixed quadcopter;
fixed (b) Custom
quadcopter; quadcopter.
(b) Custom quadcopter.

3. Measurements
The initial measurements were aimed at measuring the thrust of the UAV and at
monitoring the output raw data from the measurement system. Figure 11 shows that vi-
brations from the UAV motors were introduced into the measurements. Due to this, it was
Drones 2022, 6, 213 9 of 17

3. Measurements
The initial measurements were aimed at measuring the thrust of the UAV and at
Drones 2022, 6, x FOR PEER REVIEW 10 of 17
monitoring the output raw data from the measurement system. Figure 11 shows that
vibrations from the UAV motors were introduced into the measurements. Due to this,
it was necessary to filter the measured data to reduce the variance of the data during
of the system may
measurement andoccur and thus,
to increase theanaccuracy
undesirable reduction
of the of the system’s
measurements dynamics. As of
and determination
the time constant
thrusts. of the transfer
Finite Impulse Response function of theare
(FIR) filters system
among Ts the
= 0.300456
simplests digital
was determined
filters, both
based on the previous
theoretically results,The
and practically. it was necessary of
disadvantage to using
choosethese
the window size forthe
filters is mainly thehigh
MWM order
filter appropriately. The resulting delay, which depends on the window
of the transfer function, greater computational demands when calculating the coefficients size, should be as
small as possible so that it is not affected by the dynamics of the measurement
and determined variables, and last but not least, the long delay in processing the input system.
With the Due
sample. usedto window
the highsize n = the
delay, 11 samples of filters
use of FIR the MWM in thefilter and with
proposed a samplingsystem
measurement fre-
quency of 100 Hz, the filter delay was 0.05 s. The results of the variance
was not suitable. It was, therefore, necessary to test other filter variants. For this reason,and standard
deviation of the data before
various combinations and after
of filtering the implementation
algorithms were designedof andthetested
filtering algorithms
in previous work areand
shown
severalinmeasurements
Table 4. The proposed
were performedfilters [51].
reduceThe the varianceofofthe
efficiency thedifferent
measured data
filters wasat tested
the
output
with a by
UAV approximately
thrust of 240 g70%.
(2.35 newtons). The best results in the output data were obtained
using a double exponential moving average (DEMA) filter and triple exponential moving
Table 4. Results of the variance and standard deviation of the data at the output before and after
average (TEMA) filter which was based on the exponential moving average (EMA) filter.
filtration (DEMA and TEMA filters in combination with MWM filter) with UAV thrust of 240 g.
Exponential averaging filters are floating averaging filters, often used for data smoothing
and predictionFilter
[52,53]. Exponential filters are first order IIR
Variance filters. Deviation
Standard The EMA filter can be
(grams)
described by the following
No filter (RAW data) equation: 59.6183343 7.721291
DEMA 𝜂 = 0.2 + MWM n = 11 21.070095 4.590217
EMA = y[n] = η ·x[n] + (1 − η )·y[n − 1] (4)
TEMA 𝜂 = 0.2 + MWM n = 11 18.620764 4.315178
where y[n] is actual output value, y[n − 1] is previous output value, x[n] is actual input
valueFigure
and η11isshows
a timethe results of
constant at the
the maximum thrust
filter (number of the
value UAV before
between 0 andand
1). after data
The DEMA
filtering in the left part of Figure 11
filter can be described by the equation: and a detailed view in the right part of Figure 11 that
exemplifies how the vibrations from the UAV engines are suppressed and how the data
variance is reduced after filtering.
DEMA = 2· EMA − EMA( EMA) (5)

(a) (b)
Figure
Figure11.
11.Total
Totalmeasured
measuredthrust
thrust of of UAV and after
UAV (a) Data before and after filtration;
filtration; (b)
(b)Detailed
Detailedview
viewon
onthe
the same data before and after filtration.
same data before and after filtration.

4. Results
And finally, the TEMA filter can be described by the equation:
After the complete identification and calibration of the measuring system with the
implemented filtering = 3· EMA in
TEMAalgorithms − 3the
· EMA ( EMA) + EMA
microcontroller, ( EMA
it was ( EMAto
possible )) measure the(6)
thrust of the individual UAV motors. The thrust measurement was done by using a cus-
In our case, both filters had a time constant (η) of 0.2 and were used in combination
tomized quadcopter. The quadcopter was attached to the mounting frame with locking
with a Moving Window Median (MWM) filter with a window size of n = 11. Changing the
brackets made using 3D printing, which ensured that the quadcopter did not move during
parameters of the designed filters, especially the time constant for exponential filters, may
the measurement. The autopilot of the quadcopter had to be deactivated during the meas-
result in insufficient suppression of unwanted vibrations, if the time constant is increased
urement so that no unwanted compensations would occur, because if the control part reg-
above the value of 0.2. If the time constant is reduced below the value of 0.2, overdamping
istered that the quadcopter did not perform the required manoeuvre, it would have tried
to compensate. The thrust adjustment of the motors was realized using pulse width mod-
ulation (PWM) from the flight controller using to the electronic speed controller (ESC),
where 1000 μs pulse time was the minimum value and 2000 μs pulse time was the maxi-
mum value of the thrust. Reference thrusts of individual motors are shown in Figure 12.
Drones 2022, 6, 213 10 of 17

