11th IEEE International Conference on Cognitive Infocommunications – CogInfoCom 2020 • September 23-25, 2020 • Online on MaxWhere 3D Web

Principles of Anti-Drone Defense

Petar Čisar Robert Pinter

University of Criminal Investigation and Police Studies Subotica Tech, Department of Informatics
Belgrade-Zemun Subotica
Serbia Serbia
[email protected] [email protected]

Sanja Maravić Čisar Milan Gligorijević

Subotica Tech, Department of Informatics Ministry of Interior
Subotica Belgrade
Serbia Serbia
[email protected] [email protected]

Abstract—The large increase in the number of drones, as well The primary role in the successful defense of drones is their
as the possibility of their use for malicious (illegal) purposes, has detection. Cognitive technologies have a significant place in
raised the importance of defense from them. This paper gives an this area. Cognitive infocommunications (Cog InfoCom) is an
overview of drones, the use of different modulation techniques, interdisciplinary research field that has appeared as a synergy
remote control systems and antennas since the knowledge and between infocommunications and the cognitive sciences [2,3].
proper understanding of their characteristics enables the optimal Оne of the main characteristics of CogInfoCom is that these
choice of adequate protection methods. The paper also analyzes areas intertwine at different levels and as a result, new forms
different aspects of modern and commonly used principles of of cognitive abilities are born [4-7]. These capabilities are
anti-drone defense. Proper consideration of technical
neither completely natural (i.e. human) nor completely
characteristics in the sphere of drone communication enables the
artificial, they are “blended combinations”, so it is suggested
focus on adequate solutions, depending on the specific protection
requirements. Cognitive technologies have a significant place in that they are treated in a way that combines both an
terms of improving drone detection. engineering and a human perspective [8-11]. CogInfoCom
connects research areas such as, for instance, digital
Keywords—anti-drone defense; radio modulation; remote transformation environment for education [10, 12-19], human-
control; antennas; spoofing; jamming; radar; hacking; cognitive computer interaction [20], brain-computer interfaces and
detection human and bio-interfaces [4-5, 21], IoT [22], VR based
education [6-9, 23-29], affective computing and cognitive
I. INTRODUCTION linguistics and multimodal interaction.
A drone, in a technical sense, is an unmanned aircraft. The main characteristic of the radar system is to constantly
Drones are more generally known as unmanned aerial vehicles observe the surrounding environment, on the basis of
(UAVs). Basically, a drone is flying equipment that can be observations to make conclusions about environment nature
remotely or autonomously controlled in accordance with and to modify observations of the environment by changing
software-controlled instructions in an embedded system, the waveforms and antenna pointing direction. When we talk
working synchronized with onboard sensors and GPS. about cognitive radar, the system preserves real-time
hypotheses about the environment in order to adaptively
There are several main components that make up a control the form and parameters of the actual sensing
conventional drone: different sensors, brushless motor, procedure. Namely, prior and current knowledge are quantified
propellers, electronic speed controller (ESC), micro servos, and updated in real time in order to optimize the sensing
flight controller, receiver, transmitter, antenna, Lipo battery procedure in a closed-loop manner, thus optimizing radar
and camera. The remote control system of a drone is typically performance and making efficient use of finite sensing
based on wireless communications, such as 2.4 GHz, 5.8 GHz, resources [30].
or Wi-Fi (IEEE 802.11 a/b/g/n).
In the following sections, this paper elaborates on the
Most of the commercial drones can be divided into one of technical characteristics of the components and principles used
the following categories [1]: Linux - based drones, Arduino- in drones, with particular emphasis on the possibility of their
based drones, smartphone-controlled drones, drones controlled application to anti-drone defense.
by a little known (nameless) radio transmitter and drones
controlled by a radio transmitter from a reputable
manufacturer.

