=Paper=
{{Paper
|id=Vol-2416/paper60
|storemode=property
|title=Integrity control algorithms in the system for telemetry data collecting, storing and processings
|pdfUrl=https://ceur-ws.org/Vol-2416/paper60.pdf
|volume=Vol-2416
|authors=Viktoriya Berkholts,Arkadiy Frid,Murat Guzairov,Anastasia Kirillova
}}
==Integrity control algorithms in the system for telemetry data collecting, storing and processings ==
https://ceur-ws.org/Vol-2416/paper60.pdf
Integrity control algorithms in the system for telemetry data
collecting, storing and processings
V V Berkholts1, A I Frid1, M B Guzairov1 and A D Kirillova1
1
Ufa State Aviation Technical University, K. Marks st., 12, Ufa, Russia, 450008
e-mail: torina4@yandex.ru, kirillova.andm@gmail.com
Abstract. The issues of improving the security of the modular system for collecting, storing
and processing telemetric information on the state of the onboard subsystems of the aircraft in
automatic mode are considered. It is based on an analysis of the use of modern technologies for
the protection and processing of telemetric information to ensure certain aspects of the
guaranteeability of the system as a whole.
1. Introduction
Emerging malfunctions and pre-failure states of the onboard equipment of the aircraft can be
diagnosed based on telemetric information (TMI). This allows the specialists of ground technical
services to plan repair and preventive measures based on an assessment of the current state of the
equipment. Accumulated and processed TMI will allow specialists of the manufacturer to provide
reasonable support to engineers of ground services in making decisions in case of technical failure of
the blocks and modules of the aircraft. TMI analysis will improve the operational efficiency of the
aircraft in case of any malfunctions and attacks by intruders.
The aim of the study is to increase the security of the system for collecting, storing and processing
TMI on the state of the onboard aircraft subsystems in automatic mode. It is based on an analysis of
the use of modern (including intellectual) technologies for the protection and processing of TMI.
To achieve this goal, a structural diagram of a protected system for collecting, storing and
processing telemetric information on the state of the aircraft subsystems on the basis of a modular
principle has been developed.
2. Analysis of the problem of secure collection, storage and processing of TMI in a
geographically distributed information system
The proposed automated information system (AIS) of ground maintenance services is a set of software
and hardware. They are necessary for the reception, storage and processing of information about the
parameters of the state of complex technical products (CTP) on the aircraft. AIS is a geographically
distributed system that combines the infrastructure of the information systems of ground-based
maintenance stations and the information system of the manufacturer through secure communication
channels. Preparation TMI realized by reading a status log for CTP aircraft during inspection and
maintenance at ground stations via wireless and / or wired sensor networks.
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019)
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
The dependability of the TMI transmission systems with an aircraft allows for a comprehensive
solution of the tasks of ensuring reliability, fault tolerance, availability, security, maintainability, and
observability. An urgent task is to build a hierarchy of models that allow a comprehensive assessment
of various aspects of the TMI transmission system and the development of a methodology for
assessing the integral indicator of the system’s guaranteed performance.
Ensuring the availability of TMI transmission systems is the primary task of ensuring the effective
functioning of the aircraft (A/C). The volume of TMI collected is significant. It is an incentive for the
development of the concept of the industrial Internet of things (IIoT). It is a promising platform for use
in solving such problems [1]. For example, a jet airliner demonstrated at the Bombardier Paris Air
Show, whose engine is equipped with more than 5,000 sensors that generate up to 10 GB of data per
second. One twin-engine aircraft can generate up to 844 TB of data average in 12-hour flight [2].
The ability to transfer TMI about the actual state of individual modules during operation and the
entire A/C equipment complex in real time to the manufacturer of aeronautical engineering
components will improve the operational efficiency of the aircraft in its normal state and in the event
of malfunctions and attacks by intruders when investigating incidents. Thus, a study of ground-to-air
communication systems showed that ACARS, despite its versatility and widespread use, is vulnerable,
and if hacked in conjunction with ADS-B, an attacker can gain access to the flight control system,
download flight plans and detailed commands [3].
Ensuring the availability of telemetry information on the state of the aircraft
An automated information system (AIS) of ground maintenance services is a set of software and
hardware. They are necessary for receiving, storing and processing information on the technological
parameters of complex technical products (CTP) on board the aircraft.