of the system may occur and thus, an undesirable reduction of the system’s dynamics. As
the time constant of the transfer function of the system Ts = 0.300456 s was determined
based on the previous results, it was necessary to choose the window size for the MWM
filter appropriately. The resulting delay, which depends on the window size, should be as
small as possible so that it is not affected by the dynamics of the measurement system. With
the used window size n = 11 samples of the MWM filter and with a sampling frequency
of 100 Hz, the filter delay was 0.05 s. The results of the variance and standard deviation
of the data before and after the implementation of the filtering algorithms are shown in
Table 4. The proposed filters reduce the variance of the measured data at the output by
approximately 70%.

Table 4. Results of the variance and standard deviation of the data at the output before and after
filtration (DEMA and TEMA filters in combination with MWM filter) with UAV thrust of 240 g.

Filter Variance Standard Deviation (grams)

No filter (RAW data) 59.6183343 7.721291
DEMA η = 0.2 + MWM n = 11 21.070095 4.590217
TEMA η = 0.2 + MWM n = 11 18.620764 4.315178

Figure 11 shows the results at the maximum thrust of the UAV before and after data
filtering in the left part of Figure 11 and a detailed view in the right part of Figure 11 that
exemplifies how the vibrations from the UAV engines are suppressed and how the data
variance is reduced after filtering.

4. Results
After the complete identification and calibration of the measuring system with the
implemented filtering algorithms in the microcontroller, it was possible to measure the
thrust of the individual UAV motors. The thrust measurement was done by using a
customized quadcopter. The quadcopter was attached to the mounting frame with locking
brackets made using 3D printing, which ensured that the quadcopter did not move during
the measurement. The autopilot of the quadcopter had to be deactivated during the
measurement so that no unwanted compensations would occur, because if the control
part registered that the quadcopter did not perform the required manoeuvre, it would
have tried to compensate. The thrust adjustment of the motors was realized using pulse
width modulation (PWM) from the flight controller using to the electronic speed controller
(ESC), where 1000 µs pulse time was the minimum value and 2000 µs pulse time was
the maximum value of the thrust. Reference thrusts of individual motors are shown in
Figure 12. Based on these reference thrusts, it will be possible to verify the measurement
results on the measuring system.
Before starting the measurement, it was necessary to set the required amplification
of the circuit for processing the signal from the sensors. The gain was set to the level of
100, as the maximum reference thrust of the quadcopter is 35 N, and at the gain level of
100 we will not exceed the voltage level of the ADC used, which was 5 V. The weight of
the quadcopter was measured on the measuring system before the measurement using
our own software, and then the zero points were set to sensors. The measurement began
only after this initial setup. The thrust was gradually adjusted from 0 to 100% on the flight
controller. The measurement results are shown in Figure 13, where it is possible to see the
measured thrust of each of the motors (four quadcopter motors).
Drones 2022, 6, x FOR PEER REVIEW 11 of 17

Drones 2022, 6, 213 11 of 17

Based on these reference thrusts, it will be possible to verify the measurement results on
the measuring system.

(a) (b)

(a) (b)
(c) (d)
Figure 12. Reference thrust of individual UAV motors (a) Reference thrust of motor 1; (b) Reference
thrust of motor 2; (c) Reference thrust of motor 3; (d) Reference thrust of motor 4.

Before starting the measurement, it was necessary to set the required amplification
of the circuit for processing the signal from the sensors. The gain was set to the level of
100, as the maximum reference thrust of the quadcopter is 35 N, and at the gain level of
100 we will not exceed the voltage level of the ADC used, which was 5 V. The weight of
the quadcopter was measured on the measuring system before the measurement using
(c)our own software, and then the zero points were set to sensors. (d) The measurement began
only after this initial setup. The thrust was gradually adjusted from 0 to 100% on the flight
Figure 12. Reference thrust of individual UAV motors (a) Reference thrust of motor 1; (b) Reference
controller.
Figure The measurement
12. Reference results are
thrust of individual UAVshown in(a)
motors Figure 13, where
Reference thrustitofismotor
possible
1; (b)toReference
see the
thrust of motor 2; (c) Reference thrust of motor 3; (d) Reference thrust of motor 4.
measured
thrust thrust
of motor ofReference
2; (c) each of the motors
thrust (four3;quadcopter
of motor (d) Referencemotors).
thrust of motor 4.