978-1-7281-8213-1/20/$31.00 ©2020 IEEE 000019

Authorized licensed use limited to: Raytheon Technologies Corporation. Downloaded on December 19,2020 at 07:38:28 UTC from IEEE Xplore. Restrictions apply.
 P. Cisar et al. • Principles of Anti-Drone Defense

II. RADIO MODULATION TECHNIQUES USED IN DRONES frequency hopping, whereas DSSS applies pseudo-noise to
PWM (Pulse Width Modulation) and PPM (Pulse Position change the signal’s phase. FHSS is accomplished by allocating
Modulation) are common protocols used to transfer data from smaller channels within the large bandwidth which fit the data.
radio receiver to the flight controller. PWM is a modulation The next step is to send the signal pseudo-randomly into
technique that transmits data in the form of variable pulse another channel. At any given time momentum, just a single
width. In PPM the sample values determine the position of a channel is used. On the other hand, DSSS spreads the
narrow pulse relative to the clock signal. PAM (Pulse information across the band by inserting the pseudo-random
Amplitude Modulation) is the primary structure of pulse noise into the signal to change its phase at any given time. This
modulation. In this case, the input signal is sampled at equal results in an output that looks like static noise. Within a
time intervals and each sample is proportional to the amplitude dispreading process, the original signal can be extracted from
of the modulating signal. The principle of generating these the composite signal as long as the pseudo-random sequence is
signals is shown in Fig. 1. known.

In addition to the analog radio protocols, the drone In view of the aforementioned communication protocols,
techniques also use the SBUS (Serial Bus) serial their graphical overview can be shown in Fig. 2.
communication protocol. While first implemented by Futaba,
this protocol is generally found in numerous FrSky products as IV. ANTENNAS
well. It is not only a digital signal, but the SBUS is also An antenna is essentially a converter - a device whose
convenient due to its ability to support up to 18 channels on a function is to convert electrical energy into electromagnetic
single signal cable and is faster than PPM and PWM. radiation (or radio waves). There is no specific difference
DSMx is also a widely used radio protocol for 2.4 GHz between a transmitting and a receiving antenna. Drones can
spectrum RC transmitters and receivers. This protocol makes it have two types of antennas (based on radiation directivity):
possible to have multiple receivers/transmitters in the 2.4 GHz directional and omnidirectional. Directional antennas are most
spectrum without much interference. DSMx is in general two effective for emitting a signal over long distances and in a very
layers of multi-access techniques. DSM2 is based on code narrow direction. With omnidirectional antennas, the signal
division multiple access (CDMA) technology. This involves propagates in all directions over a relatively shorter range, but
two implementation phases: 1) the signal is spread out over a because of propagation in 360°, the losses are large.
wider frequency band and 2) each transmitter/receiver pair uses The directional antenna’s radiation pattern is conical, which
its own coding scheme to encrypt the signal. The channel in the means the tip of the cone sets out at the antenna. This fictive
2.4 GHz band (2.400 - 2.483 GHz) is used to transmit or cone is steadily becoming larger, as the distance from the
receive data. antenna is also increasing. A given object, if located within the
cone’s range, can detect the emitted signal or conversely,
III. REMOTE CONTROL SYSTEMS transmit its own signal, in the direction of the receiving
Futaba Advanced Spread Spectrum Technology (FASST) antenna. The optimal use for directional antennas is on the
and Advanced Continuous Channel Shifting Technology receiver-side, so that their position will enable the best signal
(ACCST) remote control systems are widely implemented in reception as sent by the video transmitter. The radiation pattern
UAVs. The FASST as well as the ACCST use frequency-shift of the omnidirectional antennas is torus-shaped and of equal
keying (FSK) modulation with frequency-hopping spread distribution, with the antenna’s position surrounded by the
spectrum (FHSS) and operate with in the 2.4 GHz band. torus. Except for the locations exactly above and below the
Moreover, the FASST implements Gaussian pulse-shaping antenna, the omnidirectional antenna’s radiation pattern is
filtering and direct sequence spread spectrum (DSSS) to the excellently dispersed in all directions around the antenna. Most
transmitted data. Since it has a greater number of channels as often, the omnidirectional antenna is associated with the video
compare with the FASST, the ACCST channels are arranged in transmitter, due to the fact that it seamlessly adjusts to
a more compact manner using a channel distance of 1.5 MHz. situations when the drone abruptly modifies how high it flies or
The data rate in the case of this technology is 100 kbps. in what direction.

