Increasing the Autonomy of the Unmanned Aerial Platform

Wojciech Stecz

1 a

and Marcin Chodnicki

2 b

Faculty of Cybernetics, Military University of Technology, Warsaw, Poland

Air Force Institute of Technology, Warsaw, Poland

Keywords:

UAV, Autonomy, SysML, State Machine, Hardware-in-the-Loop.

Abstract:

The article presents the principles of designing a reliable architecture supporting Unmanned Aerial Vehi-

cle (UAV) control, taking into account the need to handle hazardous situations occurring during the ﬂight.

Detailed attention was paid to the description of the UAV architecture components that affect the ability to

perform autonomous missions, understood as a ﬂight without contact with the Ground Control Station (GCS).

The method of designing UAV ﬂight algorithms in the conditions of occurrence of gusts of wind was pre-

sented. The principles of modeling the behavior of UAVs in situations of a potential air collision with another

platform or a collision with a terrain obstacle are described. Principles of modeling the hierarchy of handling

hazardous situations are presented. The developed models were tested on a computer architecture based on

ARM processors using the Hardware-in-the-Loop (HIL) technique. The presented solution uses a system of

UAV control computers in the form of a Flight Control Computer (FCC) based on a real-time operating system

(RTOS), and a Mission Computer (MC) based on a Linux system integrated with a Robot Operating System

(ROS). A method of integrating tasks related to the management of mission implementation with the algo-

rithms ensuring ﬂight safety of the air platform is presented. The research was carried out on the basis of the

UAV mathematical model, stabilization and navigation algorithms and the Dryden turbulence model.

1 INTRODUCTION

Descriptions of the architecture of modern unmanned

aerial platforms are a frequently discussed topic in

scientiﬁc research in the ﬁeld of unmanned aerial plat-

forms. A good example would be work (Sanchez-

Lopez et al., 2016). The designers of such systems

agree that each unmanned platform that can operate

autonomously, i.e. safe ﬂight without contact with

GCS, must be equipped with the following comput-

ers: FCC (Flight Control Computer) and MC (Mis-

sion Computer) (Pastor et al). FCC, depending on

the size of the UAV in which they are used, can be

advanced control systems containing, in addition to

PID-based ﬂight controllers, sensors such as GPS /

INS or ADC (Air Data Computer). The main task

of the FCC is to perform the role of an autopilot that

controls the ﬂight between successive points. MC is

a unit that supervises the order of completing tasks

speciﬁed in a given mission. Both of these devices

cooperate with each other during the ﬂight of the plat-

form. Depending on the advancement of the algo-

https://orcid.org/0000-0002-5353-5362

https://orcid.org/0000-0003-1348-289X

rithms implemented on these copunter, the UAV may

have the capability of autonomous ﬂight. The article

assumes that the UAV can operate autonomously if it

can ﬂy safely for other air platforms in the event of a

loss of communication with GCS. We ignore the le-

gal aspects of autonomous ﬂights due to the breadth

of the topic.

In this article, we focus on tasks related to the im-

plementation of the reconnaissance mission of the un-

manned platform, which is equipped with an EO / IR

head and a SAR radar. We present a method of mod-

eling the behavior of UAVs in SysML, which takes

into account the implementation of the mission in the

conditions of loss of communication with GCS. In

this case, the air platform must operate autonomously.

This means that the MC and FCC computers should

have built-in algorithms for handling the implementa-

tion of subsequent tasks planned for the mission, han-

dling tasks in the event of adverse weather conditions

and ﬂight control in the event of hazardous situations.

At this point, it is worth focusing on the types of al-

gorithms mentioned above.

In the case of the ﬁrst group of algorithms, which

includes activities related to reconnaissance tasks

Stecz, W. and Chodnicki, M.

Increasing the Autonomy of the Unmanned Aerial Platform.

DOI: 10.5220/0011140100003266

In Proceedings of the 17th International Conference on Software Technologies (ICSOFT 2022), pages 365-373

ISBN: 978-989-758-588-3; ISSN: 2184-2833

365

planned in the mission, the MC of the air platform

is designed to monitor the position of the UAV and, in

certain waypoints, turn on and off the reconnaissance

sensors. MC additionally conﬁgures individual sen-

sors, as described in (Stecz and Gromada, 2020). In

this article, we do not broadly describe the operation

of the air platform in the event of loss of contact with

the GCS, when tasks can be performed on a scheduled

basis and weather conditions do not affect the ﬂight.