A review of the main approaches to the relevance of the problem of ensuring the reliability of such
systems is considered in the works of the authors [4, 5, 6]. AIS solves the main problems associated
with the reception of TMI on the state of the onboard aircraft systems. The main methods of obtaining
data are presented in the figure (Figure 1):
1. Directly from the CTP
2. By means of reading devices of the event log from the sensors of modules CTP. When carrying
out technical inspection and maintenance, devices of this type read and store data on the state of the
modules throughout the entire previous period of operation [5].
3. Entering events into the database manually. The operator processes the information and enters
the information through the WEB application [4].
Figure 1. Methods for obtaining TMI.
In the first case, telemetry information is transmitted from the aircraft through a radio channel. To
create a transmission channel, the following approaches can be used:
• Communication satellites (IRIDIUM, SATCOM).
Existing telemetry data transmission technologies use satellite communications. For example, the
GE Aviation concern, producing aircraft engines, transmits telemetry from the aircraft in this way.
The obvious disadvantage of such way of transfer is its high cost. Streaming telemetry information
involves the transfer of significant amounts (gigabytes) of data sent. The second disadvantage is the
low noise immunity of the satellite communication channel. Incorrectly transmitted data can serve as a
signal for a false alarm, or there is a chance to miss a system failure.
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 488
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
• Channel of wireless high-speed data transmission LTE and LTE-a
Air to ground (A2G) LTE is capable of providing data rates of up to 75 Mbps for ground-to-air
communications and up to 25 Mbps for air-to-ground communications at distances of 100 kilometers
and speeds of 1,200 kilometers per hour using licenses. FDD 2x15 MHz. The standard 4G LTE can be
used for continental flights, developed by Nokia.
Despite the currently available satellite and hybrid A2G systems, there is still no low-cost, high-
throughput solution for broadband in-flight.
Thus, none of these technologies has no set of properties that allow for continuous broadcast
telemetry data. ACARS does not allow to transfer a large amount of accumulated data, satellite
communication is too expensive, and LTE-A is still at the development and implementation stage, and
in the future it will cover only the continental part of flights.
Moreover, the current TMI transmission and processing systems demonstrate vulnerabilities that
allow an attacker to gain access not only to passenger and airline data, but also to significantly affect
flight parameters.
In the second and third cases, the information enters the database through a WEB application,
which is an insulating layer between external networks and the internal structure of the AIS, since
access from the external network is one of the most vulnerable points of the system. Improving the
security of access to the database (DB) containing critical information about the product in use is
based on the development of the architecture of a secure WEB application that acts as an insulating
layer for external AIS clients, which allows for the possibility of transferring and analyzing ground-
based service points from the aircraft and provide the ability to remotely access the necessary data.
The architecture of this solution is presented in [5].
Preventing the appearance of vulnerabilities in the WEB application was carried out by
implementing measures to develop secure software established by GOST R ISO / IEC 12207.
Modeling security threats and identified vectors of possible attacks, as well as analyzing them, made it
possible to formulate countermeasures for each of the vectors at different architectural levels WEB-
applications. However, the analysis of the security of the entire TMI transmission system requires
advanced modeling and the construction of a detailed model of interaction between the onboard
information system of the aircraft and the ground-based AIS.
The growth of telemetry information forces the aviation industry to consider new approaches to the
collection and analysis of a large amount of data on the state of individual components and elements
of the aircraft. The concept of the industrial Internet of Things is being actively developed - an
expanded network consisting of a large number of devices equipped with a set of sensors that
communicate with each other through low-power and short-term wireless connections. The first step is
to collect data from the sensors. One of the most promising solutions is a protocol with low power
consumption and small radius of IEEE 802.15.4 IEEE 802.15.4e transmission. Short range is sufficient
for data transmission within the ground service station. The IEEE 802.15.4 and IEEE 802.15.4e
protocols and their architecture layers comply with IETF standards.
The question of analyzing the security of the system for collecting, transmitting and receiving
telemetry information about the state of individual elements of the onboard aircraft systems during
data transmission over the first two channels remains open.
The decomposition of the TMI transfer in the form of a hierarchical model of interacting levels of
collecting, transmitting and analyzing information with the corresponding protocol stack is the basis
for analyzing and building a system for analyzing the transmission system security (Figure 2).
In recent years, satellite communication systems in accordance with the Regulations of the
International Telecommunication Union (ITU) are switching to a higher-frequency Ka-band (15.40-
26.50 and 27.00-30.20 GHz).