Fig. 1. Generation of PAM/PWM/PPM signals [31]

Fig. 2. Overview of communication protocols used in drones (transmitter

The basic difference between FHSS and DSSS is in the way (Tx), receiver (Rx), flight controller (FC), electronic speed controller (ESC)).
of spreading data into the wider bandwidth. FHSS uses [32]

000020

Authorized licensed use limited to: Raytheon Technologies Corporation. Downloaded on December 19,2020 at 07:38:28 UTC from IEEE Xplore. Restrictions apply.
 11th IEEE International Conference on Cognitive Infocommunications – CogInfoCom 2020 • September 23-25, 2020 • Online on MaxWhere 3D Web

There are two forms of electromagnetic wave polarization: designed for it. These radars have a high resolution and
linear and circular (general case of elliptical polarization). The 3D tracking function − range, azimuth, elevation, and
main difference between linear and circular polarization is how velocity measurements of up to 200 targets
the radiation pattern propagates. A linear antenna generates simultaneously. In practice, it has been shown that they
waves along a vertical or horizontal plane (one plane). Because were able to detect small drones up to 5 km away. There
the wave is moving in one direction, linear polarization allows are different versions in practical realization: 8 − 12
more energy to be transported. But the problem arises in case GHz X-band and 12 − 18 GHz Ku-band, with applied
of obstacles which has a greater influence because of difficulty high-resolution radar signal processing technology to
a linear wave has in moving around an object that is within its locate drones accurately by applying an AESA (Active
plane of propagation. The circularly polarized antenna is more Electronically Scanned Array) technology to increase
adequate for situations that require barrier penetration because the maximum detection distance. It is often used
it emits waves in a corkscrew form (twists through both the (Scanning Surveillance Radar System) system and
vertical and horizontal planes). FMCW principle (Frequency Modulated Continuous
Situations especially convenient for circular polarization Wave) with a medium frequency of 94 GHz and an
are: flying close to large objects or in an enclosed environment output power of 100 mW. It has been shown in practice
(indoor parking spaces, stadiums), acrobatic flying (orientation that a bandwidth of up to 1 GHz allows a range
and angle are constantly changing), low altitude flying resolution of 15 cm.
(proximity flying). Situations particularly suitable for linear • Radio receiver − The passive RF detection system
polarization are: long-distance flying without any barriers, analyzes and identifies the remote control signal and
stable flying without too much roll or pitch and when antenna returns the signal from drones, detects the target and
size is taken into consideration, weight and durability are the warning. Wi-Fi is a common detection method as some
most important. drones have easily identifiable service set identifier
Antenna gain is a very important characteristic and (SSID). Detection distance can be more than 10 km,
represents a measure of how efficiently an antenna can transmit while the detection band is 20 MHz - 40 GHz.
electromagnetic waves in a certain direction and is expressed in • Audio (acoustic) sensor (microphone) − Sound-based
