=Paper=
{{Paper
|id=Vol-3732/paper01
|storemode=property
|title=Collaborative decision-making during pre-simulation training: Optimal approach and radar vectoring procedures
|pdfUrl=https://ceur-ws.org/Vol-3732/paper01.pdf
|volume=Vol-3732
|authors=Tetiana Shmelova,Yuliya Sikirda,Maxim Yatsko,Volodymyr Kolotusha
|dblpUrl=https://dblp.org/rec/conf/cmse/ShmelovaSYK24
}}
==Collaborative decision-making during pre-simulation training: Optimal approach and radar vectoring procedures==
https://ceur-ws.org/Vol-3732/paper01.pdf
Collaborative decision-making during pre-simulation
training: Optimal approach and radar vectoring procedures
Tetiana Shmelova1,†, Yuliya Sikirda2,∗,†, Maxim Yatsko3,† and
Volodymyr Kolotusha1,†
1 National Aviation University, Liubomyra Huzara Ave., 1, Kyiv, 03058, Ukraine
2 Flight Academy of the National Aviation University, Stepana Chobanu Str., 1, Kropyvnytskyi, 25005, Ukraine
3 SmartLynx Airlines "Mazrudas", Marupes novads, LV-2167, Latvia
Abstract
Experts' forecasts show a significant increase in global demand for highly qualified aviation
professionals, including pilots and air traffic controllers, by 2030. Joint simulation training of
aviation specialists is a critical stage of professional education and plays a significant role in
further ensuring flight safety. At the briefing stage, the instructor checks the readiness of
trainees for practical activities, including collaborative decision-making (CDM) between the
aircraft crew and air traffic controller in complicated, complex, emergency situations. GPS
signals are currently subject to instability. In the case of interference with GPS navigation
equipment onboard an aircraft flying according to the rules of Area Navigation (RNAV), the pilot
and air traffic controller must be able to collaborate effectively in the radar vectoring
procedure. 28.3% of GPS problems are related to the descent and landing stages. The authors
proposed to use as the elements of testing at the stage of pre-simulation training the
collaborative performance of the training tasks "Choosing the optimal landing aerodrome" and
"Vectoring". Based on the expert judgment method, models of individual and CDM in choosing
the optimal landing aerodrome are developed. By the network planning method, the
synchronized vectoring operations of the pilot and air traffic controller in complicated situation
"Unable required position" are worked out. The significance of trainees' actions for the training
task "Vectoring" is obtained by fuzzy logic methods. A comparison of the comprehensive
assessment of trainees' performance in the training task "Vectoring" by additive and
multiplicative aggregation is considered. The expert system for multiplicative assessment of the
results of pre-simulation training will allow for making an objective conclusion about the
mastery of the necessary knowledge, skills, and abilities by trainees and the acquisition of
professional practice-oriented competencies in CDM.
Keywords
air traffic controller, decision-making matrix, emergency, expert judgment method, fuzzy logic,
landing aerodrome, network planning, pilot, training task, vectoring1
CMSE’24: International Workshop on Computational Methods in Systems Engineering, June 17, 2024, Kyiv, Ukraine
∗ Corresponding author.
† These authors contributed equally.
shmelova@ukr.net (T. Shmelova); sikirdayuliya@ukr.net (Yu. Sikirda); maxim_yatsko@i.ua (M. Yatsko);
kolotusha.vp@gmail.com (V. Kolotusha)
0000-0002-9737-6906 (T. Shmelova); 0000-0002-7303-0441 (Yu. Sikirda); 0000-0003-0375-7968
(M. Yatsko); 0000-0002-6772-5140 (V. Kolotusha)
© 2024 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
1. Introduction
1.1. Introduction to the problem
According to experts, the number of passengers and cargo transported by air will double
by 2035 [1]. The forecasts of the International Air Transport Association (IATA) and the
International Civil Aviation Organization (ICAO) show that in the next 20 years, the
demand for air transport will grow by an average of 4.3% per year [1, 2]. The aviation
industry is developing dynamically, opening up new, attractive prospects for professional
development, and is confidently emerging from the crisis. New aviation companies are
being established, creating new jobs that will require qualified aviation professionals.
Already in 2019, the aviation industry provided a total of 65.5 million jobs worldwide, and
despite the challenges faced by the aviation market, it is predicted that this statistic will
increase from year to year [3].
Taking into account the forecast values, the aviation transport system in the coming
years will need many highly educated specialists: pilots, air traffic controllers (ATCOs),
and engineers. Table 1 presents the results of research by the ICAO, which compares the
average number of specialists around the world who will need to be trained annually with
the training capabilities of existing institutions [4]. This indicates an existing shortage of
specialists, namely 160 000 pilots, 40 000 ATCOs, and 360 000 engineers.
