Adv Syst Sci Appl 2022; 04:19–30
Published online at https://ijassa.ipu.ru.

Spacecraft Emergency Situations Automated Scenarios Design
within the Framework of FDIR Concept and Dynamical Neural

Networks Technique

Alexey Tuturov1, Olga Andrianova1,2*, Ernest Sleptsov, Evgeniy Yurkevich1, Lidia Kryukova1

1V. A. Trapeznikov Institute of Control Sciences of Russian Academy of Sciences, Moscow, Russia
2Tecnologico de Monterrey-Campus Guadalajara, Research Institute of Advanced Materials for Sustainable

Manufacturing, Guadalajara, Mexico

Abstract: The space complex is considered as a Socio-CyberPhysical System (SCPS) in
order to point out that the human factor is one of the most important ones in the analysis
of emergency situations using spacecraft telemetry in the conditions of the current level of
software development, information and technical resources of the space complex to ensure its
functional stability. Functional stability of the system means its functionality under the influence
of destructive factors of the external or internal environment. The concept of Fault Detection,
Isolation, and Recovery (FDIR) is described, its advantages and drawbacks are pointed out in the
context of the current problem. Based on the FDIR concept, decision support systems will be
developed in future. In the framework of FDIR, spacecraft telemetry data is proposed to be used
to design the monitor, the task of which is to identify and register emergency situations, based on
modern data processing tools, namely, Dynamical Neural Networks (DNNs).

Keywords: FDIR concept, dynamical neural networks, telemetry, socio-cyberphysical system

1. INTRODUCTION

A space complex may be considered as a Socio-CyberPhysical System (SCPS) due to the
increased influence of the human factor on the functional stability of the system in the
context of modern trends in increasing software and hardware implementations in space
technology. Functional stability is understood as the ability to preserve and/or restore the
performance of system functions, as well as to adapt information communication channels
when changing system elements under the influence of destructive factors of the external
and/or internal environment (Korolev, 2018). Thus, along with technical, informational, and
software processes and resources, a set of technical experts can be represented as human
resources, taking into account the importance of their daily participation in ensuring the
functional stability of the entire system. Increasing of personnel interaction efficiency with
software and hardware is carried out, among other things, by developing decision support
systems. In this paper, a new concept for operability diagnostics of a spacecraft using
telemetry data from its service and target systems is proposed and presented. Based on this
methodology, decision support systems are planned to be developed in future.

One of the modern methods of ensuring functional stability is the application of the
concept of Fault Detection, Isolation, and Recovery (FDIR). The main tasks of FDIR are
the automatic switch of the spacecraft to a safe mode of operation, logging, and localization

∗Corresponding author: andrianovaog@gmail.com



20 A. TUTUROV, O. ANDRIANOVA, E. SLEPTSOV, E. YURKEVICH, L. KRYUKOVA

of emergency situations. This methodology makes it possible to increase the reliability of the
target application without using excess reserves, therefore it is widely used in the rocket and
space industry (Akhmetov, Makarov, & Sollogub, 2014; Eickhoff, 2012).

To construct and implement the mentioned concept above during the system design and
after the start of the flight mission, it is necessary to classify the nature of possible faults. To do
this, methods of probability theory, reliability theory, and expert evaluation are currently used
in engineering practice. Another way is to process the telemetry information coming from the
spacecraft, to analyse and classify it, using classical and modern concepts of data processing:
the method of support vectors (Azevedo, Ambrosio, & Vieira, 2012), self-learning Kohonen
maps (Talalaev & Fralenko, 2018), artificial neural networks, one of the extensions of which
are Dynamical Neural Networks (DNNs) (A. S. Poznyak, Oria, & Poznyak, 2019), which will
be used further in the current research.