“dBi” (decibel isotropic). A drone FPV (First-Person View) techniques use arrays of microphones to extract the
antenna with a high gain is efficient at transferring unique acoustic signature of (micro) UAVs. Typically,
electromagnetic radiation and has a concentrated and narrow drones produce sounds ranging from 400 Hz to 8 kHz (
beam. If an antenna has a lower gain, it will feature a rounder due to the brushless DC motor ) [34]. Using different
radiation pattern. Despite the fact that transmitting a signal over audio analysis techniques, the sound of a drone can be
a long distance is not their strong suit, such antennas excel in separated from the environmental noise. A device for
high-quality short-range transmission and reception. drone detection is usually equipped with a Raspberry Pi
A conventional omnidirectional antenna covers the full core and with a microphone capable to pick up
hemisphere enabling it to detect the whole sky and receive up frequencies that humans can’t hear. During the
to 12 satellites simultaneously for defining the optimal processing of the collected data, the signal would be
navigation solution. A ‘smart’ antenna has the ability to focus compared with data in a database to separate it from
numerous narrow beams directly toward the satellites and track everyday noises and to confirm that it was actually a
them while showing nulls (locations where the antenna fails to drone. If the result is positive, the device would send a
receive a signal) at any interference [33]. Such devices notification that a drone was nearby. The sample is
considerably increase GPS reception even if there is jamming, compared to a proprietary database of drone signatures.
however, the cost more and are bulkier compared to The device uses acoustic technology to detect incoming
conventional antennas. drones from up to 130 meters and then send emails or
text messages to a monitoring service if a drone
approaches a specified perimeter.
V. ANTI-DRONE TECHNOLOGIES CATEGORIZATION
Along with the expansion of the industry, the use of drones • Optical sensor (camera) − This method implements
for malicious purposes also began. Therefore, many companies high-resolution cameras in order to capture drones in
follow trends and try to get into the counter drone industry. various background situations. Certain characteristics,
They have introduced many sorts of anti-UAV solutions. The including the Haar-like, local binary patterns (LBP),
counter drone technology can be broken up into two broad and histogram of gradients (HOG) can be extracted
categories: detectors and defenders. from the captured images. The next step is to feed these
features into boosted classifier cascades which are
Detectors are devices oriented to detect the presence of responsible for detection at multi-stage sequences with
drones and warning: increasing complexity. It is only the sequences that have
been subjected to the previous stage of analysis, which
• Radar − Low altitude radar is mostly used to detect
can then move on to the following stage.
ground, sea, and low-altitude targets. This method can
be problematic as some of the drones can be quite small Defenders are devices that provide some form of proactive
and require specialized radar to detect them. Small-scale function (shooting down or departure of drones):
drones can be detected by radar if the radar is specially