Before starting the measurement, it was necessary to set the required amplification
of the circuit for processing the signal from the sensors. The gain was set to the level of
100, as the maximum reference thrust of the quadcopter is 35 N, and at the gain level of
100 we will not exceed the voltage level of the ADC used, which was 5 V. The weight of
the quadcopter was measured on the measuring system before the measurement using
our own software, and then the zero points were set to sensors. The measurement began
only after this initial setup. The thrust was gradually adjusted from 0 to 100% on the flight
controller. The measurement results are shown in Figure 13, where it is possible to see the
Drones 2022, 6, x FOR PEER REVIEWmeasured thrust of each of the motors (four quadcopter motors). 12 of 17

(a) (b)

(a) (b)
(c) (d)
Figure 13. Measured filtered and unfiltered thrust of individual UAV motors in thrust mode (a)
Figure 13. Measured filtered and unfiltered thrust of individual UAV motors in thrust mode (a) Thrust
Thrust of motor 1; (b) Thrust of motor 2; (c) Thrust of motor 3; (d) Thrust of motor 4.
of motor 1; (b) Thrust of motor 2; (c) Thrust of motor 3; (d) Thrust of motor 4.

Figure 13
Figure 13 shows
shows how
howmuch
muchmore
moreaccurate
accurate thethe
data
dataareare
after filtering,
after compared
filtering, to
compared
those before filtering, against the reference thrust curves of the individual motors
to those before filtering, against the reference thrust curves of the individual motors at at max-
imum thrust
maximum values
thrust fromfrom
values Figure 12, and
Figure whatwhat
12, and a significant role data
a significant filtering
role data playsplays
filtering in thein
system.
the If we
system. compare
If we comparethethe
maximum
maximum measured
measured thrust values
thrust valuesofof
thethemotors
motorsagainst
againstthe
the
maximum reference values, we can see that the measurement using the designed
maximum reference values, we can see that the measurement using the designed measuring measur-
ing system was sufficiently accurate and the differences between the maximum thrust
values did not exceed 1% (Table 5).

Table 5. Accuracy results of the measured maximum thrust value of the motors against the maxi-
mum value of the reference thrust curves.
 Figure 13. Measured filtered and unfiltered thrust of individual UAV motors in thrust mode (a)
Thrust of motor 1; (b) Thrust of motor 2; (c) Thrust of motor 3; (d) Thrust of motor 4.

Figure 13 shows how much more accurate the data are after filtering, compared to
Drones 2022, 6, 213 those before filtering, against the reference thrust curves of the individual motors at 12
max-
of 17
imum thrust values from Figure 12, and what a significant role data filtering plays in the
system. If we compare the maximum measured thrust values of the motors against the
maximum
system wasreference values,
sufficiently we can
accurate seethe
and that the measurement
differences betweenusing the designed
the maximum measur-
thrust values
ing system was sufficiently
did not exceed 1% (Table 5). accurate and the differences between the maximum thrust
values did not exceed 1% (Table 5).
Table 5. Accuracy results of the measured maximum thrust value of the motors against the maximum
Table
value 5.
ofAccuracy results
the reference of the
thrust measured maximum thrust value of the motors against the maxi-
curves.
mum value of the reference thrust curves.
Reference Max. Measured Max.
Motor MotorReference Max. Thrust Thrust(N)
(N) MeasuredThrust Thrust (N) Difference
Max. (N) Difference (%)
(%)
1 8.57 8.53 0.57
1 8.57 8.53 0.57
2 2 9.00 9.00 8.97
8.97 0.34
0.34
3 3 8.81 8.81 8.77
8.77 0.45
0.45
4 4 8.55 8.55 8.48
8.48 0.82
0.82

The
Thelocal
localfluctuation
fluctuationofofthrusts
thrustsseen
seenininFigure
Figure1313was
wascaused
causedby byaasmall
smallamount
amountof of
freedom
freedomininthe
theclamping
clampingmechanism,
mechanism,and andthus
thusthe
thequadcopter
quadcopterreached
reachedsuch
suchaathrust
thrustatat
the
thetime
timewhen
whenthe thefluctuation
fluctuationin inthe
thegraph
graphoccurred
occurredthatthatthe
theclamping
clampingmechanism
mechanismwas was
strained
straineddue
duetotoovercoming
overcomingthe theweight
weightwithwiththrust.
thrust.
The
Thestandard
standarddeviation
deviationresults forfor
results thethe
data measured
data measured in Figure 13 are
in Figure 13shown in Fig-
are shown in
ure 14, where
Figure the red
14, where the lines represent
red lines the mean,
represent the green
the mean, area area
the green in the
ingraphs represents
the graphs the
represents
segment standard
the segment deviation
standard for thefor
deviation unfiltered RAW data,
the unfiltered RAWand theand
data, bluethe
area in the
blue areagraphs
in the
represents the segment
graphs represents standardstandard
the segment deviation for the filtered
deviation for the data.
filtered data.

Drones 2022, 6, x FOR PEER REVIEW 13 of 17

(a) (b)

(c) (d)
Figure 14. Means (red lines) and standard deviation through segments in RAW data (green area)
Figure 14. Means (red lines) and standard deviation through segments in RAW data (green area) and
and filtered data (blue area) (a) Motor 1; (b) Motor 2; (c) Motor 3; (d) Motor 4.
filtered data (blue area) (a) Motor 1; (b) Motor 2; (c) Motor 3; (d) Motor 4.