It is much more complicated to carry out a re-

connaissance mission in conditions of high wind

speed, and even more so with periodic gusts. Un-

der these conditions, the UAV must have implemented

algorithms belonging to the second group mentioned

above, periodically changing the operation of ﬂight

regulators, as shown in the Section 3. For example,

when it is necessary to recognize an object using the

SAR radar, which requires high stability of the UAV

ﬂight, in the reconnaissance section, the FCC com-

puter, due to the MC request, restricts the operation

of the tilt regulator. This is based on the assumption

that the UAV does not ﬂy exactly along the route, but

is acceptable for this type of task.

The third group of algorithms are those securing

the platform’s ﬂight and ensuring the safety of the

UAV in the air. These algorithms include algorithms

for avoiding collisions with terrain obstacles, algo-

rithms for avoiding collisions with other air platforms

and algorithms for preventing UAVs from ﬂying out-

side the permitted zone. Some of them are described

in (Stecz and Gromada, 2022). In practice, there are

many more such algorithms and they can be catego-

rized according to the hierarchy of importance of the

situations they describe. The highest category situa-

tions include the sudden loss of UAV ﬂight altitude

when communication, spatial and usually geographic

orientation is lost. In this case, the UAV mission com-

puter must immediately start the rescue procedure.

Usually it triggers an emergency procedure which, for

smaller platforms, means the parachute will be thrown

out. Lower priority situations were previously men-

tioned and are associated with potential collisions.

The lowest priority situations are those that do not af-

fect ﬂight safety, but potentially delay the implemen-

tation of the mission plan.

The rest of the article is as follows. The Section 2

describes examples and important publications from

the area presented in the article. The Section 3 shows

the method of modeling selected procedures imple-

mented on the air platform in SysML in accordance

with the basic assumptions of MBSE. In particular,

the focus was on describing two types of implemented

functions: modifying the operation of PID regulators

during the ﬂight and verifying possible collisions with

terrain and other air platforms. The Section 4 presents

exemplary results of regulators controlled by MC al-

gorithms. The Section 5 summarizes our achieve-

ments and indicates possible further directions for the

development of algorithms enhancing the autonomy

of UAV ﬂight.

2 RELATED WORKS

The articles (Sanchez-Lopez et al., 2016), (Boubeta-

Puig et al., 2018) present the general structure of

the unmanned autonomous system, which allows for

making decisions about changing the trajectory by

UAV control computers in the absence of GCS con-

trol. The system consists of several modules respon-

sible for the implementation of the mission plan in an

autonomous mode.

Another example of UAV architecture is presented

in (Ilarslan et al., 2011). In this approach, the MC

computer acts as the main control system, therefore it

is based on the RTOS real-time system, and the FCC

autopilot is a slave system.

The FCC must be equipped, as previously men-

tioned, with a state machine with built-in special and

emergency logic, thanks to which it is able to indepen-

dently determine if the MC is malfunctioning. There-

fore, the MCs of smaller platforms are developed on

the Robot Operating System software. ROS is widely

used in robotics, it allows you to divide the entire

system into individual nodes, thanks to which adding

new functionalities is much easier. MC equipped with

software to supervise the correct implementation of

the mission, taking into account the operation of the

payload, supports the FCC. This conﬁguration allows

better use of the UAV’s capabilities. Examples of de-

scriptions of special situations in the form of state ma-

chines implemented on the MC of the air platform can

be found in the works (Wang et al., 2019), (Stecz and

Gromada, 2022), (Stecz and Kowaleczko, 2021).

It is worth noting that when the platform uses the

MC computer, which acts as a computing unit, e.g.

for avoiding obstacles or modifying the ﬂight route,

in this case the ROS system is used. In (Carvalho

et al., 2017) the open-source PX4 autopilot - FCC was

combined with a computer based on Linux and ROS.

Additionally, ROS is often used to prototype mission

planning algorithms and test them in a 3D virtual en-

vironment (Zhang et al., 2015). The FCC is responsi-

ble for the basic functions of stabilization and control

and is able to work independently of the supporting

MC. The communication between the FCC and MC

usually takes place via network interfaces to which

other sensors are also connected. This connection al-

ICSOFT 2022 - 17th International Conference on Software Technologies

366

lows direct access to both FCC and MC data.