000021

Authorized licensed use limited to: Raytheon Technologies Corporation. Downloaded on December 19,2020 at 07:38:28 UTC from IEEE Xplore. Restrictions apply.
 P. Cisar et al. • Principles of Anti-Drone Defense

• Jammer − Its role is to eliminate the communication • High-energy microwave − The antenna generates
between the remote control and the drone, can microwaves and if a high-energy microwave hits a
selectively obstruct the communication channels, drone, it will be destroyed.
which covers the control (pilots) signal. It can also
disrupt navigation, the drone can lose the GPS signal, VI. JAMMERS
get lost from the planned route, or be forced to land.
More advanced drones use GPS receivers within the Jammers or commonly called signal blockers are devices
navigation and control loop which allows for some used to block and interfere with authorized radio
additional features: position hold, return to home, and communication signals. Jamming equipment interferes with the
autonomous flight. remote control circuits and navigation signals of drones and
forces them to land or return. The drone signal jammer blocks
• Anti-drone gun − It's a gun-like device that works by the communications between the enemy drone’s video,
jamming the signal between drone and pilot, in such a telemetry, command, control and navigation systems and the
way grounding unwanted UAVs. Also, it controls the operator [36]. Jammers are designed to disable a drone by
drone to land safely on the spot or returns to its point of blasting electromagnetic noise at active radio frequencies, and
origin. This device can disable drones operating on 5 at a power high enough to eliminate any communication
frequencies and has a range of up to 2500 meters. RF between the drone and its operator [37].
output power is up to 60 W.
The civilian drone jammers include two bands: Band 1 - 2.4
• Spoofer − The GPS spoofing attack is an act of fooling GHz and Band 2 - 433 MHz. The RF circuit in them consists of
the control system of the drone by sending it fake an electronic oscillator whose frequency is controlled by a
geographical coordinates. In this way, it is possible to voltage (VCO- voltage controlled oscillator) and of a linear
hijack the vehicle taking it away from the authentic power amplifier whose function is to amplify the output power
path. A similar attack is possible due to the leakage of of the VCO. The tuning circuit consists mainly of a triangular
GPS encryption rules, which is a common occurrence wave generator to generate the tuning signal along with a noise
for civilian equipment. generator to tune the VCO to the required RF jamming signal.
The power supply unit is designed to generate appropriate DC
• Hacking − Certain software exploit toolkits are focused voltages for the RF and the tuning circuits.
on hijacking drones targeting weak authentication and
encryption methods and allowing hackers to send Frequency jamming is a method that is even more subtle
commands to the victim’s drone. than the physical interception, and the anti-UAV defense
system (AUDS) is one such solution. This system scans the
• Snaggers (catchers) − a net placed under a drone, shot space on drones and jams their control signals using their own
from an air cannon or net-launching gun. high-powered radio signal.
• Drone catcher − There are situations on which An anti-drone rifle that uses targeted radio signals to disrupt
destroying drones might not be suitable. For example, drone controls (similarly to AUDS) ensures a more portable
when a drone is being used in restricted areas and the option. It currently has a range up to 1500 meters but may be
task is to land it without losing potentially accusing able to reach even farther in future versions. The standard
evidence. In this case, a special drone-catcher device output power of these devices with VCO technology is up to 75
becomes useful. This device is basically an anti-drone W and with high gain 12 dBi directional antenna.
drone, capable of catching an enemy UAV in a net from
up to 20 meters away. Characteristics of this device are:
controlled drone interception, equipped with gimbaled VII. DRONE HACKING
netgun, automatic drone detection, the target is selected Using certain communication weaknesses, various forms of
by the operator, automatic tracking without operator drone hacking are possible:
input, the unwanted drone is carried to a safe place and
1. A security researcher named Sasi infected his drone with
brought down by a parachute.
Maldrone malware. This malware creates a persistent
Destroyers are a special form of defenders which in the backdoor, enabling an attacker to remotely control the infected
physical sense eliminate drones: drone, cause it to drop from the sky, or hijack it to conduct
surveillance [38]. Main steps that have been taken: normal
• Laser − An optical device that emits a strong beam of flight controlled using API phone/computer, enabling autopilot,
light or radio energy at the drone. Relating to the power, sending code and waiting for a reverse TCP connection from
the laser can burn out or blind the hardware or camera. the drone, overriding the drone’s autopilot.
For example, ATHENA (Advanced Test High Energy
Asset) uses a specially developed tower with 30 kW 2. Some gadgets exploit DMSx, often used radio protocol
output power [35]. by most remote-controlled commercial drones, allowing the
hijacker to take control (for instance, the device called Icarus).
• Directed electromagnetic pulse − A generator emits a The exchanged public key (not encrypted) can be easily
pulse of energy that, if strong enough, can damage reconstructed after completing the binding process by
weakly shielded electronics. observing the protocol and using specific brute-force methods.
Furthermore, there is a timing attack vulnerability wherein the

000022

Authorized licensed use limited to: Raytheon Technologies Corporation. Downloaded on December 19,2020 at 07:38:28 UTC from IEEE Xplore. Restrictions apply.
 11th IEEE International Conference on Cognitive Infocommunications – CogInfoCom 2020 • September 23-25, 2020 • Online on MaxWhere 3D Web