The
Thedata
datameasured
measured this
this way at the
way at the output
output of
of the
thesystem
systemmay
maybebeapplied
appliedasasananinput
inputto
toa cybernetic
a cybernetic UAV quadcopter model in future research. The proposed system
UAV quadcopter model in future research. The proposed system allows the allows
the measurement of almost any quadcopter type UAV, the size of which does not exceed
the length of the arms of the mounting frame and whose total force does not exceed the
sum of the ranges of the used individual sensors.

5. Discussion
Drones 2022, 6, 213 13 of 17

measurement of almost any quadcopter type UAV, the size of which does not exceed the
length of the arms of the mounting frame and whose total force does not exceed the sum of
the ranges of the used individual sensors.

5. Discussion
Measuring the thrust characteristics of UAV motors is currently most often performed
by measuring the thrust of the UAV motor placed on a special measuring platform. How-
ever, with such an approach, it is necessary to dismantle it from the UAV and fix it on
a special measuring platform. This paper describes the original design of a system for
measuring the thrust characteristics of UAV motors (quadcopter type) without the need to
remove the motors from the UAV frame. The system allows the attachment of the whole
quadcopter to the measuring system, which is an indisputable advantage. However, a
disadvantage may be the limitation regarding the maximum possible dimensions of the
quadcopter, which cannot exceed the size of the mounting frame of the measuring system,
which is based on the design of the measuring system. Another limitation may be the
maximum load that the system is capable of measuring. This load depends on the type of
sensors used and their maximum possible load. The sensors used on the proposed system
have a maximum possible load of 5 kg per sensor. As the system contained eight sensors,
it could be loaded with a total weight of 40 kg, which is suitable for most widely used
quadcopters. The proposed solution can be modified in the future at the sensor level, where
different types of strain gauges can be used depending on the range of load that need to be
measured, as the proposed electronics of the system allow changing the gain, which can be
adapted to different types of sensors.
The measurements of the thrust of the motors had high accuracy, as the system was
adjusted and verified using different types of weights. However, there was a problem
during the initial measurements with vibrations from the quadcopter, which caused a high
dispersion of data at the output of the system. This was solved by implementing filtering
algorithms, which suppressed the data dispersion by approximately 70%. In future, it may
be possible to focus on the design and implementation of other advanced algorithms that
would further suppress the dispersion and thus increase the accuracy of the measurement,
but at the same time they cannot limit the dynamics of the measurement.
The measurement system currently has a resolution of 3.70961 g/LSB per sensor. This
resolution was determined by the resolution of the ADC used, which was 10 bits, which
was sufficient for current use in the case of concept design. If it is necessary to obtain
higher resolution, the measurement system can be improved by using 16-bit ADC and
the resolution of the measurement system would be 0.05796 g/LSB per sensor. Further
research will focus on optimization of the design of the system to obtain more accurate
measurement results.
Another advantage of the designed measuring system is the possibility of using it for
the diagnostics of the state of the UAV and its motors. During one of the measurements
while testing the filtering algorithms, it could be seen in the graph that the RAW data on
motor 4 showed more vibrations than the other motors (see Figure 15). The reason was a
slightly loose propeller on motor 4. If we increased the sampling frequency, this system has
the possibility to expand its functionality with diagnostics of the motors in the future.
The results show that the constructed measurement system can be used to identify UAV
parameters, especially the complete thrust characteristics of the motors. The parameters of
the UAV measured in this way can be used in future as input data for a dynamic model of
the UAV, so there is no need to model the complete motor part of the UAV from propellers
to motors and controllers. Creating a model for future research and subsequent modelling
of UAV characteristics eliminates the need for frequent test flights when tuning the UAV
controllers. With this method, setting up the on-board control electronics will be faster
and more accurate if the UAV model is known. The concept for the future development
of the complex measurement system needed for the tuning of the PID controllers can be
illustrated by a simplified block diagram (Figure 16).
 Another advantage of the designed measuring system is the possibility of using it for
the diagnostics of the state of the UAV and its motors. During one of the measurements
while testing the filtering algorithms, it could be seen in the graph that the RAW data on
motor 4 showed more vibrations than the other motors (see Figure 15). The reason was a
Drones 2022, 6, 213 slightly loose propeller on motor 4. If we increased the sampling frequency, this system14 of 17
has the possibility to expand its functionality with diagnostics of the motors in the future.

(a) (b)

(a) (b)
(c) (d)
Figure 15. The possibility of using the measuring system in motors diagnostics (a) Thrust of motor
1; (b) Thrust of motor 2; (c) Thrust of motor 3; (d) Thrust of motor 4.