The integration of FCC and MC may involve the

exchange of information aimed at modifying the con-

trols of the FCC regulators. The MC may send the

FCC a revision of the mission plan in a situation

where the analyzes conducted in the MC show that

due to changes in weather conditions, the UAV is un-

able to complete the task set on time. In this case,

the MC modules determine the mission correction by

solving optimization tasks of the VRPTW type shown

for example in the works (Siemiatkowska and Stecz,

2021). Another type of integration is the impact on

the current FCC control to enforce speciﬁc air plat-

form behavior. For example, in order to properly scan

the area with the use of SAR radar, it is important that

the UAV does not roll by more than the maximum

possible angle on a given route segment, which was

shown in the article. The MC can interfere with the

operation of the FCC in the event of collision detec-

tion and support the FCC in determining a UAV-safe

ﬂight trajectory, as described in (Stecz and Gromada,

2022).

3 MODELS AND METHODS

3.1 UAV Architecture Description

The architecture of the developed system shown in

Fig. 1 has been divided into three main compo-

nents: Mission Computer (MC), Flight Control Com-

puter (FCC) and Wireless Communication Subsys-

tems (WCS). This article does not deal with the radio

link subsystem. Human-Machine-Interface (HMI) is

dedicated software for managing the unmanned sys-

tem located at the Ground Control Station (GCS). The

ﬂight controller (High-Level Controller) consists of

three main subsystems:

• Stabilization of the platform - algorithms based on

PID regulators, which are designed to stabilize the

angular position of the UAV in space.

• Stabilization and control of the height, vertical

speed and ﬂight speed of the platform - these al-

gorithms are also based on PID controllers sup-

ported by state machines and mathematical algo-

rithms based on energy estimation.

• Navigation algorithms - course control, ﬂight

along a given route, line, etc. These algorithms

were also based on PID regulators supported by

mathematical algorithms and input/output signal

shaping systems.

The High-Level Controller, on the basis of the set

values and determined by the State Observer, deter-

mines the required deﬂections of the controls and the

propulsion system, so that the platform performs the

set ﬂight parameters.

Figure 1: Diagram of the architecture of the unmanned

aerial platform.

The Low-Level Controller is responsible for con-

verting the signals developed by the High-Level Con-

troller from angular and percentage values to appro-

priate hardware values enabling the control of servos,

motor controllers or other actuators.

In the architecture presented in the work, MC is

based on the ROS. The algorithms for determining

the ﬂight trajectory are supported by geodetic data,

other reconnaissance data and external auxiliary sen-

sors. ROS is a dedicated suite of libraries and soft-

ware for developing robotics software. Thanks to it,

in the presented architecture it was possible to divide

the software architecture into smaller modules, the so-

called nodes.

3.2 Modeling Autonomy

The functional requirements for UAV usually come

from two sources. The ﬁrst of them are customer

requirements in terms of system functionality. The

second source is the requirements imposed by safety

standards. They are deﬁned as part of the Functional

Risk Analysis (FHA) documents. All the deﬁned re-

quirements are the basis for the development of the

system architecture and a detailed description of the

scenarios of the operation of the air platform per-

forming the reconnaissance task. Each functional re-

quirement is transformed into a system performance

scenario or a single function of the designed system.

The following parts of the article present examples of

UAV operation models that take into account three se-

lected ﬂight scenarios:

• ﬂight in contact with the GCS (the radio link is

not disturbed and the GPS is working properly)

• ﬂight without communication with GCS in the

presence of gusts of wind on the route segment

where SAR radar is used

Increasing the Autonomy of the Unmanned Aerial Platform

367

• ﬂight in the event of detection of a potential air

collision of two air platforms equipped with ADS-

B (Automatic Dependent Surveillance - Broad-

cast) systems.

Figure 2 shows an example of a Use Case dia-

gram covering the scenarios presented in the article

(Use Case is a way of modeling the fulﬁllment of re-

quirements by system functions). The article presents

the UAV architecture that allows the MC to modify

the operation of PID controllers built into the FCC.