synchronization to the target transmission is done and

transmission of offensive control packets in front of the target,
the receiver accepts information and rejects the targets.
3. There are devices featuring Xbee chips through which
the drone receives the commands. They specifically implement
a telemetry module featuring an Xbee radio chip. This module
modifies the Wi-Fi commands that a flight planning software
sent into low-frequency radio signals, to be further transmitted
Fig. 3. Using Xbee chips for drone hacking. [39]
to a different Xbee chip located on the drone itself. The
operator will therefore be able to control the drone even when
Superior defense techniques include the following:
considerably removed, further than otherwise possible.
• Consistency of navigation - inertial measurement unit
VIII. SPOOFING ATTACK (IMU-outputs gyroscopes, accelerometers and
magnetometers raw data) cross-check
When spoofing a certain receiver, signals from numerous
satellites must be reproduced by the potential attacker, • Polarization and angle-of-arrival discrimination
followed by the transmission of the spoofing signal with the
aim of capturing the target GPS receiver. If the original satellite • Cryptographic authentication
signals and spoofed signals cannot be distinguished by the The basic prevention against GPS attacks is the use of
targeted receiver, the spoofing has succeeded, the receiver is cryptographic techniques, as well as the use of authentication
fooled into seemingly appear in another position or another processes (WAAS (Wide Area Augmentation System) message
point of time [40]. The final result of this attack is changed authentication) between transmitters and receivers avoiding
(spoofed) path of the drone relative to the desired plan. The interferences of external sources. These techniques are
approximate range (radius) of a spoofing device is 2500 m. implemented only in the military sector and aviation.
In order to prevent potential misuses, the regular GPS As an efficient protection against spoofers, civilian users
signal may include an encrypted binary P-code, known as Y- can rely on numerous receivers with the ability to track several
code, which is transmitted on L1 and L2 frequencies. This GNSS (for instance GPS, GLONASS, Galileo, and BeiDou
encrypted code has a frequency of 10.23 MHz, and it does not simultaneously), given that it would be necessary for the
repeat over the course of an entire week. The P (precise) code attacker to generate and transmit all possible GNSS signals
is encrypted into the Y (or P(Y)) code when anti-spoofing simultaneously to spoof the target.
mode is on. Given that without being in possession of the
encryption key values, actually the generation of the Y-code is
IX. GPS SIGNAL JAMMING
not possible, therefore, neither is it possible to spoof a receiver
device into tracking Y-code. Signals stemming from the satellites are significantly
weakened so that the elimination of GPS coverage for a
Apart from other features, the GPS signals contain a C/A particular territory is possible by a 10 W jammer (or less) [33].
(coarse/acquisition) code, whose was supposed to help with A weak received signal power means that CDMA (Code
acquiring the Y-code, though it is currently applied with all Division Multiple Access) signals (not only GNSS), have
civilian GPS receivers. The C/A code is not encrypted, the become vulnerable to either unintended or deliberate
binary symbols (logical 1s and 0s) have a frequency of 1.023 interference (jamming). In case the level of interference
MHz, and the code itself (a unique sequence) repeats each surpasses a specific threshold, the GNSS signal is lost within
millisecond. Since the C/A code structure is openly published the interfering signal.
in a signal-in-space (SIS) interface specification, a half-way
capable attacker could reproduce it and then generate a spoofed Applying jamming methods enables the interruption of
version of the GPS signal. GPS signal reception. In such a case, the drone may lose the
ability to follow the route and calculate where it is, how high it
Spoofing attacks countermeasures – A reliable way of is and where it is flying to. Because the communication itself is
warding off spoofing is by means of directly tracking the also jammed, the operators lose their control over the drone,
encrypted Y-code. However, the precondition of such type of allowing the UAV to fly uncontrolled up to the point when it
protection is that a GPS receiver be equipped with a Selective smashes into the ground, makes a landing based on an
Availability Anti-Spoofing Module (SAASM). It is impossible emergency procedure, or simply runs out of fuel. The estimated
for the SAASM receivers to track Y-code unless they are range (radius) of such jamming equipment is 1500 m.
loaded with the functioning decryption key of that moment and
the modules have protection so as to avoid attackers’ reverse A solution to these attacks is the production of equipment
engineering. Only government authorized customers have that has spoofing or jamming detectors and uses encrypted
access to such receivers, the purchases and distribution of such GPS signals (like the military does).
devises are closely monitored. Another solution against jamming is using backup chips
Anti-spoofing countermeasures implemented in software on that contain gyroscopes, accelerometers, and a master clock -
receivers can be grouped into two categories: amplitude and Time and Inertial Measurement Unit (TIMU). It is able to
time-of-arrival discrimination. produce a record of vehicle movements, maintaining