The results show that the constructed measurement system can be used to identify
UAV parameters, especially the complete thrust characteristics of the motors. The param-
eters of the UAV measured in this way can be used in future as input data for a dynamic
model of the UAV, so there is no need to model the complete motor part of the UAV from
propellers to motors and controllers. Creating a model for future research and subsequent
modelling of UAV characteristics eliminates the need for frequent test flights when tuning
(c)the UAV controllers. With this method, setting up the on-board (d) control electronics will be
faster and more accurate if the UAV model is known. The concept for the future develop-
Figure 15. The possibility of using the measuring system in motors diagnostics (a) Thrust of motor
Figure
ment of Thecomplex
15.the possibility of using the measuring
measurement systemfor
system needed in motors diagnostics
the tuning of the (a)
PIDThrust of motor 1;
controllers
1; (b) Thrust of motor 2; (c) Thrust of motor 3; (d) Thrust of motor 4.
(b) Thrust
can of motor by
be illustrated 2; (c) Thrust of motor
a simplified block3;diagram
(d) Thrust of motor
(Figure 16).4.
The results show that the constructed measurement system can be used to identify
UAV parameters, especially the complete thrust characteristics of the motors. The param-
eters of the UAV measured in this way can be used in future as input data for a dynamic
model of the UAV, so there is no need to model the complete motor part of the UAV from
Figure 16. Use of measurement system in process of creating of the UAV mathematical model.
propellers
Figure to motors
16. Use and controllers.
of measurement system inCreating a model
process of creating for
offuture
the UAV research and subsequent
mathematical model.
modelling of UAV characteristics eliminates the need for frequent test flights when tuning
6. Conclusions
6.
theConclusions
UAV controllers. With this method, setting up the on-board control electronics will be
The
fasterThe objective
andobjective of this
of
more accurate thisif research
research
the UAV was
was to design
to
model design aa measuring
is known.measuring
The concept system
system that
forthat could
the could identify
futureidentify
develop- the
the
UAV UAV parameters,
ment parameters,
of the complex i.e., determine
i.e.,measurement
determine thesystem the total thrust
total thrust
needed andand
forthe the thrust
thethrust
tuning characteristics
characteristics ofofthe
theindi-
of the PID controllers
vidual
can be UAV motors.
illustrated by The measuring
a simplified system
block described
diagram (Figurein 16).
this research was designed for this
purpose. The initial design and construction of the measuring system was optimized sev-
eral times. It was necessary to design the concept and to create a 3D model of the measuring
system. Based on this 3D model, a simple rod model was created, in which the distribution
of forces in the structure was simulated, based on possible UAV flight modes. Subsequently,
the selection
Figure 16. Useof
of sensors
measurementfor the measurement
system in process system and
of creating of the
the construction
UAV mathematicalof themodel.
measuring
system began. Electronics for processing signals from sensors were designed, as well as
software for the microcontroller, which ensured the calculation of the motors thrusts at the
6. Conclusions
output of the system.
The objective After
of this the system
research was was designed,
to design a mathematical
a measuring model
system that of the
could system
identify
was created. Subsequently, filter algorithms were designed for the
the UAV parameters, i.e., determine the total thrust and the thrust characteristics of themeasurement system,
which helped reduce the dispersion of data at the output of the system by almost 70%,
which contributed to higher measurement accuracy. Thrust measurements of individual
quadcopter motors were performed on the measuring system. These measurements were
compared with the reference thrust curves of the individual motors used on the quadcopter.
The measurements showed that the differences between the measured and the reference
maximum thrusts of the motors were less than 1%, which means that the accuracy of the
measurement system was more than 99%. A measurement system with such accuracy can
be used in future UAV parameter determination research and studies.
The system forms the basis for the design of a mathematical model of the UAV
(quadcopter type). The thrusts of the individual motors measured by the system can be
used in further research focused on the dynamic cybernetic model of the UAV as inputs of
Drones 2022, 6, 213 15 of 17

this model. The UAV model will be used for conversion of the measured thrust to position
angles, calculation of position coordinates in space and to tune the PID controllers of the
autopilot, based on static laboratory measurements. The main benefits of this modelling
method are the reduction of the time required for tuning PID controllers, reduction of the
number of UAV flights needed to verify the set constants of the PID controllers and the fact
that it is possible to use the proposed methodology for the diagnostics, for example for the
UAV testing during the manufacturing process.

7. Patents
For the solution mentioned in the article, a patent application for a measuring system
with application number 50020-2021 was filed at the Office of Industrial Property of the
Slovak Republic.

