=Paper=
{{Paper
|id=Vol-2387/20190454
|storemode=property
|title=None
|pdfUrl=https://ceur-ws.org/Vol-2387/20190454.pdf
|volume=Vol-2387
|dblpUrl=https://dblp.org/rec/conf/icteri/ShmelovaSK19
}}
==None==
https://ceur-ws.org/Vol-2387/20190454.pdf
Applied Artificial Intelligence for Air Navigation
Sociotechnical System Development
Tetiana Shmelova1[0000-0002-9737-6906], Yuliya Sikirda2[0000-0002-7303-0441],
Mykola Kasatkin3[0000-0002-2501-1756]
1
National Aviation University, Komarova av., 1, 03058, Kiev, Ukraine
shmelova@ukr.net
2Flight Academy of National Aviation University,
Dobrovolskogo Street, 1, 25005, Kropyvnytskyi, Ukraine
sikirdayuliya@ukr.net
3Kharkiv National University of Air Forces named by I. Kozhedub,
Sumska Street, 77/79, 61023, Kharkiv, Ukraine
kasatik_79@ukr.net
Abstract. Nowadays the evolution of Human Factor's models has included im-
plementation of Artificial Intelligence (AI) in aviation as an innovative technol-
ogy for enhancing security and the characteristic ability to learn, improve, and
predict. The AI is presented in models of decision making in Air Navigation
Sociotechnical system as Expert Systems, Decision Support Systems for pilots
of manned and unmanned aircraft, for air traffic controllers in the emergencies.
Keywords: Air Traffic Controller, Collaborative Decision Making, Expert Sys-
tem, Human-Operator, Unmanned Aerial Vehicle.
1 Introduction
Nowadays in documents of International Civil Aviation Organization (ICAO) defined
new added approaches for achieving the main goal of ICAO enhancing the effective-
ness of global aviation security, and improving the practical and sustainable imple-
mentation of preventive aviation security measure. The Global Aviation Security Plan
(GASP) identifies five key outcomes for improving effectiveness, such as [1]:
enhancing awareness and response of risk;
development of security culture and human capability;
improving technological resources and foster innovation;
improving oversight and quality assurance;
increasing cooperation and support between states.
So, the quality of decisions dependences from the development and using of inno-
vative technology in aviation nowadays such as Artificial Intelligence (AI) [2]. De-
veloping of AI in Air Navigation System (ANS) as Sociotechnical system (STS) such
as Expert Systems (ES), Decision Support Systems (DSS), are considering new con-
cepts in aviation need with using modern information technologies and modern cours-
es: Data Science, Big Data, Data Mining, Multi-Criteria Decision Analysis, Collabo-
ration Decision Making (CDM), Blockchain, etc.
The purposes of the work are: analysis of the benefits of using AI models in the
Air Navigation Sociotechnical System (ANSTS); AI models and methods; problems
of CDM by ANS personnel (human-operators (H-O): pilots of manned and unmanned
aerial vehicles (UAV), air traffic controllers (ATC), engineers).
2 Artificial Intelligence Systems in Air Navigation
Sociotechnical System
Today, AI capabilities are proliferating across the transport sector. The AI systems
have high potential in Air Traffic Management (ATM), specifically in areas which
involve decision making (DM) under uncertainty (e.g. conflict detection and resolu-
tion) and prediction with limited information (e.g. trajectory prediction) [3; 4]. For
example, validation based on one month of ADS-B data, the AI system is able to pre-
dict ATC actions, for complex traffic scenarios, at an accuracy – AN-Conf/13-
WP/232 of above 70%. In addition to developing solutions to support en-route opera-
tions, AI can be applied in speech recognition to act as an additional safety net to
detect read-back errors; trajectory synchronization of aircraft ground movements that
provide optimised taxiing strategies that comprehensively accounts for arrivals and
departures as well; and predicting the most optimal runway configuration for a given
arrival sequence and departure schedule so as to maximize the runway throughput [2].