Table 1
ICAO comparative analysis of the number of specialists who need to be trained annually
with the training capabilities of existing institutions [4]
Type of Actual number Number of Need for Training Annual
specialists of specialists in specialists annual system deficit of
2010 needed by 2030 training capabilities specialists
Pilots 463 386 980 799 52 506 44 360 8 146
ATCOs 67 024 139 796 8 718 6 740 1 978
Engineers 580 926 1 164 969 70 331 52 260 18 071
A comparable situation with staff is observed in the European region [5]. According to
forecasts, by 2036, the aviation segment will need [6]:
• 620 000 new pilots (67 pilots per day for aircraft with more than 100 seats).
• 125 000 ATCOs (13 new ATCOs every day).
Simulator training is a set of forms and methods of education that allow trainees to
develop practical skills using the theoretical knowledge of several academic disciplines by
performing complex tasks and exercises under the guidance of an instructor [7]. The
purpose of the training is to improve the work of aviation specialists and develop practical
skills in standard and non-standard situations. The quality and quantity of exercises, and
objective assessment of exercises affect the effectiveness of simulator training.
To optimize the effectiveness of the training following EUROCONTROL
recommendations, theoretical and practical training are combined from the very
beginning of the training using a pre-training system [7]. The training process starts with
the acquisition of skills on simulators (SA – skill acquisition), then the performance of
individual tasks (PTP – part-task practice) is practiced and continues with simulator
training. Guided SA (G-SA) is currently relevant – the acquisition of skills accompanied by
interactive assessment, commenting, and control over the trainees' actions. Guided
practice of partial task completion (G-PTP) is the practical implementation of specific
tasks, accompanied by comments, display of results, assessment of the trainees' actions,
and the possibility of feedback [7].
Effective use of simulators makes it possible: 1) to provide a gradation "from simple to
complex" of professional competence formation in safe conditions, especially when
practicing unpredictable and emergency flight conditions; 2) to optimize training
resources (involving instructors who are not actively involved in training); 3) to repeat
training exercises several times; 4) to develop skills and improve decision-making abilities
in conditions of uncertainty and time pressure (development of professional confidence).
Significant changes in the aviation system are "passed" through their modeling on
simulator equipment. Based on the correct strategy for using simulators, in which a
significant role is assigned to the instructor of practical training, it is possible to design a
systematic approach to the development of a set of professional competencies required by
the employer.
1.2. Motivation
Joint simulation training of aviation specialists is a critical stage of professional education
and plays a significant role in further ensuring flight safety. At the briefing stage, the
instructor checks the readiness of trainees for practical activities, including collaborative
decision-making (CDM) between the pilot and ATCO in complicated, complex, emergency
situations. However, there are currently no objective methods for assessing the
interaction between operators at all stages of the flight. This fact motivates to develop the
expert system for multiplicative assessment of aviation specialists' (pilot and ATCO) CDM
during pre-simulation training.
1.3. Contribution
This research contributes to the enhancement of the measurement of the aviation
specialists’ interaction in complicated, complex, emergency situations based on the expert
system for multiplicative assessment of the results of pre-simulation training. It will allow
making an objective conclusion about the mastery of the necessary knowledge, skills, and
abilities by trainees and the acquisition of professional practice-oriented competencies in
CDM in emergencies.
1.4. The organization of the paper
The paper consists of six sections. The first section includes an introduction, it
concentrates on the analysis of related works and problem statements. The second section
considers the development of the models of individual and CDM by the pilot and ATCO
when performing the training task "Vectoring". The third section discusses the expert
system for quantitative assessment of CDM by the pilot and ATCO during pre-simulation
training. The fourth part is results and discussions. The fifth section is the conclusion. The
sixth section describes the future research directions.
1.5. Formulation of the problem
The ever-increasing volume of air traffic places new demands on airspace capacity [2]. The
concept of Performance-Based Navigation (PBN) is being implemented to ensure the
efficiency of airspace use by providing direct routes, track accuracy, and high accuracy of
navigation systems [8]. Two main types of navigation procedures exist within PBN: Area
Navigation (RNAV) and Required Navigation Performance (RNP) [9, 10]. In this context,
RNAV refers to a particular navigation specification with a given lateral accuracy that must
be maintained for 95% of the flight time. RNP includes onboard Receiver Autonomous
Integrity Monitoring (RAIM) – a technology for the assessment of the Global Positioning
System (GPS) signals’ integrity. A satellite may be transmitting slightly incorrect
information, resulting in erroneous navigation data, but the GPS receiver cannot detect
this using standard methods. RAIM uses redundant signals to obtain and compare multiple
GPS coordinates, and a statistical function determines whether the error can be attributed
to any of the signals. However, despite the use of modern satellite navigation equipment,
the current problem is the instability of GPS signals, especially near war zones, which
negatively affects the correctness and accuracy of aircraft positioning (complicated flight
situation). For 46 hours, 873 aircraft in the Baltic region had problems with GPS signals
and related equipment, according to a Swedish OSINT analyst. [11]. In the case of 43
aircraft, GPS navigation was unavailable for more than two hours. Der Spiegel journalist
notes that the massive loss of signal in the Baltic Sea region may be the result of Russia's
"hybrid war" [11]. In such cases, the pilot can request help from the ATCO to receive
instructions on the correct flight course – a common task that is solved by the radar
vectoring procedure [12, 13].