DNNs have the aim of approximating uncertain systems represented by differential
or difference equations or measuring (telemetry) information. In case of dynamical
representations, both unknown parameters in the right-hand side of the system or its unknown
form may be recovered, meanwhile, if the system is represented by time series, omissions
or errors in the input sequence can be restored due to the flexible structure of DNNs.
The main difference between dynamical neural networks and well-known data processing
methods is time-dependence of their coefficients (weighting matrices), which allows them to
be used in approximation problems for dynamical plants. DNNs have proven to be efficient
approximate representations of different classes of uncertain models (A. Poznyak, Sanchez,
& Yu, 2001; Wu, 1998; T. Poznyak, Chairez, & Poznyak, 2019). They can also use their
internal memory to process arbitrary sequences of inputs. When analyzing time series, this
property is important not only to recover the missing data, but also to detect non-standard
behavior, that can lead to faults and emergency situations.

The following main contributions of this study are:

1. The space complex is considered as a SCPS and the concept of its functional stability is
introduced and explained;

2. The telemetry data of the spacecraft is analyzed from the point of view of detecting and
classifying anomalies;

3. In the FDIR concept, DNNs are proposed to be used as a monitor on telemetry data.

The rest of the paper is organised as follows. In Section 2, components of the space
complex are described as well as a SCPS and its functional stability are defined. Section 3
analyzes the spacecraft telemetry data, introducing two types of anomalies and emergency
engine start, that can occur the system. In Section 4, basics of FDIR concept are given
and its modification for the current problem is discussed. Section 5 introduces theoretical
background for DNN identification, using telemetry data. Section 6 presents some concluding
remarks and ideas for future work.

2. THE SPACE COMPLEX AS A SOCIO-CYBERPHYSICAL SYSTEM

This section describes the main components of the space complex and their functions, as
well as definitions of functional stability and telemetry. The necessity of transition from a
cyberphysical system to a socio-cyberphysical one is justified when describing the space
complex and ensuring its functioning.

The structure, characteristics, and operational standards of the equipment for a particular
spacecraft are directly related to its purpose as part of the Space Complex (SC) defined as
a set of interconnected orbital and ground-based technical devices that provide solutions to
various tasks based on the use of outer space.

Copyright © 2022 ASSA. Adv Syst Sci Appl (2022)



SPACECRAFT EMERGENCY SITUATIONS AUTOMATED SCENARIOS DESIGN 21

Out of all the variety of SC components, only the following will be used in the current
research:

• Satellite Constellation (SCon) as a set of spacecrafts located on the orbits in accordance
with the ballistic structure and united by the commonality of the problems to be solved
as part of the SC;

• Ground Control Complex (GCC) as a set of technical devices and structures designed
to control the functioning of the spacecraft of SCon from the moment of launching to
orbit;

• Mission Control Center (MCC) as structures with technical systems and technological
tools of command-software, telemetry, and ballistic-navigation support, external
information exchanges, trunk and special communications, mapping, designed to ensure
the activities of service personnel for the formation, transmission, reception, processing,
storage, documentation of information during the continuous process of spacecraft flight
control;

• Payload Data Management Center (PDMC) as a set of interconnected ground-based
technical devices with software designed to provide the customer and its consumers
with target information obtained on the basis of space data (Piro et al., 2018).

Due to the concept of telemetry, the space complex may be represented and described
as a single system. Telemetry is defined as a field of science and technology that deals
with the development and operation of automated devices complex that ensures the
receipt, conversion, transmission via a communication channel, reception, processing, and
registration of measuring information and information about events in order to monitor at a
distance the state and functioning of technical and biological systems of various objects and
the study of natural phenomena. Thus, there is necessarily a telemetry system as a part of
any spacecraft, considered as an automated system. A typical model of a telemetry system
functionating is shown in Figure 2.1.

Fig. 2.1. Hierarchical model of a telemetry system using CCSDS (the Consultative Committee for Space Data
Systems) recommendations

Copyright © 2022 ASSA. Adv Syst Sci Appl (2022)



22 A. TUTUROV, O. ANDRIANOVA, E. SLEPTSOV, E. YURKEVICH, L. KRYUKOVA

Functional stability is understood as a necessary condition for ensuring the fulfillment
of the SC and its components objectives. For correct operation of the concept of functional
stability, the SC should be considered as a complex object, represented by a CyberPhysical
System (CPS), meaning an intelligent system based on engineering interacting networks of
physical and computational components.