Author Contributions: Conceptualization, P.L., M.Š. and M.F.; methodology, J.N., M.Š. and P.L.;
software, J.N. and P.L.; validation, J.N. and M.Š.; formal analysis, P.L.; investigation, J.N., M.F., M.Š.
and P.L.; data curation, J.N., M.F. and P.L.; writing—original draft preparation, J.N.; writing—review
and editing, P.L.; visualization, J.N., M.F. and P.L.; supervision, P.L.; project administration, P.L. All
authors have read and agreed to the published version of the manuscript.
Funding: This research was funded by the Cultural and Educational Grant Agency MŠVVaŠ SR,
grant number 045TUKE-4/2022, Research Agency, ITMS code number 313011AUP1 and ITMS code
number 313011T557.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Mesquita, G.; Mulero-Pázmány, M.; Wich, S.; Rodriguez-Teijeiro, J.D. A practical approach with drones, smartphone and tracking
tags for potential real-time tracking animal. Curr. Zool. 2022, 68, zoac029. [CrossRef]
2. Saunders, D.; Nguyen, H.; Cowen, S.; Magrath, M.; Marsh, K.; Bell, S.; Bobruk, J. Radio-tracking wildlife with drones: A viewshed
analysis quantifying survey coverage across diverse landscapes. Wildl. Res. 2022, 49, 1–10. [CrossRef]
3. Tansuriyavong, S.; Koja, H.; Kyan, M.; Anezaki, T. The Development of Wildlife Tracking System Using Mobile Phone Commu-
nication Network and Drone. In Proceedings of the 2018 International Conference on Intelligent Informatics and Biomedical
Sciences (ICIIBMS), Bangkok, Thailand, 21–24 October 2018; pp. 351–354.
4. Yavuz, D.; Akbıyık, H.; Bostancı, E. Intelligent drone navigation for search and rescue operations. In Proceedings of the 24th
Signal Processing and Communication Application Conference (SIU), Zonguldak, Turkey, 16–19 May 2016; pp. 565–568.
5. Iob, P.; Frau, L.; Danieli, P.; Olivieri, L.; Bettanini, C. Avalanche Rescue with Autonomous Drones. In Proceedings of the 2020
IEEE 7th International Workshop on Metrology for AeroSpace, MetroAeroSpace, Pisa, Italy, 22–24 June 2020; pp. 319–324.
6. Falorca, J.; Miraldes, J.; Lanzinha, J. New trends in visual inspection of buildings and structures: Study for the use of drones.
Open Eng. 2021, 11, 734–743. [CrossRef]
7. Yaprak, S.; Ömer, Y.; Susam, T. UAV Based Agricultural Planning and Landslide Monitoring. TeMA-J. Land Use Mobil. Environ.
2017, 11, 325–338.
8. Maini, P.; Tokekar, P.; Sujit, P.B. Visual Monitoring of Points of Interest on a 2.5D Terrain Using a UAV with Limited Field-of-View
Constraint. IEEE Trans. Aerosp. Electron. Syst. 2021, 57, 3661–3672. [CrossRef]
9. d’Oleire-Oltmanns, S.; Marzolff, I.; Peter, K.; Ries, J. Unmanned Aerial Vehicle (UAV) for Monitoring Soil Erosion in Morocco.
Remote Sens. 2012, 4, 3390–3416. [CrossRef]
10. Lipovský, P.; Draganová, K.; Novotňák, J.; Szőke, Z.; Fil’ko, M. Indoor Mapping of Magnetic Fields Using UAV Equipped with
Fluxgate Magnetometer. Sensors 2021, 21, 4191. [CrossRef]
11. Lipovský, P.; Fil’ko, M.; Novotňák, J.; Szöke, Z.; Košuda, M.; Draganová, K. Concept of Magnetic Microwires Based Magnetometer
for UAV Geophysical Survey. In Proceedings of the International Conference on New Trends in Signal Processing (NTSP),
Demanovska Dolina, Slovakia, 14–16 October 2020; pp. 1–5.
12. Burkert, F.; Fraundorfer, F. UAV-based monitoring of pedestrian groups. ISPRS-Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci.
2013, 40, 67–72. [CrossRef]
13. Avola, D.; Foresti, G.L.; Martinel, N.; Micheloni, C.; Pannone, D.; Piciarelli, C. Aerial video surveillance system for small-scale
UAV environment monitoring. In Proceedings of the 2017 14th IEEE International Conference on Advanced Video and Signal
Based Surveillance (AVSS), Lecce, Italy, 29 August–1 September 2017; pp. 1–6.
Drones 2022, 6, 213 16 of 17