According to the CDM concept [14], the effective interaction of aviation specialists is a
precondition for ensuring safety at all stages of an aircraft flight in standard and non-
standard situations. Aviation specialists must closely follow the regulatory documents that
have been reviewed in the course of their professional training and activities. At the same
time, the content of professional training documents and guidance documents often
differs, making it difficult to develop a single algorithm for joint actions, especially in
emergencies. Joint pre-simulation and simulation training of aviation specialists (pilots
and ATCOs) is used to prevent conflicts between decisions and actions of CDM participants
in real flight conditions [15].
Publication [16] discusses integrated models for training aviation specialists (pilots
and controllers), [17] – a game-based approach to implementing CDM at a leading
European airport, and [18] – partnership programs between aviation educational
institutions and airlines. The authors have presented new approaches to the improving of
aviation specialists’ professional activity (intelligent decision support system [19, 20, 21])
and practical training (artificial neural network for pre-simulation training [22], machine
learning [23], intelligent integrated training system "CDM – Education" [24]) in
emergencies using CDM.
Methods of CDM, proposed by the authors [19–24]:
• Method for integrating decision-making models in certainty, risk, and uncertainty
(deterministic and stochastic models).
• Method of subjective-objective CDM based on individual and collaborative
decision-making models.
• Method for managing the development of the situation by using the integration of
decision-making models (non-stochastic, stochastic, and deterministic models) and
CDM models (individual and collaborative decision-making models).
• Method of CDM modeling based on the priority of the factors influencing decision-
making.
• Method of CDM modeling based on the priority of the Hurvitz criteria.
• Method of Collaborative Decision-Making in education "CDM-E".
• Method of CDM in an emergency with a multi-step (multi-stage) decision-making
process.
However, no objective methods for assessing the interaction between operators at all
stages of flight in complicated, complex, emergency situations have yet been proposed.
Therefore, an urgent problem is the development of the expert system for multiplicative
assessment of aviation specialists' (pilot and ATCO) CDM during pre-simulation training.
The purpose of the work is to solve the following problems:
• Development of the models of individual and collaborative decision-making by the
pilot and ATCO when performing training tasks based on the expert judgment
method.
• Calculating the significance of trainees' actions for the training task "Vectoring" by
fuzzy logic methods.
• Comparison of the comprehensive assessment of trainees' performance in the
training task "Vectoring" by additive and multiplicative aggregation.
2. Models of individual and collaborative decision-making by the pilot
and air traffic controller when performing training tasks
According to [25], in the vast majority of reports from the aircraft, regarding problems
with obtaining navigation information, the crews required radar guidance (vectoring)
from ATC. In its turn, European Union Aviation Safety Agency (EASA) explicitly
recommends that in case of problems with receiving GPS signals, pilots should be
prepared to request and receive guidance (vectoring) from the ATCO as long as necessary
[26]. Based on IATA data [27], 28.3% of the problems in obtaining navigation information
from GPS are related to the descent and landing stages.
Input data based on real situation for training task:
1. Aircraft Boeing 737-8MAX.
2. Flight from Istanbul Sabiha Gökçen LTFJ (departure aerodrome) to Rize-Artvin
LTFO (arrival aerodrome). Alternate aerodrome in bad weather conditions (BWC)
Trabzon LTCG.
3. Runway in use at destination Rize-Artvin LTFO RW24 due to strong wind.
4. Suspected arrival route ZUBRE 1K following approach RNP Z RW24 (conventional
arrival ZUBRE 1B following approach VOR Z or NDB Z RW 24).
5. Weather:
• Wind 230/16 VRB 210V260.
• Visibility 10 km.
• Clouds BKN 2800’ SCT 1000’.
• Temperature 15, dew point 11.
• Pressure QNH 1021.
• No significant changes (with the subsequent deterioration of weather conditions).
6. There are three operators in the CDM: pilot (O1 ), ATCO (O2 ), and expert/Artificial
Intelligence (AI) (O3 ).
7. Factors are taken into account by all operators while decision-making:
• {aj } – factors influencing decision-making by operator O1 (pilot).
• {bj } – factors influencing decision-making by operator O2 (ATCO).
• {cj } – factors influencing decision-making by operator O3 (expert/AI).
The common objective factors for all operators (𝑎𝑗 , 𝑏𝑗 , 𝑐𝑗 ):
• a1 , b1 , c1 – the weather conditions.
• a2 , b2 , c2 – the distance to the applicable aerodromes/the quantity of fuel onboard.
• a3 , b3 , c3 – the tactical and technical characteristics of the runways.
• a4 , b4 , c4 – the flight and technical characteristics of the aircraft.
• a5 , b5 , c5 – the approach systems and navigation aids at the applicable aerodromes.
• a6 , b6 , c6 – Threat and Error Management (TEM).
The usual situation during the flight over Turkey territory is unstable GPS signal due to
military conflicts in Ukraine and Syria.