Taking into account the interaction of the SCon ballistic structure features and the
geography of Ground-based Communications (GC), the unity of the information space for a
particular spacecraft and the entire SC is possible only within a time-limited communication
session over a physical connection via a radio signal, by establishing a communication
channel. Usually, the communication channel between the spacecraft and the GC is
considered as the basic level when presenting the SC as a CPS. With this approach, all
resources and processes in the CPS belong to one of the categories: information, software
or technical.

However, in this paper, this functionality is proposed to be extended by adding the
category “personnel”, designed to reflect the knowledge, experience, intuition, and ability
to improvise of the personnel teams of the GCC, MCC, and PDMC in the form of expert
assessments described by weight functions. Thus, in this paper it is proposed to consider
a SC as a SCPS. The idea of considering a human (social) factor in automatic control
systems is not new, associated with increasingly complex control and navigation tasks.
For example, a concept of control subject has been introduced by V. Lefebvre (Lefebvre,
1977) and developed by V.E. Lepskiy (Lepskiy, 1998), D.A. Novikov (D.A. & E., 2012).
According to this new general statement of control problems, control subject and control
object should always be considered as integral parts. In (Sleptsov & Andrianova, 2021),
the control subject’s goals are described as follows: “Control subjects task is to ensure the
required state of the system from its point of view, taking into account information about
external disturbances. Naturally, the control system functioning this way is closed-loop as
control object and control subject form a closed cycle.”

Consideration of SC as a SCPS will make it possible to introduce into the control process
a human factor not as a statistical disturbance, but as one of the vectors determining the
development of SC. To solve this problem, methods of scenario tree construction are used to
develop emergency situation scenarios. However, this method requires a sufficient amount of
accumulated event statistics, i.e. recorded facts of qualitative or quantitative changes in the
parameters of the internal and/or external environment of the spacecraft, which is directly
related to the level of wear of the spacecraft systems components and contradicts the trend
of high functional stability of the spacecraft for short periods of its operation. Thus, in this
paper, the scenario tree method will not be directly applied.

3. SPACECRAFT TELEMETRY AS AN INFORMATION ENVIRONMENT FOR
CONSTRUCTING THE SCENARIOS

Consider the communication channel between the spacecraft and the GCC as the only way
for a single information space existence between the spacecraft and the SC. Characteristics
of the communication channel under this consideration act as boundary conditions, and the
main limitation of information exchange is the duration of the communication channel stable
existence, mainly due to the mutual radio visibility of the spacecraft and the GSS.

Figure 3.2 shows the telemetry data of an element of the Spacecraft Orientation and
Stabilization (SOS) system for 14 days: from 305 to 319. The SOS data are selected because
the functional stability of the entire spacecraft directly depends on the operation of this
system, as well as the quality of its target tasks: in this example, the telemetry information
from the Earth remote sensing satellite is used.

The color of plots in Figure 3.2 shows the change of day in the considered interval. From
the color change, it can be seen that the oscillation of the plots is cyclical: the data of the first
days differ qualitatively from the last by the presence of a shift equal to about 17 minutes,

Copyright © 2022 ASSA. Adv Syst Sci Appl (2022)



SPACECRAFT EMERGENCY SITUATIONS AUTOMATED SCENARIOS DESIGN 23

Fig. 3.2. Telemetry data from the element of SOS system

Fig. 3.3. Telemetry data from the magnetometer: X axis

which corresponds to the natural precession of the orbit for this spacecraft. This pattern can
also be traced for other elements. Thus, Figures 3.3, 3.4, and 3.5 show telemetry data from
one magnetometer along three axes, but the tendency to cyclicity persists here, despite the
difference in the nature of the recorded characteristic.

In each of Figures 3.2–3.5, a characteristic area of data dominance is clearly visible,
and it is not difficult to find mathematical functions to describe its boundaries. Denote this
statistical area by the “norm” for the time period of observations under consideration. Two
types of anomalies in relation to the “norm” can be distinguished.