Modiﬁcations to the setting of the maximum allow-

able platform roll in ﬂight were presented. In the fur-

ther part of the work, we assumed that the functional

analysis of the system operation (FHA) showed that

during the ﬂight of the platform, a collision may oc-

cur with another aircraft equipped with ADS-B. We

also assume that the UAV may not be in contact with

the GCS in the time preceding the collision. If, during

detection of a potential collision, the UAV had radio

contact with the GCS, the platform will not take any

action and will wait for the controls sent by the pi-

lot. We also assume that the UAV is equipped with

ADS-B with the ability to receive the signal gener-

ated by other aviation platforms. Otherwise, the sys-

tem should only prompt the pilot on possible actions,

but the ﬁnal decision must always be with the pi-

lot. Moreover, other ﬂying platforms must also have

ADS-B or Sense and Avoid systems. Otherwise, the

UAV will not have enough data to react. Based on

these assumptions, the model presented in the article

was developed.

Figure 2: Use Case that aggregates activities supporting an

autonomy of UAV.

Figure 2 shows the Use Case model, which de-

scribes the possible activities during the UAV ﬂight

along the given waypoints (often referred to as the

WAYPOINT mode). The ﬁgure shows the basic case

(basic scenario) - Flight in WAYPOINT mode, which

includes scenarios related to autonomous operation

(without contact with GCS) and actions required in

the event of hazardous situations. The autonomous

scenarios are represented by the SAR Object Recog-

nition in Autonomous Mode scenario. This scenario

includes support for the FCC PID Parameter Correc-

tion scenario that allows UAV roll limitation during

target recognition using SAR.

Hazardous scenarios are described by the collec-

tion of UAV Collision Avoidance scenarios, which is

included in the main scenario. This group includes

procedures for handling a sudden loss of ﬂight alti-

tude by UAVs, response to a potential air collision

with another platform, and response to a terrain ob-

stacle, etc.

Figure 3 shows a state machine model that de-

scribes the operation of the system in ﬂight along

predeﬁned route points. The model shows three or-

thogonal states. Actions described in these states can

be performed in parallel. In the presented simpli-

ﬁed case, in the ﬁrst orthogonal state, the actions per-

formed during the ﬂight over the points were deﬁned.

In this state, the FCC and MC cooperate with each

other, with the assumption that the MC takes over the

reconnaissance service using SAR and the FCC con-

trols the ﬂight of the UAV. Details are presented in

Section 3.3.

The second state shows the handling of the highest

level tactical situation related to the uncontrolled loss

of ﬂight altitude by the UAV. In this state, the FCC

software activates the parachute if it is not possible to

stop the uncontrolled fall of the UAV. This is the usual

procedure for small drones.

In the third orthogonal state, there are procedures

for handling situations related to potential collisions.

Due to the fact that in this case detection of collisions

and determination of a new ﬂight trajectory requires

some computing power, MC is responsible for han-

dling this group of situations. Critical functions from

this group are performed serially in a separate thread

of the MC computer, which is required by the ﬂight

safety rules.

The basic scenario carried out by a UAV that per-

forms a ﬂight along a predeﬁned route consists in go-

ing through the following states in sequence:

FW M1 → FW M2 → FW M3 → FW M2 →

SUP → FW M1 (selection of the next waypoint, ﬂight

to a point, optional conﬁguration of the recognition

sensor, checking the threat status and going to the se-

lection of the next waypoint).

In the case when the UAV starts recognition with

the use of SAR at a given point, the weather condi-

tions are tested at the input in the FWM3 state (en-

ICSOFT 2022 - 17th International Conference on Software Technologies

368

Figure 3: State machine presenting some typical excep-

tional situation handling.

try:checkWindParams()). The UAV’s allowable roll

angle is then calculated in the main state routine

(do:calculateFCCparams()) (see the 3.3). In case of

strong wind, the MC sends a request to the FCC to

minimize the roll angle. On the input of the FWM2

state in the function (entry:checkRecognition()) it is

checked whether the recognition will be performed

on the route segment with the use of a given sensor

type. If so, the do:setFCCparams() procedure is run,

which sets the maximum allowed UAV roll angle. On

exiting the state in the exit:resetFCCparams() proce-

dure, the default parameters of the sensor settings are

restored.

The described approach is very simpliﬁed and

presents only the most important elements related to

the reconnaissance carried out on the ﬂight segment.

Note that the procedures described in the states (en-

try:, do:, exit:) are then detailed in the form of se-

quence or activity diagrams.