Trainees are performing two tasks during pre-simulation training (Figure 1):
• Task 1: Decision-making in uncertainty – choosing the optimal landing aerodrome
in BWC.
• Task 2: Decision-making in certainty – vectoring.
Black Sea
Task1
DM in uncertainty – Task 2
2 hours DM in BWC DM in certainty
VECTOR
Turkey
Figure 1: The tasks that trainees are performing during pre-simulation training.
2.1. Task 1: Decision-making in uncertainty – choosing the optimal landing
aerodrome
For rational CDM, all operators (pilot (𝑂1 ), ATCO (𝑂2 ), and expert/Artificial Intelligence
(AI) (𝑂3 )) are analyzing and considering the current flight situation. The operators are
composing the decision-making matrices, where alternative solutions are the departure
aerodrome Istanbul Sabiha Gökçen LTFJ, arrival aerodrome Rize-Artvin LTFO, and
alternate aerodrome Trabzon LTCG. All operators are taking into account the identical
factors in the present situation, but with varying advantages for themselves.
The decision-making matrices for operators in BWC are in Tables 2–4. Optimal
solutions are based on the Wald (𝑊), Laplace (𝐿), Hurwitz (𝐻), and Savage (𝑆) criteria.
Table 2
The decision-making matrix for operator O1 (pilot)
Alternative decisions Factors Solutions
{𝐴} a1 a2 a3 a4 a5 a6 W L H, 𝑆
α=0.5
Departure Istanbul (A1) 8 2 9 9 9 8 2 7.5 8.3 7
aerodrome
Arrival Rize (A2) 5 9 8 9 7 8 5 7.7 8.6 4
aerodrome
Alternate Trabzon (A3) 6 7 7 9 5 7 5 6.8 8.6 4
aerodrome
MAX 5 7.7 8.6 4
The optimal landing aerodrome follows the pilot's decision based on the Wald (𝑊)
criterion – Rize (A2) and Trabzon (A3), Laplace (𝐿) criterion – Rize (A2), Hurwitz (𝐻)
criterion – Rize (A2) and Trabzon (A3), Savage (𝑆) criterion – Rize (A2) and Trabzon (A3).
The optimal landing aerodrome follows the ATCO's decision based on the Wald (𝑊)
criterion – Rize (A2) and Trabzon (A3), Laplace (𝐿) criterion – Rize (A2), Hurwitz (𝐻)
criterion – Rize (A2), Savage (𝑆) criterion – Trabzon (A3). The optimal landing aerodrome
follows the expert/AI's decision by all criteria – Rize (A2).
Table 3
The decision-making matrix for operator O2 (ATCO)
Alternative decisions Factors Solutions
{𝐴} b1 b2 b3 b4 b5 b6 W L H, S
α=0.5
Departure Istanbul (A1) 8 2 9 8 9 8 2 7.3 5.5 7
aerodrome
Arrival Rize (A2) 5 9 8 8 7 8 5 7.5 7.0 4
aerodrome
Alternate Trabzon (A3) 6 7 7 8 5 7 5 6.7 6.5 3
aerodrome
MAX 5 7.5 7.0 3
Table 4
The decision-making matrix for operator O3 (expert/AI)
Alternative decisions Factors Solutions
{𝐴} c1 c2 c3 c4 c5 c6 W L H, S
α=0.5
Departure Istanbul (𝐴1 ) 7 2 10 10 9 8 2 7.7 6.0 8
aerodrome
Arrival Rize (𝐴2 ) 6 9 8 10 7 8 6 8.0 8.0 4
aerodrome
Alternate Trabzon (𝐴3 ) 6 7 8 10 5 7 5 7.2 7.5 5
aerodrome
MAX 6 8.0 8.0 4
A matrix of collective decision-making is built to determine the consistency of
operators (Table 4). It contains identical objective factors for operators (pilot (𝑂1 ), ATCO
(𝑂2 ), expert/AI (𝑂3 )), and the solutions of the operators from individual matrices. The
CDM matrix uses the opinions of operators (subjective factors). The optimal CDM for
operators in complicated situation "Unable required position" based on the Wald (𝑊),
Laplace (𝐿), and Hurwitz (𝐻) criteria is presented in Table 5. The optimal CDM is
determined by objective and subjective factors that influence the decisions of all operators
(pilot, ATCO, expert/AI) by Wald (𝑊) criterion – Rize (𝐴2 ) and Trabzon (𝐴3 ), Laplace (𝐿)
and Hurwitz (𝐻) criteria – Rize (𝐴2 ). The optimal landing aerodrome by all criteria – Rize
(𝐴2 ).