Copyright © 2022 ASSA. Adv Syst Sci Appl (2022)



24 A. TUTUROV, O. ANDRIANOVA, E. SLEPTSOV, E. YURKEVICH, L. KRYUKOVA

Fig. 3.4. Telemetry data from the magnetometer: Y axis

Fig. 3.5. Telemetry data from the magnetometer: Z axis

The first type of anomalies are long straight lines in the near-horizontal plane of the
plot. These anomalies are nothing but gaps in the data. These gaps indicate the presence
of problems in the recording equipment on board of the spacecraft and/or in the data
transmission over the communication channel. In any case, the analysis of possible reasons
for the presence of these gaps is the prerogative of specialists and experts, and not automatic
tools, due to the lack of data on which it is possible to make a mathematically sound forecast
about the cause of this phenomenon.

The second type of anomalies are vertical bursts that go beyond the boundaries of the
“norm”. These situations can be either single values as in Figure 3.3, or blue indications

Copyright © 2022 ASSA. Adv Syst Sci Appl (2022)



SPACECRAFT EMERGENCY SITUATIONS AUTOMATED SCENARIOS DESIGN 25

similar in nature in Figure 3.2. In the future, the anomalies of the second type will be called
events, i.e. the occurrence of conditions or disturbances.

Fig. 3.6. Example of telemetry with an emergency engine start

Figure 3.6 shows a fragment of telemetry, which represents the emergency activation of
the spacecraft engine (orange color), which can be classified as an emergency situation. The
sequence of events in the telemetry of various spacecraft systems leading to an emergency
situation will be called an emergency scenario. The development of automated software tools
for localization and relief of the emergency situation consequences, followed by the recovery
of the spacecraft operability, is one of the main tasks, solved in the framework of the FDIR
concept.

4. DESCRIPTION OF FDIR METHODOLOGY

The FDIR concept is based on the principles that emergency situations of varying criticality
occur with a certain probability. The probability and criticality of the occurrence of an
emergency situation is determined from previous experience, the results of game-theoretic
modeling, expert knowledge. The parrying of emergency situations depends on its criticality
and can be carried out with the help of predefined scenarios developed on the basis of expert
opinion and experience, see Figure 4.7.

The FDIR also provides for the possibility of parrying single emergency situations. This
is achieved through the use of watchdog timers. A watchdog timer is an electronic timer that
is used to detect and recover faults. During normal operation, this timer is constantly restarted
to prevent the entire system from restarting, while in case of a software or hardware fault, this
timer stops restarting and initiates system restart.

This concept is based on the following principles (Akhmetov et al., 2014; Eickhoff, 2012;
Wander & Förstner, 2013; SalarKaleji & Dayyani, 2013):

Copyright © 2022 ASSA. Adv Syst Sci Appl (2022)



26 A. TUTUROV, O. ANDRIANOVA, E. SLEPTSOV, E. YURKEVICH, L. KRYUKOVA

Fig. 4.7. Levels of FDIR

• design a precise hierarchy of types of faults, determination of the level of software and
hardware that ensures the elimination of every fault;

• construction and implementation of algorithms for automatic transfer of the system to
the safe mode;

• exploitation of autonomous onboard radio command and information-telemetry systems
in the design of the spacecraft, providing reliable control and guaranteed receiving of
telemetry information;

• ensuring the protection of the issuance and realization of potentially dangerous
commands;

• providing the possibility of full or partial cleaning of the onboard RAM and restoring
the original state of the onboard software;

• identification of the faults using explicit and implicit characteristics;
• usage of flexible configuration of software and hardware in system design.

The safe mode of operation of the spacecraft assumes:

• automatic transfer of the spacecraft to the orientation mode, which provides guaranteed
power from solar panels, disconnection of the electrical load associated with the
intended use and with other energy-consuming processes;

• ensuring the operation of the thermal regime system, onboard control system;
• use of redundancy: functional and/or substitution;
• the ability to restore onboard software using a new configuration of onboard hardware

and software, which provides a mode of restoring normal functioning.