The grouping of satellites in geostationary orbit and ground control centers make it possible to
form a network infrastructure with high reliability indicators and the possibility of building distributed
state networks. Channels of transmission of such networks provide a fairly high level of encryption
and data protection.
Transmission of information in such networks is characterized by a low level of errors - no more
than one per 10 million transmitted information bits and reliable operation - up to 100 thousand hours.
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 489
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
The speed of work on the satellite channel is from 16 Kbps to 10 Mbps and more, which is comparable
with the data transfer rate in the terrestrial channel.
The main methods for ensuring the security of telemetry information transmission in a wireless
satellite channel is the use of software and hardware means of information protection. Widely used
standards for secure protocols IPsec.
IPsec protocol (set of protocols) provides:
• integrity of the virtual connection, authentication of the source of information using the AH
protocol (Authentication Header);
• encryption of information transmitted via the ESP (Encapsulating Security Payload) protocol;
• initial connection setup, mutual authentication and confidential key exchange.
Figure 2. Extended scheme of data transmission from the aircraft.
At present, modern bilateral satellite communication networks use coding systems at the software
and hardware level, which makes the interception and decoding of information over the radio channel
almost impossible.
All data transmitted via satellite channel pass through a multi-stage system of transformation and
encryption. The result of this:
• application of proprietary data encryption algorithms;
• terminal authentication when it is registered on the operator’s network (hardware key);
• encryption of both the entire session (software key) and each session separately (session keys);
• application of proprietary algorithms for converting source data into internal data formats
(structures), which are then transmitted via satellite channel; thus, the tasks of additional
protection of information, delivery of service information and error correction are solved;
• in the created virtual channels, source data in TCP sessions are grouped, compressed, and
prioritized.
Satellite channels in the direction from A/C to TMI processing centers are reverse satellite
channels. Currently, the most common ways of functioning of transmitters in such channels are the
principles of access with time-frequency division of TDMA / FDMA channels. Each reverse channel
is located in a certain frequency range or has a carrier with frequency modulation and with a given
coding algorithm for detecting and correcting errors of transmitted data - Turbo Coding.
To transmit TMI from the aircraft, it is necessary to provide a mechanism for changing the
frequencies of the carrier reverse channels, which makes it much more difficult to intercept the
transmitted data.
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 490
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
Data encryption in the satellite channel is carried out with the participation of both satellite
terminals on board the aircraft and specialized high-performance servers at the TMI ground collection
station. The server of the ground station for collection and processing of TMI hosts a secure database
of encryption keys and session keys of all satellite terminals. In order for the aircraft board to operate
in the transmission network, the information in the key database must match the hardware onboard
key.
Telecommunications systems using the UMTS (Universal Mobile Telecommunications System)
standards are third-generation mobile communication systems - 3G. For mobile communication of the
third generation, the decimeter frequency band is used (about 2 GHz), and data transmission is
provided at a speed of 2 Mbit / s.
All information security threats in the UMTS network can be distributed depending on the location
of the impact of their respective attacks:
• on the radio access area (radio interface);
• on other parts of the network.
The radio section between the aircraft and the service network is one of the most vulnerable points
of attack in UMTS. The threats related to this site and described below are divided into the following
categories:
• unauthorized access to data;
• threats to data integrity;
• “denial of service”;
• unauthorized access to services.
Table1. Describes some of the threats to unauthorized access to data on the radio site.
Threat
Threat name Threat description
designation
Т1а Interception of user traffic Violators can intercept user traffic
Т1b Interception of alarm and Violators can intercept alarm data and control data
control data
Т1с Masking as a participant Violators can be disguised as a network element
Т1d Passive traffic analysis Violators can monitor the characteristics of messages
Т1е Active traffic analysis Violators can actively initiate a connection and then
access information
Table 2. Threats to the integrity of information.
Threat
Threat name Threat description
designation
Т2а User traffic manipulation Violators can modify, insert, repeat or destroy user
traffic. This manipulation may be accidental or
intentional.
Т2b Alarm data manipulation The intruder can modify, insert, repeat, or destroy
alarm or control data.
Table 3. Denial of service threats.
Threat
Threat name Threat description
designation
Т3а Physical intervention Violators can physically interfere with the
transmission of user traffic, signaling data and control
data.
Т3b Protocol Intervention Violators may introduce special protocol failures.