It is very important to create highly intelligent joint DM systems for engineers, pilots,
air traffic controllers and new airspace users, such as unmanned aircraft systems. As a
rule, systems are significant for work and for personal training.
Therefore, it is necessary to present ANS as STS and to applicate AI methods for
development capacity of ANSTS. Examples of applied AI in education course “In-
formatics of Decision Making” for aviation students and future personal of ANS in
Table 1. AI is the simulation of human intelligence processes by modeling, computer
systems, and machines. These processes include: learning (the acquisition of infor-
mation and rules for using the information); reasoning, estimation, and modeling (us-
ing rules to reach conclusions (approximate or definite results)); self-correction (esti-
mation of obtained models); particular applications of AI include ES; DSS; automated
systems; systems of pattern recognition, speech recognition, and machine vision, etc.
Many cases of using AI technology are driven by emergence, availability, and acces-
sibility. The differentiating factor of an AI system from a standard software system is
the characteristic ability to learn, improve, and predict. Through training, an AI sys-
tem is able to generate knowledge and apply it to novel situations not encountered
before. In AI, an ES is a computer system that simulates the decision making ability
of a human. The ESs are designed to solve complex problems by reasoning through
bodies of knowledge, represented mainly as if-then rules rather than through conven-
tional procedural code. The first ESs were created in the 1970s and then proliferated
in the 1980s. Expert systems were among the first truly successful forms of AI soft-
ware. The ICAO documents recommend developing Intelligent Expert Systems in
aviation to support DM of operators [1; 2]. For the education of aviation personals in
AI developed training, such as:
Expert Judgment Method (EJM) / Multi-criteria decision problems.
Deterministic models. Network planning. Decision making in an emergency.
Stochastic models. Decision making in Risk. Decision making in an emer-
gency.
Game Theory. DM in uncertainty. Optimal aerodrome of landing.
Dynamic programming (DP) and GRID analyzes of the problem. The DP
method to solve the problem of minimal cost, climbing an aircraft.
Table 1. Applied AI in education course “Informatics of DM”.
Content Pilot ATC Engineer UAV operator
Aviation man- Analysis and synthesis of aviation using the theory of automatic control
machine system MMS “pilot – MMS MMS MMS “Operator of
(MMS) aircraft” Analysis “ATC- UAV”
and synthesis of aircraft”
aviation using
theory of automat-
ic control
Aviation expert sys- Expert Judgment Method / Multi-criteria decision problems
tem of quantitative Significance (com- Controller’s Significance of the Significance of the
estimation plexity) of the workload for Landing System UAVs, phases of
phases of flight of aircraft (GNSS, ILS, VOR, flight of the UAVs
the aircraft service VOR/DME)
Decision Support Models of DSS: deterministic and stochastic models of DM
System of H-O in DM of a pilot in DM of ATC DM of an engineer DM of unmanned
ANS the emergency in emergen- in service / emer- pilot
cy gency
Decision making in Network planning of action of H-O in service/ emergency
certainty Graph of proce- Graph of Graph of proce- Graph of proce-
dures of the pilot procedures dures in the service dures of an un-
in the emergency of ATC in of equipment manned pilot in an
emergency emergency
Design making in risk Design tree of forecasting of action of H-O in emergency
DM in emergency DM in DM in service of DM in emergency
emergency equipment
Decision making in Criteria Wald. Laplace, Hurwitz, Savage optimal DM in uncertainty
uncertainty Optimal landing Optimal Optimal action in Optimal landing
aerodrome in landing emergency aerodrome/place in
emergency aerodrome emergency
in emergen-
cy
Neural Networks Forecasting of outcomes in emergencies
Markov Networks Neural network admission student to simulator training
GERT-models Development and forecasting the emergency situation / Preventing cata-
strophic situation
Fuzzy logic Quantitative estimation of the outcomes /risk in emergencies
ANSTS Analysis of ANS as STS and diagnostics, monitoring of the factors (profes-
sional and non-professional) that influence on DM by the H-O in STS (indi-
vidual-psychological, socio-psychological and psychophysiological factors)
For example, training 1 “Expert Judgment Method (EJM) / Multi-criteria decision
problems” [5; 6]:
Theory. Basic of EJM for ANS (Classification of methods of DM. The algorithm
of EJM. The matrix of individual and group preferences. Coordination of experts’
opinion. Multi-criteria decision problems).