FDIR has several levels of settings:

1. The settings that are embedded in ROM during the preparation of the spacecraft on the
ground;

2. Software settings in RAM that are implemented and changed during flight tests;

Copyright © 2022 ASSA. Adv Syst Sci Appl (2022)



SPACECRAFT EMERGENCY SITUATIONS AUTOMATED SCENARIOS DESIGN 27

3. Flexible and adaptive settings, laid down according to certain cycles: year, month,
decade.

The advantages of this concept include:
1. Adaptability and flexibility of spacecraft systems;
2. Autonomous and prompt elimination of emergency situations or automatic transfer to

safe mode for further actions of the MCC;
3. The FDIR concept implies the use of the onboard computer resource, and, accordingly,

the analysis of telemetry information on board, which, on the one hand, reduces the
load on the MCC, allows for the analysis of information promptly and allows you
to make decisions during a communication session with the Earth, and on the other
hand, selectively record the most interesting fragments of telemetry information for long
periods of time.

Accordingly, the disadvantage is the complication of the onboard hardware and software
complex, which requires:

1. complication of initial design;
2. increased requirements on the reliability of the electronic component base;
3. increased requirements on the quality of the program code.

As an implementation of the FDIR concept, the presence of the Automatic
RecOnfiguration (ARO) mechanism in the onboard software can be considered. The essence
of the work of this software mechanism is that each working element of all spacecraft systems
is expertly assigned with weight values, which are summed up in the ARO indicator. When the
value of the ARO indicator due to faults in the operation of the spacecraft elements falls below
the expert-defined value, the procedure for changing the orientation mode of the spacecraft
is automatically started, aimed at stabilizing the functioning of the spacecraft. The telemetry
data of the ARO software mechanism is shown in Figure 4.8.

Fig. 4.8. Telemetry data from ARO

The FDIR concept involves the use of several execution units: a monitor, a script block,
and a logging block. They can be both software and hardware. Later in this article, the design
of the software monitor will be described. Monitor’s task is to identify and register emergency
situations.

Copyright © 2022 ASSA. Adv Syst Sci Appl (2022)



28 A. TUTUROV, O. ANDRIANOVA, E. SLEPTSOV, E. YURKEVICH, L. KRYUKOVA

The main tools for monitoring used in the aerospace industry are various types of Kalman
filters (Pontuschka & da Fonseca, 2014; Lu, Eykeren, Kampen, Chu, & Yu, n.d.; Zolghadri,
2017), and also time series (Troiano, Tipaldi, Hoping, De Pasquale, & Bruenjes, 2012). Time
series have no predictive power. Kalman filter makes it possible to predict, but it can diverge.
DNNs that will be described in the next section have predictive ability, and their convergence
is proved using Lyapunov methods. The speed of their convergence depends on the selection
of parameters.

5. DNN IDENTIFICATION FOR ELECTRO-MECHANICAL PART OF THE
SPACECRAFT SYSTEM

DNNs have the aim of approximating the right-hand side of uncertain systems represented
by dynamical equations. For continuous-time systems, differential neural networks have
proven to be efficient approximate representations of different classes of uncertain
models (A. Poznyak et al., 2001; Wu, 1998; T. Poznyak et al., 2019), meanwhile up to now
recurrent neural networks with time-dependent weighting matrices have not been used much
to represent discrete-time systems.

Measurements from the spacecraft sensors are received discretely, which is why in
the development of this study it is planned to consider the electro-mechanical part of the
measured data as a discrete-time system of the “black box” type, i.e. with unknown motion
laws, and based on the available sensors data to approximate this system using recurrent
neural networks. Generalization of existing methods to discrete-time systems is not a difficult
task, since the mathematical basis for the approximate representation of systems using DNN
is the Lyapunov stability theory, which is sufficiently developed for both continuous and
discrete-time cases. Move on to the description of the system under consideration.

5.1. Problem description
In Figure 5.9, the spacecraft is represented by an unknown electro-mechanical model, which
has the following input and output signals: αi, being the spacecraft rotation angle; ωi, being
the angular velocity of the spacecraft; ωM

i , being the angular velocity of the wheel; ui =
ui

(
ωM
i (kT )

)
, being the control signal; ξi, being the norm-bounded disturbances, i = 1, 3.