14. Grote, M.; Cherrett, T.; Oakey, A.; Royall, P.G.; Whalley, S.; Dickinson, J. How Do Dangerous Goods Regulations Apply to
Uncrewed Aerial Vehicles Transporting Medical Cargos? Drones 2021, 5, 38. [CrossRef]
15. Escribano Macias, J.J.; Angeloudis, P.; Ochieng, W.Y. Optimal hub selection for rapid medical deliveries using unmanned aerial
vehicles. Transp. Res. Part C Emerg. Technol. 2020, 110, 56–80. [CrossRef]
16. Barnawi, A.; Chhikara, P.; Tekchandani, R.; Kumar, N.; Alzahrani, B. Artificial intelligence-enabled Internet of Things-based
system for COVID-19 screening using aerial thermal imaging. Future Gener. Comput. Syst. 2021, 124, 119–132. [CrossRef]
[PubMed]
17. Bogatov, S.; Mazny, N.; Pugachev, A.; Tkachenko, S.; Shvedov, A. Emergency Radiation Survey Device Onboard the UAV.
ISPRS-Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2013, XL-1/W2, 51–53. [CrossRef]
18. Lassak, M.; Draganova, K.; Blistanova, M.; Kalapos, G.; Miklos, J. Small UAV Camera Gimbal Stabilization Using Digital Filters
and Enhanced Control Algorithms for Aerial Survey and Monitoring. Acta Montan. Slovaca 2020, 25, 127–137.
19. Wang, G.; Han, N.; Lv, Y.; Zhang, D. UAV onboard electronic detection system based on simulation test technology. In Proceedings
of the 2013 IEEE 11th International Conference on Electronic Measurement & Instruments, Harbin, China, 16–19 August 2013;
pp. 608–611.
20. Braun, J.; Braunova, H.; Suk, T.; Michal, O.; Petovsky, P.; Kuric, I. Structural and Geometrical Vegetation Filtering-Case Study on
Mining Area Point Cloud Acquired by UAV Lidar. Acta Montan. Slovaca 2021, 26, 661–674.
21. Bozek, P.; Al Akkad, M.A.; Blistan, P.; Ibrahim, N.I. Navigation control and stability investigation of a mobile robot based on a
hexacopter equipped with an integrated manipulator. Int. J. Adv. Robot. Syst. 2017, 14, 6. [CrossRef]
22. Valiga, M.; Novotňák, J.; Šmelko, M.; Čekanová, A.; Košuda, M. Control System Design of UAV Starter-Generator Power Unit. In
Proceedings of the International Conference on New Trends in Aviation Development (NTAD), Chlumec nad Cidlinou, Czech
Republic, 26–27 September 2019; pp. 204–208.
23. Shvetsova, S.V.; Shvetsov, A.V. Ensuring safety and security in employing drones at airports. J. Transp. Secur. 2021, 14, 41–53.
[CrossRef]
24. Shvetsov, A.V. Evaluation of the Prospects for the Admission of Electric Multicopters to Operation at Airports, Subject to the Joint
Use of Modern Methods of Ensuring Safety of the Movement. World Electr. Veh. J. 2022, 13, 79. [CrossRef]
25. Michalowski, B.; Varano, N. UAV flight test characterization using minimal test equipment. In Proceedings of the 2017 Interna-
tional Conference on Unmanned Aircraft Systems (ICUAS), Miami, FL, USA, 13–16 June 2017; pp. 1737–1741.
26. Ambroziak, L.; Gosiewski, Z. Preliminary UAV Autopilot Integration and In-Flight Testing. Solid State Phenom. 2013, 198, 232–237.
[CrossRef]
27. Lipovský, P.; Szőke, Z.; Moucha, V.; Jurč, R.; Novotňák, J. Data Acquisition System for UAV Autopilot and Operator Evaluation.
In Proceedings of the 2019 Modern Safety Technologies in Transportation (MOSATT), Košice, Slovakia, 28–29 November 2019;
pp. 98–103.
28. Xuan Mung, N.; Nguyen, N.P.; Pham, D.B.; Dao, N.N.; Hong, S.K. Synthesized Landing Strategy for Quadcopter to Land Precisely
on a Vertically Moving Apron. Mathematics 2022, 10, 1328. [CrossRef]
29. Lee, J.W.; Xuan-Mung, N.; Nguyen, N.P.; Hong, S.K. Adaptive altitude flight control of quadcopter under ground effect and
time-varying load: Theory and experiments. J. Vib. Control 2021, 2021, 10775463211050169. [CrossRef]
30. Waliszkiewicz, M.; Wojtowicz, K.; Rochala, Z. Experimental method of controller tuning for quadcopters. In Proceedings of the
2017 6th International Conference on Systems and Control (ICSC), Batna, Algeria, 7–9 May 2017; pp. 165–170.
31. Obias, K.C.U.; Say, M.F.Q.; Fernandez, E.A.V.; Chua, A.Y.; Sybingco, E. A Study of the Interaction of Proportional-Integral-
Derivative (PID) Control in a Quadcopter Unmanned Aerial Vehicle (UAV) Using Design of Experiment. In Proceedings of the
2019 IEEE 11th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control,
Environment, and Management, Laoag, Philippines, 29 November–1 December 2019; pp. 1–4.
32. Cedro, L.; Wieczorkowski, K. Optimizing PID controller gains to model the performance of a quadcopter. Transp. Res. Procedia
2019, 40, 156–169. [CrossRef]
33. Abdalla, M.; Albaradie, S. Real Time Optimal Tuning of Quadcopter Attitude Controller Using Particle Swarm Optimization. J.
Eng. Technol. Sci. 2020, 52, 745–764. [CrossRef]
34. Chehadeh, M.; Boiko, I. Design of Rules for In-Flight Non-Parametric Tuning of PID Controllers for Unmanned Aerial Vehicles. J.
Frankl. Inst. 2018, 356, 474–491. [CrossRef]
35. Munro, N.; Söylemez, M. Fast Calculation of Stabilizing PID Controllers for Uncertain Parameter Systems. IFAC Proc. Vol. 2000,
33, 549–554. [CrossRef]
36. Pounds, P.; Mahony, R.; Corke, P. Modelling and control of a quad-rotor robot. Control Eng. Pract. 2006, 18, 691–699. [CrossRef]
37. Castillo, P.; Lozano, R.; Dzul, A. Stabilization of a mini rotorcraft with four rotors. IEEE Control Syst. Mag. 2005, 25, 45–55.
38. Gremillion, G.; Humbert, J. System Identification of a Quadrotor Micro Air Vehicle. In Proceedings of the AIAA Atmospheric
Flight Mechanics Conference, Toronto, ON, Canada, 2–5 August 2010; p. 7644.
39. Mustapa, M. Altitude Controller Design for Quadcopter UAV. J. Teknol. 2015, 74, 85–88. [CrossRef]
40. Budnyaev, V.; Filippov, I.; Vertegel, V.; Dudnikov, S. Simulink-based Quadcopter Control System Model. In Proceedings of the
2020 1st International Conference Problems of Informatics, Electronics, and Radio Engineering (PIERE), Novosibirsk, Russia,
10–11 December 2020; pp. 246–250.
Drones 2022, 6, 213 17 of 17