Practice. Tasks: Quantitative estimation of the complexity of the aircraft flight.
Definition of significance (complexity) of the phases of flight of the aircraft (Fig. 1).
Fig. 1. The process of applying solutions to many similar tasks in the quantitative estimation of
the complexity of the aircraft’s flight stages.
To show the DSSs we use tasks of alternate aerodrome/place in the case of an emer-
gency landing (difficult meteorological conditions, etc.) by the method of DM in un-
certainty is means of the criteria of DM in uncertainty: Wald, Laplace, Savage, Hur-
wicz [7]. For example, the results matrix of DM for choosing of landing aero-
drome/place of UAV for route of UAV flight from Bila Tserkva aerodrome to
Konotop with possible alternate destinations at Vasylkiv, Berezan’ Nizhyn and
Pryluky (Table 2). Input data are: λ1 – is an availability of fuel/energy onboard of
UAV; λ2 – is a distance from UAV to ADest, ADep, AAP; λ3 – are the tactical and tech-
nical characteristics of the runways of ADep, ADest, AAP; λ4 – are the meteorological
conditions at ADep, ADest, AAP; λ5 – is a reliability of C2 lines for connection with
UAV; λ6 – is a possibility of communication with ATC units; λ7 – are the navigational
aids at ADep, ADest, AAP; λ8 – is a possibility of communication with ATC units; λ7 –
are the lighting systems at ADep, ADest, AAP; etc. Formation of possible results in ma-
trix determined with the EJM by rating scales according to the regulations.
Table 2. The results matrix of DM for choosing of landing aerodrome/place of UAV.
Alternative decisions Factors that influence DM Solutions
Аi ААAs λ1 λ2 λ3 λ4 λ5 λ6 λ7 W L H S
А1 Bila Tserkva 9 2 5 8 0 3 9 0 5,14 4,5 9
А2 Konotop 3 5 7 9 2 4 9 2 5,57 4,5 7
А3 Vasylkiv 2 8 8 9 2 4 10 2 6,14 6 8
А4 Berezan’ 7 1 8 7 1 7 7 1 5,43 5 7
А5 Nizhyn 6 4 8 6 6 5 8 4 6,14 4,5 4
А6 Pryluky 4 8 9 8 4 6 6 4 6,43 6,5 5
For the last years, the authors have developed computer programs for DSS of the
aircraft pilot, air traffic controllers, flight dispatcher, UAV’s operator, etc. In Masters
Diploma “Remote expert air traffic management system“ decision making in a com-
mon environment FF-ICE presented decentralized-distributed UAVs control system
using blockchain technology (Fig. 2). There are many advantages of this approach
such as enhanced security, security, big data analysis, and record keeping, real-time
constant data exchange, etc.
Fig. 2. Decentralized-distributed UAV’s control system
For today, the key to ensuring the safety of flights is the problem of the organization
of Collaborative Decision Making (CDM) by all the operational partners – airports,
air traffic control services, airlines and ground operators – on the basis of general
information on the flight process and ground handling of the aircraft in the airport [8].
Blockchain technology is ideal as a new infrastructure to secure, share, and verify
learning achievements and CDM too.