Fig. 5.9. Spacecraft as a black box representation

Dynamical representation of the system, given in Figure 5.9, may be proposed in the
following form:

Copyright © 2022 ASSA. Adv Syst Sci Appl (2022)



SPACECRAFT EMERGENCY SITUATIONS AUTOMATED SCENARIOS DESIGN 29

αi(kT ) = fi
(
αi((k − 1)T ), ωi((k − 1)T ), ωM

i ((k − 1)T )
)
+ ξi(kT ), (5.1)

ωi(kT ) = gi
(
αi((k − 1)T ), ωi((k − 1)T ), ωM

i ((k − 1)T )
)
+ (5.2)

ui

(
ωM
i ((k − 1)T )

)
+ ξi(kT ),

ωM
i (kT ) = hi

(
αi((k − 1)T ), ωi((k − 1)T ), ωM

i ((k − 1)T )
)

(5.3)

where functions fi
(
αi, ωi, ω

M
i

)
, gi

(
αi, ωi, ω

M
i

)
, and hi

(
αi, ωi, ω

M
i

)
are unknown and should

be approximated; k is a non-negative integer value; T > 0 is a time step.

5.2. Dynamical (recurrent) neural network identifier

Denoting the state vector of system (5.1)–(5.3) by x =
[
αi, ωi, ω

M
i

]
, supposing that the right-

hand side of the system is unknown or partly known, represent system (5.1)–(5.3) in the form

x(kT ) = Ax((k − 1)T ) +W ∗
1 σ1(x((k − 1)T ))+

W ∗
2 σ1(x((k − 1)T ))ui

(
ωM
i (k − 1)T )

)
+ f̃(x((k − 1)T ))

where A ∈ R9×9 is a Hurwitz matrix, W ∗
1 ∈ R9×9 and W ∗

2 ∈ R9×m are time-invariant
(unknown) weighting matrices (m can be chosen), σ1(x((k − 1)T )) ∈ R3×1 and σ2(x((k −
1)T )) ∈ Rm×1 are activation functions (for example, sigmoidal ones).

DNN identifier takes the following form:

x̂(kT ) = Ax̂((k − 1)T ) +W1(kT )σ1(x̂((k − 1)T ))+

W2(kT )σ2(x̂((k − 1)T ))ui

(
ωM
i (k − 1)T )

)
.

By means of the Lyapunov stability theory for discrete systems and using the concept of
ultimate boundedness, difference equations for the weighting matrices may be found in the
form

Wi(kT ) = Σi(x((k − 1)T ), x̂((k − 1)T ), u(ωM
i (k − 1)T )), i = 1, 2.

The values of W1(kT ) and W2(kT ), obtained from the solution of the minization problem,
connected with Lyapunov inequality, starting from some moment k0T do not or slightly
change, that is why these values may be fixed and used further for solving the prediction
problem. The main task of prediction is not only to restore the future dynamics of the system
using previous states and fixed weighting matrices, but also to detect possible faults in the
system in the future and prevent them.

6. CONCLUSION

The purpose of this work was to explain the necessity of representation of SC in the form of
SCPS and, in this context, to describe the tools that will be used for detection and localization
of anomalies and emergency situations in the system. For this purpose, the experience of
controlling the spacecraft SCon was used, in particular, the analysis of telemetry and the
application of the FDIR concept, as an example of the introduction of expert participation of
the MCC, GCC, and PDMC teams in controlling the spacecraft during the communication
session. In order to achieve the goal in the mentioned operating conditions of the spacecraft, it
was proposed to consider the use of DNN as a mechanism for automated analysis of telemetry
for constructing emergency scenarios, which in the future will become the basis for creating
a decision support system that allows increasing the functional stability of the entire SC and
saving the limited time of the spacecraft communication session for detection, localization,
and recovery of the consequences of an emergency situation on board the spacecraft by
creating an automated archive of emergency scenarios.