An alternative processing scenario will occur

when a dangerous situation is detected in one of the

other orthogonal states. At the same time, it should

be remembered that the detection of a sudden loss

of height performed by the FCC has a higher prior-

ity than the detection of a potential collision (which

is supervised by the MC). It’s not so obvious why

the FCC-designated state takes precedence. How-

ever, when the reader realizes that the sudden loss of

UAV ﬂight altitude is associated with the immediate

loss of communication with GCS, it becomes clear

that the UAV must react to such a state faster than to

any other. The supervisor module is responsible for

the appropriate assessment of the importance of the

states, which is shown on the model in the form of the

SUPERVISOR state.

When a dangerous situation is detected regarding

a possible collision with another platform or a colli-

sion with a terrain obstacle, the processing sequence

is as follows.

FW M1 → FW M2 → SUP → FW M4 →

FW M2 → SUP → FW M1 (following waypoint

selection, ﬂight to point, emergency test, emergency

collision avoidance, a continuation of the ﬂight to a

point, a test of the occurrence of an emergency, and

selection of the next waypoint). do:testCollision()

and do:ﬁndWPT() are given numerical algorithms

that are used to implement them. Collision testing

is performed according to algorithms presented in

(Stecz and Gromada, 2022). It is worth noting that

the determination of the collision situation using the

geometric methods takes a short time so that indi-

vidual tests can be performed sequentially without

risk.

As part of handling emergency situations, poten-

tial collisions between air platforms and a collision

with a terrain obstacle are investigated (see states

ExS1 and ExS2). Of course, this set also includes the

emergency situation related to the UAV crossing the

border of the mission area. Since this group of states

may have the same priority, the MC algorithms must

work efﬁciently enough to verify in a very short time

which of the situations may occur in the foreseeable

future. For example, in the event of a potential colli-

sion with a terrain obstacle, UAV algorithms must de-

termine the route point above the obstacle. By default,

the platform can lower the ﬂight altitude in certain sit-

uations, which is not possible in the event of a poten-

tial collision with a terrain obstacle. The processing

itself is strictly dependent on the adopted principles

of UAV operation and conditions beyond the scope of

the article.

When an uncontrolled loss of height is detected

directly threatening the UAV, the processing sequence

is as follows.

FW M1 → FW M2 → SUP → FW M5 →

FW M2 → SUP (after selecting a waypoint, ﬂight to

a point, emergency test, UAV ﬂight stabilization or

parachute release, ﬂight continuation to a point after

successful UAV stabilization, test the occurrence of

an emergency situation). If the attempt to stabilize

the ﬂight fails, the parachute discharge ends the UAV

ﬂight.

The reader can see that the state machines do not

need to show a clear separation of the tasks of the in-

dividual computers (in this case the FCC and MC).

Increasing the Autonomy of the Unmanned Aerial Platform

369

State machines are an abstract description of the op-

eration of a system that allows for the presentation

of the concurrent operation of its components. If it

is necessary to assign individual procedures to equip-

ment, a diagram detailing the data processing method

should be drawn for each of the procedures listed on

the basis of state machines. Only such a diagram can

be assigned to a speciﬁc resource on which processing

takes place. It is important to emphasize once again

that according to the concept of UML and SysML, the

actions described in states are processing algorithms

detailed in the form of sequence or activity diagrams.

When procedures belonging to states can be de-

scribed with numerical algorithms, there is no need

to generate additional activity diagrams and connect

them with state machines. It is worth bearing in mind,

however, that adding an additional, very simpliﬁed di-

agram, in which one or two activities is described, is

useful for the purposes of mapping low-level require-

ments into implemented functions. Therefore, it is not

worth skipping this step.

3.3 MC and FCC Integration

The integration of FCC and MC will be described

on the example of correcting the PID parameters (see

Fig. 4) by the MC during the ﬂight on the route seg-

ment where the SAR radar was used, when the con-

trol of the UAV’s roll angle is very important. De-

tailed rules of SAR operation are described in the

work (Stecz and Gromada, 2020). It is worth men-

tioning that in order to perform a correct surface scan,

the SAR must scan through a ﬂight segment of the

length equal to the so-called synthetic aperture. The

greater the distance and the more accurate the scan,

the longer the ﬂight segment must be - usually at least

several hundred meters. In this section, the roll angle

should be minimal and not more than a few degrees.

It is acceptable, however, that during the preparation

of the scan, the UAV will not keep the course and the

platform will be carried away by the wind.