Т3с Denial of service due to Violators may refuse to serve a legitimate user.
masking as a participant in
communication
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 491
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
Below are tables with descriptions of information security threats. The accepted designations of the
threat TAn correspond to: T - the first letter of the English word "threat" (threat): A - the number
corresponds to the number of the threat group (table number); n - the letter corresponds to the ordinal
number of the threat in the threat group (in accordance with the list of threats in the ETSI document.
Table 4. Threat of unauthorized access to services.
Threat
Threat name Threat description
designation
Т4а Masking as another user The intruder is disguised as another network user.
First, the intruder is disguised as a base station with
respect to the user.
3. Threats related to attacks on other parts of the system
Although attacks on a radio channel represent the most serious threats, attacks on other parts of the
system also require analysis from the point of view of information security.
Table 5. Threats to unauthorized access to data.
Threat
Threat name Threat description
designation
T5e Unauthorized access to data Violators (by physical influence or logical control)
on the system object can gain access to local or remote data.
T5f Compromising information A legitimate user of a UMTS service may obtain
about the location information about the location of other users of the
system
T6с Manipulation of masking as a Violators can be disguised as a network element in
communication partner order to modify, insert, repeat or destroy traffic
T6f Manipulation of data on the Violators can modify, insert, destroy data that is
objects of the system contained in the objects of the system.
T7а Physical intervention Violators may interfere with transmission on any
system interface (wired or wireless). For example, the
physical method of an obstacle on a wired interface
could be a broken wire.
T7b Protocol intervention Violators can interfere with the transmission of user
traffic or signaling data on any interface of the system
(wired or wireless) or by signaling the protocol to fail.
T7с Denial of service by masking Violators can deny service to users by impeding the
communication partners transmission of user traffic and signaling data,
controlling them by blocking as a result of masking as
a network element.
T7d Incorrect use of emergency Violators can interfere with access to the services of
services other users and at the same time cause disruption of
the equipment to perform functions in emergency
situations.
T8а Disagreement with the Disagreement with the submitted invoice. This may be
submitted invoice expressed in the refusal of the service or in the refusal
that the service was actually provided.
T8b Failure of user traffic source The user can refuse to send traffic.
T9а Custom masking Violators can introduce themselves as a user in order
to use the authorized services of this user. The
intruder was able to get this opportunity from other
objects such as the serving network, home
environment and even the user himself.
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 492
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
T9b Masking under the serving Violators can introduce themselves as a service
network network or part of a service network infrastructure.
T9с Home environment masking Violators can introduce themselves as a home
environment in order to obtain information that
enables them to disguise themselves as users.
T9d Misuse of user priorities Users may misuse their assigned priorities in order to
gain unauthorized access to services or simply use
their subscription intensively for free.
T9е Incorrect use of serving Service networks may misuse their priorities to gain
network priorities unauthorized access to services.
Table 6. Threats related to attacks on the terminal and UICC / USIM.
Threat
Threat name Threat description
designation
T10h Masking to receive data on Violators can be disguised as USIM or a terminal in
the UICC interface - terminal. order to intercept data on the interface of a UICC
terminal.
T10j Confidentiality of certain user Violators may wish to gain access to personal user
data in the terminal and UICC data stored in the terminal or UICC, for example, the
/ USIM telephone book of interacting subscribers.
T4а, It must be possible to prevent Requirements for security access to service.
Т9а, unauthorized access to 3G
Т9с services by disguising
themselves as legitimate
users.
T4а, An alarm should be provided Security Requirements
Т8а, to the provider informing him
Т9d, of the security event. Provide
T9e service providers with the
ability to authenticate users
upon request and during the
provision of the service.
T7b, Protection against System integrity requirements
Т7с unauthorized modification of
user traffic should be
provided.
T7а, Protection against System integrity requirements
Т7b, unauthorized modification of
Т7с certain signaling and control
data must be provided.
T1a, The user should be provided Personal data protection requirement
Т1b with the ability to verify that
his traffic and information
about his calls are
confidential.
T10h, It should not be possible to USIM security requirement
T10k access USIM data that is
intended for use only within
USIM, such as authentication
keys and algorithms
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 493
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
4. Development of a block diagram of a secure system for collecting, storing and processing
telemetric information on the state of the aircraft subsystem
The generalized structure of a geographically distributed hierarchical system for the collection, storage
and processing of TMIs arriving from airplanes based on ground maintenance stations is presented in
Figure 4.