Copyright © 2022 ASSA. Adv Syst Sci Appl (2022)



30 A. TUTUROV, O. ANDRIANOVA, E. SLEPTSOV, E. YURKEVICH, L. KRYUKOVA

REFERENCES

Akhmetov, R., Makarov, V., & Sollogub, A. (2014). Problems of unmanned spacecraft
reengineering in abnormal flight situations and their knowledge-based solutions.
VESTNIK of the Samara State Aerospace University, 07, 9.

Azevedo, D. R., Ambrosio, A. M., & Vieira, M. (2012). Applying data mining for detecting
anomalies in satellites. In 2012 ninth european dependable computing conference
(p. 212-217).

D.A., N., & E., R. (2012). Foundations of control methodology. Advances in systems science
and applications, 12, 238-257.

Eickhoff, J. (2012). Onboard computers, onboard software and satellite operations..
Korolev, A. (2018). Functional stability of navigation information systems. Izvestiâ vysših

učebnyh zavedenij. Priborostroenie, 61, 559-565.
Lefebvre, V. (1977). The structure of awareness : toward a symbolic language of human

reflexion..
Lepskiy, V. (1998). The concept of subject-oriented computerization of management

activities (in russian). Institute of Psychology RAS, Moscow.
Lu, P., Eykeren, L. V., Kampen, E.-J. V., Chu, Q. P., & Yu, B. (n.d.). Adaptive hybrid

unscented kalman filter for aircraft sensor fault detection, isolation and reconstruction.
In Aiaa guidance, navigation, and control conference.

Piro, A., Casella, D., Pinori, S., Diciolo, L., Cappelletti, C., Graziani, F., & Battistini, S.
(2018, 05). Hybris: Analysis and design of a hyper-spectral cubesat mission for
multiple remote sensing applications and earth observation synergies..

Pontuschka, M., & da Fonseca, I. (2014). Fdir for the imu component of aocs systems.
Mathematical Problems in Engineering, 09, 1-9.

Poznyak, A., Sanchez, E., & Yu, W. (2001). Differential neural networks for robust nonlinear
control (identification, state estimation and trajectory tracking). World Scientific.

Poznyak, A. S., Oria, I. C., & Poznyak, T. (2019). A survey on artificial neural networks
application for identification and control in environmental engineering: Biological and
chemical systems with uncertain models. Annu. Rev. Control., 48, 250-272.

Poznyak, T., Chairez, I., & Poznyak, A. (2019). Ozonation and biodegradation in
environmental engineering: Dynamic neural network approach (1st ed.). Elsevier.

SalarKaleji, F., & Dayyani, A. (2013). A survey on fault detection, isolation and recovery
(fdir) module in satellite onboard software. In 2013 6th international conference on
recent advances in space technologies (rast) (p. 545-548).

Sleptsov, E., & Andrianova, O. (2021). Control theory concepts: Analysis and design, control
and command, control subject, model reduction. IFAC-PapersOnLine, 54, 204-208.

Talalaev, A., & Fralenko, V. (2018, 01). Fault-tolerant system for organizing highperformance
computing for solving data processing problems. Program Systems: Theory and
Applications, 9, 85-108.

Troiano, L., Tipaldi, M., Hoping, M., De Pasquale, D., & Bruenjes, B. (2012, 10). Satellite
fdir practices using timed failure propagation graphs. In (Vol. 11).

Wander, A., & Förstner, R. (2013). Innovative fault detection, isolation and recovery
strategies on-board spacecraft: State of the art and research challenges..

Wu, J. (1998). Symmetric functional differential equations and neural networks with memory.
Transactions of the American mathematical society, 350(12), 4799–4838.

Zolghadri, A. (2017, 11). The challenge of advanced model-based fdir for real-world flight-
critical applications. Engineering Applications of Artificial Intelligence, 68.

Copyright © 2022 ASSA. Adv Syst Sci Appl (2022)


	Introduction
	The space complex as a socio-cyberphysical system
	Spacecraft telemetry as an information environment for constructing the scenarios
	Description of FDIR methodology
	DNN identification for electro-mechanical part of the spacecraft system
	Problem description
	Dynamical (recurrent) neural network identifier

	Conclusion
	References