The problem of optimizing the interaction between pilot and ATC can be solved by
the way of development and synchronization (maximal alignment over time) of de-
terministic models of H-O CDM, which will minimize the critical time needed to
solve EC, by definition the optimal sequence of execution of technological proce-
dures. The parallel process of simultaneous execution of pilot and ATC technological
operations in the emergency case can be represented as a consolidated dual-channel
network. For a consistent optimization of such a network in order to achieve the
cross-cutting efficacy of joint decisions, it is advisable to use a multi-criteria ap-
proach: achieving a minimum time for parity of emergency case with maximum safe-
ty / maximum harmonization over the time of H-O actions.
In this context, the use of flight simulators during ATC professional training is rel-
evant. They will help ATC’s to get acquainted with the situation in the flight crew
cabin and the parameters of the aircraft's devices during the emergency case. At the
same time the ATC: will receive the experience of the crew members during the
emergency case; will pay attention to how the intervention of the dispatcher can dis-
rupt crew members; will complete exercises on the use of radio during the emergency
case; will complete the checklist in the emergency case; will participate in captain
decision making during the emergency case; will observe the features of the go-
around procedure. In the emergency case, ATC is advised to use a checklist that will
help to handle incidents in order to establish optimal actions to achieve better cooper-
ation between pilot and ATC. A supervisor who works with ATC, using a checklist,
can provide better support as it will more clearly understand the traffic control in
emergencies.
3 Conclusion
The AI technologies in aviation were clustered in the following seven capabilities:
Machine Learning (ML); Natural Language Processing (NLP); Expert Systems; Vision
and Speech; Planning; Robotics.
Further research should be directed to the solution of the problem in prerequisites
of emergency situations and preventing catastrophic situations too. Models of flight
emergency development and of DM by an operator in-flight emergency will allow
predicting the operator’s actions with the aid of the informational-analytic and diag-
nostics complex for research of operator behavior in extreme situations. It is neces-
sary to develop modern DSSs of Air Navigation System’s operator (pilots, air traffic
controllers, flight dispatchers, UAV’s operators) in-flight emergencies and in other
situations, to investigate applied tasks of the DM in Sociotechnical System by an
operator of aviation system, chemical production, energy, military industry, etc.
Developing of Intelligent ES, DSSs considering new concepts in aviation (FF-ICE,
PBA, SMART, CDM, SWIM, etc.) for different operators and each stage, process,
which are problems, with using modern information technologies Data Science, Big
Data, Data Mining, Multi-Criteria Decision Analysis, etc. It is necessary to analyze all
factors influencing the DM of operators in these systems in order to predict the devel-
opment of the technogenic catastrophe and prevent it with AI.
References
1. Global Aviation Security Plan (GASP). International Civil Aviation Organization, Canada,
Montreal (2017)
2. Potential of Artificial Intelligence (AI) in Air Traffic Management (ATM). In: Thirteenth
Air Navigation Conference ICAO, Montréal, Canada, 9-19 October 2018 (2018)
3. International Civil Aviation Organization Global Air Traffic Management Operational
Concept: Doc. ICAO 9854 (1st ed.). International Civil Aviation Organization, Montreal,
Canada (2005)
4. Air Traffic Management: Doc. ICAO 4444-RAC/5011 (5th еd.). International Civil Avia-
tion Organization, Montreal, Canada (2007)
5. Shmelova, T., Sikirda, Yu., Rizun, N., Lazorenko, V., Kharchenko, V.: 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:
manuscript. USA, Hershey, IGI Global, 1-50 (2019)
6. Kharchenko, V., Shmelova, T., Sikirda, Y.: Methodology of Research and Training in Air
Navigation Socio-technical System. Proceedings of the National Aviation University, vol.
1, 8–23 (2018)
7. Shmelova, T., Sikirda, Y., Rizun, N., Salem, A.-B. M., Kovalyov, Y.: Socio-Technical
Decision Support in Air Navigation Systems: Emerging Research and Opportunities: man-
uscript. USA, Hershey, IGI Global (2018)
8. Manual on Collaborative Decision-Making (CDM): Doc. ICAO 9971 (2nd ed.). Interna-
tional Civil Aviation Organization, Montreal, Canada (2014)