Figure 4. Generalized structural diagram of a protected system for collecting, storing and processing
TMI (1 – level of TMI collection from wired and wireless sensors of ground-based aircraft servicing
systems; 2 – level of primary surveillance and preparation of TMI for transmission to the AIS of the
manufacturer’s enterprise (AIS EM); 3 – level of data transmission over secure channels through
global data networks in the AISEM; 4 – level of organization of reception and distribution of TMI on
EM; 5 – level of storage and processing of TMI in the CIS).
The creation of a secure channel through global communication networks and the transfer of TMI
to a part of the AIS EM is realized at the transmission level of accumulated data. Organization levels
of reception and distribution of information at the enterprise are realized according to the three-layer
CISCO model. There is a level in the corporate information network of EM. It includes subsystems for
storage and processing of TMI. Also, there is a segment designed to support and implement the
business processes of the enterprise.
5. Development of the structure of the collection and storage subsystem TMI at the ground
stations of aircraft maintenance
The vast majority of Industrial Ethernet protocols do not have built-in security mechanisms.
Consequently, the actual problem is the security of industrial networks.
To ensure the security of subsystems that implement the first two levels of the proposed structure,
it is necessary to be guided by the normative documents of the international and federal standards.
When designing the wireless sensor network collection subsystem of the TMI, take into account the
requirements of GOST R ISO / IEC 27033-1-2011 and GOST R ISO / IEC 27033-3-2014.
The physical architecture of the TMI collection and storage subsystem at the ground is presented in
Figure 5.
Mechanisms for collecting and storing a large amount of TMI on the state of individual
components and elements of aircraft should take into account the actively developing concepts of the
industrial Internet of things (IIoT). It is proposed to use heterogeneous wired (physical RS-485
interface) and wireless sensor networks (IEEE 802.15.4, IEEE 802.15.4e) to collect protocol-based
TMI using embedded Modbus over TCP mechanisms to ensure the protection of transmitted data
(streaming encryption). The IEEE 802.15.4 and IEEE 802.15.4e protocols and their architecture levels
follow the IETF standards.
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 494
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
Figure 5. The subsystem of data collection and storage at aircraft service stations.
6. Development of the structure of the subsystem for receiving, storing and processing TMI in
AIS
The organizations for receiving and distributing TMIs on PIs are implemented according to the three-
tier CISCO model and Security Architecture for Enterprise (SAFE) design methodology, which allows
to take into account modern experience in deploying secure networks based on the deep-echelon
defense against external and internal attacks.
The main element of the TMI distributed processing system is the distributed file system HDFS.
Additional measures to ensure the confidentiality of stored data is encryption at the level of individual
database columns. To audit access to big data, you need to apply Database Activity Monitoring class
solutions.
7. The concept of data integrity and verification of data sources
The actual problem is the security of industrial networks, which is not solved by existing approaches,
since the attacker's intervention is possible not only at the network level from outside or inside, but
also at the level of the data sources themselves. The main hardware and software part of the CTP is
free from possible "bookmarks", which is guaranteed by the manufacturer, but it is necessary to
comprehensively analyze the progress of the object, identifying abnormal situations not related to
equipment breakdowns or failure of individual components and assemblies, but potentially caused by
the intervention of an attacker. It is necessary to improve the monitoring system of CTP as an element
of the intrusion detection system, considering the complexity of the control object, the nonlinearity of
the processes and the possible conditions of the equipment that lead to emergency or catastrophic
situations. The system of monitoring the condition of CTP, implemented as a component of the
intrusion detection system, involves continuous monitoring of the parameters of CTP to identify
significant deviations from the" normal behavior", which in turn will indicate possible malicious
intentions. This approach is a development of the concept of Data Centric Security [7], which implies
the security of the data itself. To determine deviations from the" normal behavior " it is proposed to
use the system model of the object – the CTP, which is the development of the concept of Fault
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 495
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
Detection and Identification [8]. The process of monitoring the state of the CTP is a sequential
operation of collection, processing and analysis of technological information, the main of which is to
detect the impact of an attacker on the course of the CTP and the components of the information
system by comparing the mathematical model of the CTP and the current performance of the real
object will improve the security of the object. Using the proposed concept of monitoring the CTP
comparing the fixed state of the object with the real model is a tool that allows you to control the
presence of hardware and software interventions in the infrastructure of the information system.
The algorithm is based on a comparison of the characteristics of time series arriving from the
aircraft, and time series generated by the gas turbine engine (GTE) model which is simulating the
same mode and the same flight conditions in which the GTE operates in real time. Information
receiving from A/C is the result of GTE and its ACS operation.