Table 5
The CDM matrix for all operators (pilot, ATCO, expert/AI)
Alternative Operators/Solutions
decisions O1 O2 O3 W O1 O2 O3 L O1 O2 O3 H, β = 0.5
{𝐴}
A∗1 2 2 2 2 7.5 7.3 7.7 7.3 8.3 5.5 6.0 5.5
A∗2 5 5 6 5 7.7 7.5 8.0 7.5 8.6 7.0 8.0 7.0
∗
A3 5 5 5 5 6.8 6.7 7.2 6.7 8.6 6.5 7.5 6.5
2.2. Task 2: Decision-making in certainty – vectoring
By the method of subjective-objective CDM based on individual and CDM models the
optimal landing aerodrome is defined – Rize (𝐴2 ). During arrival, it was the message
"Unable required position" (complicated situation), which meant that it was impossible to
perform the RNAV approach. The crew requested VECTORS to VOR ART to perform a
conventional VOR approach from the right seat according to MEL 34-51-01 – VOR1
UNSERVICEABLE. Due to the unstable GPS signal for proceeding to the landing aerodrome,
it is necessary to use the radar vectoring procedure [28]. Simplified analysis of the pilot
and ATCO technological operations during the vectoring is presented in Table 6.
Table 6
Structure-time table of the pilot (𝑝𝑖 ) and ATCO (𝑐𝑗 ) technological operations during the
vectoring – an example
Pilot’s The sequence of Previous ATCO’s The sequence of the Previous
operations, the technological operations operations, technological operations
𝑝𝑖 operations 𝑐𝑗 operations
𝑝1 The message - 𝑐1 Acceptation and 𝑝1
“Unable required confirmation of the
position” information,
reporting to the
supervisor
𝑝2 The decision 𝑝1 𝑐2 Meteorological and 𝑝2 , 𝑐1
about landing technical
aerodrome information about
landing aerodrome
𝑝3 Following the 𝑝2 𝑐3 Vector instructions: 𝑝2 , 𝑐2
ATCO’s heading, altitude,
instructions and speed
restrictions
𝑝4 Landing 𝑝3 𝑐4 The final approach 𝑝3 , 𝑐3
and landing
clearance
According to Table 5, the network graphs of the synchronized technological operations
of the pilot and ATCO during the vectoring are built (Figure 2).
p3 p4
p2
p1
c0
c1
c2
c3 c4
Pilot
ATCO
RNAV
t
Figure 2: The network graphs of the technological operations of the pilot (𝑝𝑖 ) and ATCO
(𝑐𝑗 ) during the vectoring.
Each time the ATCO performs radar guidance by vectors, he must assume
responsibility for all navigation parameters: heading, altitude, and speed (rate of
climb/descent) until the final approach track is intercepted.
The ATCO must not risk the safety of the aircraft when using the radar vectoring
procedure. He must not gamble with the risk of losing echeloning. ATCO must ensure that
the appropriate distance is maintained each time a flight clearance is granted.
3. Expert system for quantitative assessment of collaborative decision-
making by pilot and air traffic controller during pre-simulation
training
Quantitative assessment of the CDM by pilot and ATCO during pre-simulation training was
carried out with the help of an expert system (Figure 3).
Knowledge acquisition Dialog Assessment
component component
Knowledge base Database
Expert information Static and dynamic
CDM models information
Solver
(logical conclusion unit)
Explanatory
component
Figure 3: The scheme of an expert system for quantitative assessment of the CDM by pilot
and ATCO during pre-simulation training.
The database contains information on the flight plan of the aircraft and its changes;
tactical and technical characteristics of the aircraft; geographical, technical, and
meteorological information on the air traffic control area and aerodromes.
The knowledge base contains expert data obtained based on an expert survey of
aviation specialists (values of the parameters of the CDM models) and rules for using this
data (CDM models). The solver (logical inference unit) generates scenarios of training
tasks based on the initial data from the database and knowledge from the knowledge base.
The trainee assessment is performed based on determining the discrepancy between
the standard and actual actions of the trainee.
The significance of trainees' actions for the radar vectoring procedure (setting (ATCO)
and holding (pilot) of navigation parameters: heading, altitude, and speed (rate of
climb/descent)) is obtained by fuzzy logic methods [29] (Figure 4). The use of
membership functions in the context of fuzzy information allows for formalizing
qualitative characteristics.
According to Figure 4, the quantitative indicators of the level of significance of trainees'
actions for the radar vectoring procedure (setting (ATCO) and holding (pilot) of navigation
parameters: heading, altitude, and speed (rate of climb/descent)) are:
• Heading – 70 units.
• Altitude – 40 units.
• Speed (rate of climb/descent) – 10 units.
Assessment of the CDM by the pilot and ATCO during pre-simulation training is a rather
complex and responsible task. For pre-simulation training assessment, the artificial neural
networks are used [22, 30].
1,1
Speed (rate of
climb/descent)
0,9
Altitude
0,7
Heading
0,5
0,3
0,1
-0,1 0 10 20 30 40 50 60 70 80 90 100
Units
Figure 4: The membership functions μ for determining the significance of trainees'
actions for the radar vectoring procedure.
Additive or multiplicative aggregation of the results of pre-simulation training task
performance for comprehensive assessment of the CDM by the pilot and ATCO is
proposed. The results of the assessment of the radar vectoring procedure (setting (ATCO)
and holding (pilot) of navigation parameters: heading, altitude, and speed (rate of
climb/descent)) by additive and multiplicative aggregation are shown in Table 7: 𝑅 – rank;
𝑅𝑎𝑣 – average rank; 𝑤 – weight coefficient; 𝐴, 𝐵 – marks of the trainees; 𝑊𝑎𝑑 – additive
assessment; 𝑊𝑚 – multiplicative assessment.