41. Sabir, A.; Zakriti, A. Modeling of a Quadcopter Trajectory Tracking System Using PID Controller. Procedia Manuf. 2019, 32,
564–571.
42. Salameh, I.; Ammar, E.; Tutunji, T. Identification of Quadcopter Hovering Using Experimental Data. In Proceedings of the
2015 IEEE Jordan Conference on Applied Electrical Engineering and Computing Technologies (AEECT), Amman, Jordan,
3–5 November 2015.
43. Zhang, Y.; Chen, Z.; Zhang, X.; Sun, Q.; Sun, M. A novel control scheme for quadrotor UAV based upon active disturbance
rejection control. Aerosp. Sci. Technol. 2018, 79, 601–609. [CrossRef]
44. Dief, T. Review: Modeling and Classical Controller of Quad-rotor. Int. J. Comput. 2015, 5, 314–319.
45. Elkholy, H.; Habib, M.K. Dynamic Modeling and Control Techniques for a Quadrotor. In Handbook of Research on Advancements in
Robotics and Mechatronics; Habib, M.K., Ed.; American University: Cairo, Egypt, 2015; p. 408.
46. Al-Shabi, M.A.; Hatamleh, K.S.; Asad, A.A. UAV dynamics model parameters estimation techniques: A comparison study. In
Proceedings of the 2013 IEEE Jordan Conference on Applied Electrical Engineering and Computing Technologies (AEECT),
Amman, Jordan, 3–5 December 2013; pp. 1–6.
47. Jakubowski, A.; Kubacki, A.; Minorowicz, B.; Nowak, A. Analysis Thrust for Different Kind of Propellers. In Advances in Intelligent
Systems and Computing Progress in Automation, Robotics and Measuring Techniques; Institute of Mechanical Technology: Poznań,
Poland, 2015; Volume 350, pp. 85–90.
48. Zabunov, S.; Mardirossian, G. Scales for measuring UAV micro-motor static thrust. Aerosp. Res. Bulg. 2018, 30, 96–102. [CrossRef]
49. Niemann, H.; Miklos, R. A Simple Method for Estimation of Parameters in First order Systems. In Proceedings of the 11th
European Workshop on Advanced Control and Diagnosis, Berlin, Germany, 13–14 November 2014; Volume 570, p. 012001.
50. Novotňák, J.; Šmelko, M.; Andoga, R.; Lipovský, P.; Fil’ko, M. Modeling of the tensometric measuring system. Acta Avion. 2019,
21, 18–23. [CrossRef]
51. Novotňák, J.; Szöke, Z.; Šmelko, M.; Lipovský, P.; Fil’ko, M.; Košuda, M. Simple Filtering Algorithms for the Needs of Measuring
UAV Parameters. In Proceedings of the International Conference on New Trends in Signal Processing (NTSP), Demanovska
Dolina, Slovakia, 14–16 October 2020; pp. 1–5.
52. Belyaev, A.; Tutov, I.; Butuzov, D. Analysis of noisy signal restoration quality with exponential moving average filter. In
Proceedings of the International Siberian Conference on Control and Communications (SIBCON), Moscow, Russia, 12–14 May
2016; pp. 1–4.
53. Varshavskiy, I.E.; Krasnova, A.I.; Polivanov, V.V. Efficiency estimation of the noise digital filtering algorithms. In Proceedings of
the 2019 IEEE Conference of Russian Young Researchers in Electrical and Electronic Engineering (EIConRus), Saint Petersburg
and Moscow, Russia, 28–31 January 2019; pp. 724–727.