Figure 6. Method of analyzing the consistency parameter of model and field data.
Figure 7 shows the block diagram of the monitoring system of GTE A/C TMI parameters. The AIS
of the manufacturer receives the vector of specified values of the controlled coordinates of the object
Y0, the vector of perturbing factors F and the vector of measured perturbing factors F’. The control
object receives the vector of control actions U generated by the control system. Data from the
monitoring system via a communication channel is sent to the manufacturer. The communication
channel can be exposed both to an attacker (the Z vector) and to the external environment that
generates noise (the N vector). Similar control signals are received in a model that simulates signals
from A/C sensors.
A set of GTE parameters was selected to analyze the data discrepancies obtained from the model
and data obtained from the aircraft. The k-means clustering method was performed on the training
sample, during which nine types of GTE behavior dynamics were identified.
For carrying out preliminary experiments and learning the decision block for each type of
dynamics, its own neural network NARX model was built.
For real-time integrity monitoring, a multidimensional time series (TS) of the mismatch parameters
of model data and TMI indicators for a sliding window is constructed [9].
To make a final decision on the state of the data transmission system from the aircraft and the
presence or absence of intruder interventions into the data channel, a decision block was built that
takes into account not only the type of data mismatch from the aircraft and the data generated by the
model, but also the signal control systems about the state of the GTE, obtained from the aircraft in real
time, as well as the type of dynamics of the GTE at a given time. All these three parameters are taken
into account for the block to make a decision on the state of the data transmission channel ("break",
"normal operation", "integrity violation", etc.). The second output signal of the decision block is an
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 496
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
estimation of the degree of confidence about the decision made, namely the estimate of the probability
of such a state, calculated from the three input parameters [10].
Figure 7. Block diagram of the integrity monitoring system of GTE A/C TMI parameters.
Figure 8. Simulation experiment №1 of hardware failure.
8. Simulation of possible situations arising during the A/C operation
Experiment 1.
In this example, a hardware failure or interruption in signal transmission occurs. From the figure it is
clear that, starting with iteration 85, the values of the data received from the object have changed
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 497
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
dramatically and the value of their average value has decreased relative to the average value of the
previous values obtained from the sensor of the technological process. The changes in the values are
shown in Fig. 8. The fall of the amplitude of the received signal occurred almost at the last iterations
in the current time window.
When calculating the correlation and determination coefficients, the following results were
obtained:
rxy = 0.56
𝑅 = 0.34
The average value of the coefficients indicates possible problems with the data obtained from the
process.
For an additional test of this data, the MAP coefficient was calculated:
𝑀𝐴𝑃𝐸 = 11,88%
Check the value of the MAPE coefficient before the jump:
𝑀𝐴𝑃𝐸 = 4.6%
As can be seen, the MAPE coefficient has changed, but not critically. Such a small change in the
level of MAPE error and the average value of the correlation and determination coefficients is due to
the late signal jump in the time window.
Experiment 2.
In this example, the signal transmission is also interrupted. At about 77 iterations, the average value of
the transmitted signal has changed dramatically in comparison with previous data. As in the previous
example, in the current time series, signal distortion occurs at late iterations, which will reduce the
sensitivity of the coefficients of determination and correlation.
Figure 9. Simulation experiment №2 of hardware failure.
When calculating the correlation and determination coefficients, the following results were
obtained:
𝑟𝑥𝑦 = 0.19
𝑅 = 0.08
The values of both coefficients are too low, which indicates a serious problem with the received
data. Calculate the coefficient MAPE:
MAPE=16,31%
The MAPE error rate is high, which confirms the hypothesis that there are problems in the data
transmission channel.
Calculate the MAPE coefficient before the start of problems with the signal:
MAPE = 4.11%
In this example, the signal drop occurred earlier than in the previous one. This explains the lower
values of the correlation coefficients and determination, and the percentage of errors, on the contrary,
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 498
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
is higher. It is also worth noting that the percentage of errors before the start of the signal jumps is
approximately the same in both examples.
Experiment 3.
This experiment also illustrates problems with signal transmission or equipment failure. The fall in the
average value occurs in the time window fairly early, at about 37 iterations.
When calculating the correlation and determination coefficients, the following results were
obtained:
𝑟𝑥𝑦 = 0,36
𝑅 = 0.11
The values of both coefficients are too low, which indicates a serious problem with the received
data. Calculate the coefficient MAPE:
MAPE=41.81%
The MAPE error rate is high, which confirms the hypothesis that there are problems in the data
transmission channel.