Two trainees took part in the experiment: trainee 𝐴 who did not miss classes and
trainee 𝐵 who missed classes. Let's compare the results of trainee 𝐵’s comprehensive
assessments obtained by additive and multiplicative aggregation (he received a mark of 0
– completely unable to set/hold the altitude) during pre-simulation training.
Comprehensive assessment of trainee 𝐵’s performance in the training task "Vectoring",
obtained based on the additive aggregation:
𝑊𝑎𝑑 = ∑𝑛𝑖=0 𝑤𝑖 𝑓𝑖 = 0.49 ⋅ 5 + 0.32 ⋅ 0 + 0.19 ⋅ 4 = 2.23.
Table 7
An example of a comprehensive assessment of the radar vectoring procedure: additive and
multiplicative aggregation
Navigation Assessment of the trainees Additive Multiplicative
parameters aggregation aggregation
R R av C w A B Wad (A) Wad (B) Wm (A) Wm (B)
Heading, f1 1 1.21 0.93 0.49 5 5 2.45 2.45 2.20 2.20
Altitude, f2 2 2.16 0.61 0.32 4 0 1.28 0.00 1.56 0.00
Speed (rate), 3 2.89 0.37 0.19 4 4 0.76 0.76 1.30 1.30
f3
Sum/ – – 1.91 1.00 13 7 4.49 3.21 4.46 0.00
Aggregation
Comprehensive assessment of trainee 𝐵’s performance in the training task "Vectoring",
obtained based on the multiplicative aggregation:
𝑊𝑚 = ∏𝑛𝑖=1 𝑓𝑖 𝑤𝑖 = 50.49 ⋅ 00.32 ⋅ 40.19 = 0.
During pre-simulation training of the pilot and ATCO, the multiplicative aggregation of
practical results should be used, as it does not provide for mutual compensation due to
positive (higher) results in other elements of the assessment, as is the case with additive
aggregation.
4. Results and discussions
The illustrative example for pre-simulation training of the pilot and ATCO about flight
Boeing 737-8MAX from departure aerodrome Istanbul Sabiha Gökçen (𝐴1 ) to arrival
aerodrome Rize-Artvin (𝐴2 ) with alternate aerodrome Trabzon (𝐴3 ) is considered.
Trainees are performing two tasks during pre-simulation training:
• Task 1: Decision-making in uncertainty – choosing the optimal landing aerodrome
in BWC.
• Task 2: Decision-making in certainty – vectoring.
The decision-making matrices for choosing the optimal landing aerodrome in BWC by
the pilot, ATCO, and expert/AI with the help of the expert judgment method are built. Six
common objective factors are taken into account by all operators while decision-making:
the weather conditions; the distance to the applicable aerodromes/the quantity of fuel
onboard; the flight and technical characteristics of the aircraft; the approach systems and
navigation aids at the applicable aerodromes; the tactical and technical characteristics of
the runways; Threat and Error Management (TEM). Optimal solutions are based on the
Wald, Laplace, Hurwitz, and Savage criteria. By the subjective-objective CDM method,
using individual and collaborative decision-making models, the optimal landing
aerodrome Rize (𝐴2 ) is determined.
During arrival, it was the message "Unable required position" (complicated situation),
which meant that it was impossible to perform the RNAV approach. Due to the unstable
GPS signal for proceeding to the landing aerodrome, it is necessary to use the radar
vectoring procedure. Structure-time table and network graphs of the synchronized
technological operations of the pilot and ATCO during the vectoring are presented.
The scheme of the expert system for quantitative assessment of the CDM by the pilot
and ATCO during pre-simulation training is designed. The trainee assessment is
performed based on determining the discrepancy between the standard and actual actions
of the trainee. The significance of trainees' actions for the radar vectoring procedure
(setting (ATCO) and holding (pilot) of navigation parameters: heading, altitude, and speed
(rate of climb/descent)) is obtained by fuzzy logic methods. The quantitative indicators
are:
• Heading – 70 units.
• Altitude – 40 units.
• Speed (rate of climb/descent) – 10 units.
An experiment was conducted with two trainees: trainee 𝐴 who did not miss classes
and trainee 𝐵 who missed classes. The results of trainee comprehensive assessments
during pre-simulation training obtained by additive and multiplicative aggregation are
compared. The multiplicative aggregation of practical results more preferable because it
does not provide for mutual compensation due to positive (higher) results in other
elements of the assessment, as is the case with additive aggregation.
5. Conclusions
Joint simulation training of aviation specialists is a critical stage of professional education
and plays a significant role in further ensuring flight safety. At the briefing stage, the
instructor checks the readiness of trainees for practical activities, including CDM between
the aircraft crew and ATCO in complicated, complex, emergency situations. GPS signals are
currently subject to instability. In the case of interference with GPS navigation equipment
onboard an aircraft flying according to the RNAV rules, the pilot and ATCO must be able to
collaborate effectively in the radar vectoring procedure. 28.3% of the problems in
obtaining navigation information from GPS are related to the descent and landing stages.