The most important thing when performing a SAR

scan is minimizing rolls. Therefore, the MC that con-

trols the scanning process must be able to interfere

with the maximum range of platform roll set by the

FCC in its regulators when ﬂying between waypoints.

This is what the External Roll Command Limit con-

trol input is for. The scheme of the roll angle control

procedure for the described situation in the SysML

modeling language in the form of a sequence diagram

is shown in Fig. 5. The MC software checks in a

loop whether the UAV has reached the point that is

the beginning of the SAR recognition segment. If so,

MC, having data on wind parameters, sets the maxi-

mum allowable platform roll. Additionally, it sets the

value of the Heading parameter, which indicates the

direction of the UAV ﬂight.

Figure 4: Block diagram of FCC built-in PID controllers.

Two types of control blocks are visible: the UAV navigation

block and the UAV stability control block.

When the UAV leaves the reconnaissance route

segment, the default values of the allowable platform

roll angles are restored, which allows you to return to

the commanded trajectory.

Figure 5: Diagram of the procedure for controlling the UAV

roll angle during the recognition of the object by the SAR.

ADC provides data on wind conditions. Data is exchanged

between MC and FCC. MC is responsible for setting the

maximum allowable roll angle.

In some situations it is also useful to maintain a

given UAV course, which the MC may impose on the

FCC controllers. Course control is used in practice

when the MC detects a dangerous situation and ana-

lyzes it for a speciﬁed period of time. In this case, the

UAV switches to the ﬂight mode for the set course at

the minimum allowable speed, which is safe for the

UAV under the given conditions. The MC calculates

and sends the desired course to the PID controller in

the control block. Depending on the weather condi-

tions, the speed relative to the ground may also be

reduced, which allows to limit the length of the UAV

ﬂight, during which the UAV carries out self-testing

procedures of the systems. This is what the Exter-

nal Heading command control input is for. When MC

ICSOFT 2022 - 17th International Conference on Software Technologies

370

sets a heading, the Enable signal is set for Distance to

Heading PID Controller.

4 RESULTS

Operation of the described platform was tested in a

simulation environment, the architecture of which is

presented in Fig. 6. A Hardware-in-the-Loop envi-

ronment was built in which the physical FCC and MC

devices (on which the target software was installed)

were integrated. Only the position reported to the

FCC and MC by GPS/INS and ADC was simulated.

The operation of other on-board equipment was not

simulated as it was not necessary for the purpose of

the research. Changes in wind strength and direction

as well as wind gusts were simulated by a component

imitating ADC. It was assumed that the wind direc-

tion and speed were within the limits set for this type

of UAV. This also applied to gusts of wind. Such as-

sumptions are correct because the UAV has a greater

tolerance to the wind force than the SAR radar, which

should work in conditions without platform roll.

Figure 6: HIL simulation architecture of the unmanned

aerial platform.

Fig. 7 shows result of simulation performed in

HIL test of route planned and control algorithms de-

scribed above. For the clarity, only UAV ﬂights on

one of the route segments of the designated route have

been considered.

Another simulation was carried out with a view to

checking the functioning of the bank angle limitation

algorithm for the implementation of the SAR mission.

The Fig. 8 shows the ﬂight trajectory without limiting

the bank angle value. This ﬁgure shows that despite

the disturbances in the form of wind, UAV follows

Figure 7: Simulated UAV trajectory for predeﬁned wind

speed and direction.

the route according to the designated path. However,

in the case of ﬂight with a bank angle limitation of 5

degrees for SAR missions, the UAV continues to try

to follow the given path shown in Fig. 9. However,

the predeﬁned path is achieved after a much greater

adjustment time. This time is due to the fact that the

minimum turning radius of the air platform has in-

creased.

On the other hand, UAV is still able to follow

the given route. However, in this case, the MC must

take into account the change in maneuverability of the

UAV platform when correcting the mission.

Increasing the Autonomy of the Unmanned Aerial Platform

371

Figure 8: An example of a UAV ﬂight simulation on a sec-

tion when FCC has no restrictions on platform roll.

Figure 9: An example of a UAV ﬂight simulation on a route

segment, when the maximum roll angle of the UAV has

been deﬁned.

5 CONCLUSIONS

The article presents the principles of designing reli-

able architecture supporting UAV control, taking into

account the need to handle exceptional situations oc-

curring during the ﬂight. At the same time, the set of

exceptional situations includes all situations that en-

sure the safety of the UAV ﬂight while ensuring its

correct operation in the event of loss of contact with

the GCS.