Calculate the MAPE coefficient before the start of problems with the signal:
MAPE = 4.61%
Figure 10. Simulation experiment №3 of hardware failure.
Figure 11. Simulation experiment №4 of hardware failure.
The correlation coefficient is not too different from the previous example, but the MAPE error rate
is very high. Such a large value is due to the early appearance of signal distortion in the time window.
Experiment 4.
This experiment is different from previous ones. Here, starting from the 69th iteration, there is a
smooth growth of data values obtained from the technological process in parallel with the growth of
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 499
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
model values; however, the average value of the received signal is significantly higher than the
reference data. Such data behavior may have a weak effect on the values of the correlation coefficient
and determination.
When calculating the correlation and determination coefficients, the following results were
obtained:
𝑟𝑥𝑦 = 0.98
𝑅 = 0.97
The values of both coefficients are too low, which indicates a serious problem with the received
data. Calculate the coefficient MAPE:
MAPE=23.11%
The MAPE error rate is high, which confirms the hypothesis that there are problems in the data
transmission channel.
Calculate the MAPE coefficient before the start of problems with the signal:
MAPE = 5.01%
As shown by calculations, the coefficients of determination and correlation take a very high value.
However, when paired with a high MAPE, the values obtained should raise suspicions. There may be
problems with the equipment, as well as falsification of the transmitted data.
Experiment 5.
This experiment is a case where nothing happened to the data and it is transmitted normally.
Figure 12. Simulation experiment №5 of hardware failure.
When calculating the correlation and determination coefficients, the following results were
obtained:
𝑟𝑥𝑦 = 0.94
𝑅 = 0.90
Calculate the coefficient MAPE:
MAPE=4.72%
The combination of these three factors indicates that there are no problems with the signal. The
resulting data can be trusted.
Experiment 6.
This example is similar to the previous one. Here the data obtained from the aircraft is just noisy
and not distorted in any way.
When calculating the correlation and determination coefficients, the following results were
obtained:
𝑟𝑥𝑦 = 0.76
𝑅 = 0.58
Calculate the coefficient MAPE:
MAPE=2.38%
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 500
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
Figure 13. Simulation experiment №6 of hardware failure.
The correlation and determination coefficients show the average value of the relationship between
the data obtained from the model and the data obtained from the object of the technological process.
However, the MAPE value is low. This combination of the three coefficients allows us to say that
everything is in order with the obtained data and that they can be trusted.
Experiment 7.
This experiment explains the low correlation coefficient from example two. Here, as well as in the two
previous experiments, data is presented that has not been changed by the attacker.
Figure 13. Simulation experiment №6 of hardware failure.
When calculating the correlation and determination coefficients, the following results were
obtained:
𝑟𝑥𝑦 = 0.21
𝑅 = 0.04
Calculate the coefficient MAPE:
MAPE=4.25%
Such a low correlation coefficient and determination may appear not only in case of equipment
failure or fake messages. In the case when the model readings are close to a constant, and the data
from the technological object undergo analog-digital transformations and undergo slight distortions
during transmission, the correlation coefficient takes very low values. However, the MAPE coefficient
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 501
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
shows a small percentage of errors, which means that the deviations of the model and TP object values
are insignificant. Data can be trusted.
9. The rule base for decision making
For the decision block, the following set of rules was developed, on the basis of which the decision on
the integrity of the transmitted TMI was made.