The authors proposed to use as the elements of testing at the stage of pre-simulation
training the collaborative performance of the training tasks "Choosing the optimal landing
aerodrome" and "Vectoring".
Based on the expert judgment method, models of individual and CDM on choosing the
optimal landing aerodrome in BWC by the pilot, ATCO, and expert/AI are developed. The
optimal solutions are calculated using the Wald, Laplace, Hurwitz, and Savage criteria
(training task 1). By the network planning method based on structure-time table and
network graphs, the synchronized vectoring operations of the pilot and ATCO in
complicated situation "Unable required position" are worked out (training task 2).
The significance of trainees' actions when performing the training task "Vectoring"
(setting (ATCO) and holding (pilot) of navigation parameters: heading, altitude, and speed
(rate of climb/descent)) for quantitative assessment in the expert system is obtained by
fuzzy logic methods.
A comparison of the comprehensive assessment of trainees' performance in the
training task "Vectoring" by additive and multiplicative aggregation is considered, and the
expediency of using multiplicative aggregation of practical results is proved.
The expert system for multiplicative assessment of the results of pre-simulation
training will allow for making an objective conclusion about the mastery of the necessary
knowledge, skills, and abilities by trainees and the acquisition of professional practice-
oriented competencies in CDM.
6. Future scope
Further researches are aimed at developing an Intelligent System for Collaborative
Decision-Making while Training (IS CDM-T) for joint education of aviation specialists
(pilots, operators of drones, flight dispatchers, ATCOs, handling agents, maintenance
personnel, etc.) using big data analysis and machine learning. To control AI solutions,
aviation professionals need to use hybrid intelligent systems that integrate human and
machine capabilities.
References
[1] Future of the airline industry 2035, International Air Transport Association, Montreal,
Canada, 2018. URL: https://www.iata.org/contentassets/086e8361b2f4423e
88166845afdd2f03/iata-future-airline-industry.pdf.
[2] Global Air Navigation Plan 2016-2030, Doc. 9750, 5th ed., ICAO, Montreal, Canada,
2016. URL: https://www.iata.org/contentassets/1be2bec28b3d45f9ae7780d6ebea7
be9/icao_ganp_doc209750_5ed_en.pdf.
[3] J. Benson, A. Andreeva, S. Hansen, T. K. Wolynski, Overall investment outlook for
global aviation finance, White & Case, 14 February 2022. URL:
https://www.whitecase.com/insight-our-thinking/overall-investment-outlook-
global-aviation-finance-0.
[4] ICAO Study Reveals Strong Demand for Qualified Aviation Personnel up to 2030,
International Civil Aviation Organization, 8 March 2016. URL:
https://www.icao.int/Newsroom/Pages/icao-study-reveals-strong-demand-for-
qualified-aviation-personnel-up-to-2030.aspx.
[5] O. Gill, Europe needs 6,000 new pilots a year for next two decades, warns Boeing, The
Telegraph, 27 July 2022. URL: https://www.telegraph.co.uk/business/2022/
07/27/europe-needs-6000-new-pilots-year-next-two-decades-warns-boeing/.
[6] About NGAP, International Civil Aviation Organization, 2017. URL:
https://www.icao.int/safety/ngap/pages/ngapinitiatives2.aspx.
[7] Specification for the ATCO Common Core Content Initial Training, 2.0 ed.,
EUROCONTROL, Brussels, Belgium, 2015. URL:
https://www.eurocontrol.int/publication/eurocontrol-specifications-atco-common-
core-content-initial-training.
[8] Global Performance of the Air Navigation System (Doc. 9883), in: ICAO Workshop on
the Application and Development of the Vol. III of the CAR, International Civil Aviation
Organization, Virtual Meeting, 2020. URL: https://www.icao.int/SAM/Documents/
2020-RLA06901-ANPVOLIII/1_2_Doc.%209883%20GPM.pdf.
[9] А. Damiba, Changes to the ICAO Doc. 9613, in: Performance-based Navigation (PBN)
Route Laboratory Workshop, Nairobi, 22-26 May 2023, International Civil Aviation
Organization. URL: https://www.icao.int/ESAF/Documents/meetings/2023/
PBN%20ROUTE%20LAB%2022-26%20May%202023/Presentations/prl_changes%
20to%20icao%20doc9613.pdf.
[10] V. Konin, O. Pogurelskiy, I. Prykhodko, T. Maliutenko, O. Sushych, O. Ishchenko, Check
for updates multi-GNSS in limited navigation satellite, Lecture Notes in Networks and
Systems (LNNS), vol. 736, 2023, pp. 126–140. doi: 10.1007/978-3-031-38082-2_22.
[11] Problems with air navigation in the Baltic Sea, Radio Liberty, International News, 18
March 2024. URL: https://www.radiosvoboda.org/a/news-baltyka-gps-sygnal-
problemy/32865919.html.