The architecture of modern unmanned aerial plat-

forms with built-in capability to carry out autonomous

missions was characterized in detail. Architecture

requires the integration of UAV systems with ﬂight

route planning and correction algorithms. Particular

attention was paid to UAV ﬂight methods in the con-

ditions of wind gusts, the occurrence of which causes

great difﬁculties related to the recognition of objects

with the use of SAR radars.

The article refers to the algorithms that ensure,

above all, the safety of the autonomous UAV ﬂight.

All these algorithms must be implemented on the

platform operating in changing weather conditions in

zones where other air platforms may also appear. Of

course, the presented approach does not ensure the

safety of the UAV while ﬂying in any terrain. Cer-

tain assumptions were made in the work, which do

not always have to be met. For example, in order for

a UAV to avoid a collision with another UAV, it must

be equipped with ADS-B and each platform moving

in the area where the UAV is operating must also have

a position warning system. Without it, it is impossible

to detect a potential air collision.

An important element pointed out in the article are

the principles of modeling the hierarchy of handling

exceptional situations. Situations such as the sudden

loss of UAV altitude or the detection of a collision

with another air platform have the priority of service

much higher than the detection of problems with the

completion of the task in the assumed time.

The integration of mission management methods

and methods of handling emergency situations, tak-

ing into account wind conditions, is an innovative ele-

ment of the work. The research was carried out on the

basis of the UAV mathematical model developed for

the needs of the platform designed by ITWL, which

took into account the stabilization and navigation al-

gorithms and the wind turbulence model according to

Dryden’s concept.

The direction of further work concerns greater in-

tegration of the developed methods with the methods

of automatic detection of failures of devices constitut-

ing the equipment of the air platform.

ICSOFT 2022 - 17th International Conference on Software Technologies

372

ACKNOWLEDGEMENTS

The presented results were prepared with the use of

UAV models developed by ITWL.

REFERENCES

Boubeta-Puig, J., Moguel, E., S

anchez-Figueroa, F.,

Hern

andez, J., and Carlos Preciado, J. (2018). An au-

tonomous uav architecture for remote sensing and in-

telligent decision-making. IEEE Internet Computing,

22(3).

Carvalho, J., Juc

a, M., Menezes, A., Olivi, L., Marcato, A.,

and dos Santos, A. (2017). Autonomous uav outdoor

ﬂight controlled by an embedded system using odroid

and ros. 402:58–67.

Ilarslan, M., Bayrakceken, M. K., and Arisoy, A.

(2011). Avionics system design of a mini vtol uav.

IEEE Aerospace and Electronic Systems Magazine,

26(10):35–40.

Sanchez-Lopez, J., Pestana, J., and de la Puente, P. (2016).

A reliable open-source system architecture for the fast

designing and prototyping of autonomous multi-uav

systems: Simulation and experimentation. J Intell

Robot Syst, 84:779–797.

Siemiatkowska, B. and Stecz, W. (2021). A framework for

planning and execution of drone swarm missions in a

hostile environment. Sensors, 21(12).

Stecz, W. and Gromada, K. (2020). Determining uav ﬂight

trajectory for target recognition using eo/ir and sar.

Sensors, 20.

Stecz, W. and Gromada, K. (2022). Designing a reliable

uav architecture operating in a real environment. Appl.

Sci., 12(294).

Stecz, W. and Kowaleczko, P. (2021). Designing oper-

ational safety procedures for uav according to nato

architecture framework. In Proceedings of the 16th

International Conference on Software Technologies -

ICSOFT,, pages 135–142. INSTICC, SciTePress.

Wang, Y., Lei, H., Hackett, R., and Beeby, M. (2019).

Safety assessment process optimization for integrated

modular avionics. IEEE Aerospace and Electronic

Systems Magazine, 34(11):58–67.

Zhang, M., Qin, H., Lan, M., Lin, J., Wang, S., Liu, K., Lin,

F., and Chen, B. M. (2015). A high ﬁdelity simulator

for a quadrotor uav using ros and gazebo. In IECON

2015 - 41st Annual Conference of the IEEE Industrial

Electronics Society, pages 002846–002851.

Increasing the Autonomy of the Unmanned Aerial Platform

373