CMS is the channel monitoring system. In case if CMS = 1, everything is normal with the channel,
otherwise - CMS = 0;
Rule 1. �𝑟𝑥𝑦 =′ middle′ �𝐴𝑁𝐷(𝑅 =′ middle′ )𝐴𝑁𝐷 (𝑀𝐴𝑃𝐸 =′ middle′ ) AND(CMS=0) THEN
signal breakage occurred;
Rule 2. �𝑟𝑥𝑦 =′ low ′ �𝐴𝑁𝐷(𝑅 =′ low′)𝐴𝑁𝐷(𝑀𝐴𝑃𝐸 =′ high′)AND (CMS=0) THEN signal
breakage occurred;
Rule 3. �𝑟𝑥𝑦 =′ low ′ �𝐴𝑁𝐷(𝑅 =′ low′)𝐴𝑁𝐷(𝑀𝐴𝑃𝐸 =′ high′)AND (CMS=0) THEN signal
breakage occurred;
Rule 4. (𝑟𝑥𝑦 =′ high′)𝐴𝑁𝐷 (𝑅 =′ high′ )𝐴𝑁𝐷(𝑀𝐴𝑃𝐸 = ′high′) 𝐴𝑁𝐷 (CMS=1) THEN data fraud
is possible;
Rule 5. (𝑟𝑥𝑦 =′ high′)𝐴𝑁𝐷( 𝑅 =′ high′)𝐴𝑁𝐷 (𝑀𝐴𝑃𝐸 = ′high′) AND (CMS=1) THEN normal
operation;
Rule 6. (𝑟𝑥𝑦 =′ средний′)И(𝑅 =′ средний′)И(𝑀𝐴𝑃𝐸 = ′низкий′) И (CMS=1) THEN normal
operation;
Rule 7. (𝑟𝑥𝑦 =′ low′) 𝐴𝑁𝐷(𝑅 =′ low′)𝐴𝑁𝐷 (𝑀𝐴𝑃𝐸 = ′low′) AND (CMS=1) THEN normal
operation.
10. Conclusion
A block diagram of a protected system for collecting, storing and processing telemetric information on
the state of aircraft subsystems based on the modular principle is proposed. The difference of the
proposed solution is that it contains rather large subsystems with a high degree of connectivity of the
components inside and a sufficient degree of autonomy at the level of interaction of the subsystems
themselves. Each subsystem is built on the basis of organizational principles specific to the specifics
of the problem being solved, and is governed by existing regulatory documents to ensure specific
aspects of the system's reliability.
An algorithm for monitoring the integrity of the TMI on the state of the GTE in service has been
developed, which determines the type of TS dynamics coming not only from the aircraft, but also the
type of the TS model and GTE mismatch dynamics, which allows evaluating the actual state of the
GTE ACS and detecting intrusion interference.
11. References
[1] Internet of Things Volume G4: Security Framework URL: http://www.iiconsortium.org/
pdf/IIC_PUB_G4_V1.00_PB-3.pdf (10.11.2018)
[2] Rapolu B 2016 Internet of Aircraft Things: An Industry Set to be Transformed AVIATION
WEEK NETWORK URL: http://aviationweek.com/connected-aerospace/internet aircraft-things-
industry-set-be-transformed (10.11.2018)
[3] Aircraft Hacking Practical Aero Series n.runs Professionals Stations URL: http://
www.sita.aero/file/3744/Aircom (10.11.2018)
[4] Frid A I, Vulfin A M, Zakharov D Ju, Berkholts V V and Mironov K V 2017 Architecture of
the security access system for information on the state of automatic control systems of aircraft
Proc. of the 19th Int. Workshop on Computer Science and Information Technologies 2 21-27
[5] Frid A I, Vulfin A M and Berkholts V V 2018 Analysis of the methods of constructing
information attack models for the system of telemetric information transmission Proc. of the
Information Technology Intelligent Decision Support (Russia: Ufa) 226-229
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 502
Data Science
V V Berkholts, A I Frid, M B Guzairov and A D Kirillova
[6] Rapolu B 2016 Internet of aircraft things: an industry set to be transformed AVIATION WEEK
NETWORK URL: http://aviationweek.com/connected-aerospace/internet-aircraft-things-
industry-set-be-transformed (13.11.2018)
[7] Zhang L, Afanasyev A, Burke J, Jacobson V, Crowley P, Papadopoulos C, Wang L and
Zhang B 2014 Named Data Networking ACM SIGCOMM Computer Communication Review
44(3) 66-73
[8] Choi S W, Lee C, Lee J M, Park J H and Lee I B 2005 Fault detection and identification of
nonlinear processes based on kernel PCA Chemometrics and Intelligent Laboratory Systems
75(1) 55-67
[9] Kropotov Yu A, Proskuryakov A Yu and Belov A A 2018 A method for predicting changes in
the parameters of time series in digital information control systems Computer Optics 42(6)
1093-1100 DOI: 10.18287/2412-6179-2018-42-6-1093-1100
[10] Plotnikov D E, Kolbudaev P A and Bartalev S A 2018 Identification of dynamically
homogeneous areas with time series segmentation of remote sensing data Computer Optics
42(3) 447-56 DOI: 10.18287/2412-6179-2018-42-3-447-456
Acknowledgments
This work is partially supported by the Russian Science Foundation under grants № 17-07-00351.
V International Conference on "Information Technology and Nanotechnology" (ITNT-2019) 503