[12] Procedures for Air Navigation Services-Air Traffic Management (PANS-ATM), Doc.
4444, 16th ed., International Civil Aviation Organization, Montreal, Canada, 2016.
[13] L. Ren, M. Castillo-Effen, Air Traffic Management (ATM) operations: A review, Report
number 017GRC0222, GE Global Research, Niskayuna, New York, United States, 2017.
[14] Manual on Collaborative Air Traffic Flow Management (ATFM), Doc. 9971, 3rd ed.,
International Civil Aviation Organization, Montreal, Canada, 2018.
[15] Pilot training, Lufthansa aviation training, 2024. URL: https://www.lufthansa-
aviation-training.com/en/pilot-training.
[16] Improving pilot-air traffic control collaboration, Hong Kong International Aviation
Academy, 09 May 2024. URL: https://www.hkiaacademy.com/en/air-traffic-
management/professional-
courses/atcs1971_improving_pilot_air_traffic_control_collaboration.html.
[17] S. Corrigan, G.D.R. Zon, A. Maij, N. Mcdonald, L. Mårtensson, An approach to
collaborative learning and the serious game development, Cognition, Technology &
Work 17(2) (2015) 269–278. doi:10.1007/s10111-014-0289-8.
[18] R. Lutte, R.W. Mills, Collaborating to train the next generation of pilots: Exploring
partnerships between higher education and the airline industry, Industry and Higher
Education 33(6) (2019) 448–458. doi:10.1177/0950422219876.
[19] T. Shmelova, Yu. Sikirda, N. Rizun, A.-B. M. Salem, Yu. Kovalyov, Socio-technical
decision support in air navigation system: Emerging research and opportunities, IGI
Global Publ., Hershey, USA, 2018. doi: 10.4018/978-1-5225-3108-1.
[20] Yu. Sikirda, T. Shmelova, V. Kharchenko, М. Kasatkin, Intelligent system for supporting
collaborative decision making by the pilot/air traffic controller in flight emergencies,
CEUR Workshop Proceedings 2853 (2021) 127–141. URL: http://ceur-ws.org/Vol-
2853.
[21] Т. Shmelova, Yu. Sikirda, М. Yatsko, М. Kasatkin, Collective models of the aviation
human-operators in emergency for Intelligent Decision Support System, CEUR
Workshop Proceedings 3156 (2022) 160–174. URL: http://ceur-ws.org/Vol-
3156/paper10.pdf.
[22] T. Shmelova, Yu. Sikirda, T. Jafarzade, Artificial neural network for pre-simulation
training of air traffic controller: Chapter 2, in: T. Shmelova, Yu. Sikirda, N. Rizun, D.
Kucherov (Eds.), Cases on modern computer systems in aviation, USA, Hershey, IGI
Global, 2019, pp. 27–51. doi: 10.4018/978-1-5225-7588-7.ch002.
[23] T. Shmelova, Yu. Sikirda, N. Rizun, V. Lazorenko, V. Kharchenko, Machine learning and
text analysis in an artificial intelligent system for the training of air traffic controllers:
Chapter 1, in: T. Shmelova, Yu. Sikirda, N. Rizun, D. Kucherov, K. Dergachov (Eds.),
Automated systems in the aviation and aerospace industries, USA, Hershey, IGI Global,
2019, pp. 1–50. doi: 10.4018/978-1-5225-7709-6.ch001.
[24] T. Shmelova, Yu. Sikirda, М. Yatsko, V. Kolotusha, Intelligent integrated training
system for the aviation specialists "Collaborative Decision-Making – Education"
(CDM-E), CEUR Workshop Proceedings 3538 (2023) 168–180. URL: https://ceur-
ws.org/Vol-3538/Paper_16.pdf.
[25] FAA warning issued, further serious navigation failures reported, OPSGROUP, 28
September 2023. URL: https://ops.group/blog/faa-warning-navigation-failures.
[26] J. Franclin, Global Navigation Satellite System Outage, EASA Community Network, 06
November 2023. URL: https://www.easa.europa.eu/community/topics/global-
navigation-satellite-system-outage.
[27] GNSS/GPS Interference, Reported in MENA Region 2022, Global Aviation Data
Management, 06 March 2023. URL: https://www.icao.int/MID/Documents/2023/
MIDAMC% 20and%20CNS/CNS%20SG12-PPT11.pdf.
[28] Radar vectoring procedure and method, IVAO Documentation Library, 2024. URL:
https://wiki.ivao.aero/en/home/training/documentation/Radar_vectoring_procedur
e_and_method.
[29] S. Broumi (Ed.), Handbook of research on advances and applications of fuzzy sets and
logic, IGI Global, Hershey, USA, 2022. doi: 10.4018/978-1-7998-7979-4.
[30] S. Kaddoura (Ed.), Handbook of research on AI methods and applications in computer
engineering, IGI Global, Hershey, USA, 2023. doi: 10.4018/978-1-6684-6937-8.