000023

Authorized licensed use limited to: Raytheon Technologies Corporation. Downloaded on December 19,2020 at 07:38:28 UTC from IEEE Xplore. Restrictions apply.
 P. Cisar et al. • Principles of Anti-Drone Defense

coordinates accurate until the GPS connection is established

again.
Protection against jamming - There are several main
strategies to help overcoming interference:
1. Filtering the incoming traffic
2. Receiver equipped with IMU
3. Elimination of interfering signals using an adaptive
antenna array.

X. COGNITIVE APPROACH IN DRONE DETECTION TECHNOLOGY

The primary role in the successful defense of drones is their
detection. In this sense, the adequate radar system in
combination with an optical camera is effective and often used
Fig. 4. Block diagram of cognitive radar. [44]
in practice. Cognitive technologies have a significant place in
this area. The special form of radars, the so-called cognitive
radars, run on the basis of the perception-action cycle of • Active radars adjust to the environment - work in
cognition with the ability of sensing the environment, learning spectrally dense conditions and quickly modify the
from vital information about both target and background, while waveform of the transmitted signals to avoid
also adjusting the radar sensor to best meet the given mission’s interference with the original user of the channel.
requirements dictated by the goal to be accomplished [41]. The • Active radars adjust transmit signal parameters to
notion of cognitive radar was initially exclusively implemented achieve a defined level of target tracking performance.
for active radar.
• Passive radars are incapable of directly altering the
Cognitive radar (CR) “continuously learns about the transmitted waveforms, however, they can pinpoint the
environment through experience gained from interactions with best option to make both the detection and tracking
the environment, the transmitter adjusts its illumination of the better
environment in an intelligent manner, the whole radar system
constitutes a dynamic closed feedback loop encompassing the There are different advantages important for drone
transmitter, environment, and receiver” [42]. detection that could be attributed to CRs when compared to
traditional active radar (TAR) with similar antenna
The main characteristic of cognitive radar that makes it performance [45]:
stand out from the standard radar is the use of active feedback
between receiver and transmitter, which is presented in Fig. 4. 1. Increased detection range.
The classical principle of adaptivity is, in this way, extended to 2. Reduced time for parameter acquisition in target
the transmitter. tracking and faster information delivery
The key features of cognitive radar include the following 3. Enhanced information accuracy of the target position.
[43]: 4. Reduced risk in selecting a decision-making type in
the transmitter in conditions of real-time environmental
1. it makes informed decisions based on theoretic uncertainties and disturbances.
approach; 5. Better detection of smaller targets (such as drones
2. it has passive environmental and radar sensors; and stealth flying objects).
3. is is equipped with learning algorithms to make the 6. Increased capabilities of passive radar and multistatic
performance better and adjust to novel environmental radar systems in detection of stealth aircrafts better than
situations; conventional monostatic radars. Older stealth technologies
4. it includes a knowledge database full of environmental, reflected energy away from the line of sight, increasing radar
targets and other a priori information; cross-sections (RCS), which multistatic passive radars can
5. the waveform solution space, for known targets; detect.
6. receiver-to-transmitter feedback to mitigate 7. Use of waveforms difficult to be detected, including
clutter/interference and maximize target information. wideband signals.
8. Lower probability of being tricked by fake target
Cognitive radars are generally divided into active and energy (including decoys).
passive:
9. Use of built-in protection against misdirecting
Direction-of-Arrival (DOA) measurements.
10. More precise geolocalization due to networked CRs.
11. Enabling communication when other devices are
jammed.

000024

Authorized licensed use limited to: Raytheon Technologies Corporation. Downloaded on December 19,2020 at 07:38:28 UTC from IEEE Xplore. Restrictions apply.
 11th IEEE International Conference on Cognitive Infocommunications – CogInfoCom 2020 • September 23-25, 2020 • Online on MaxWhere 3D Web

Several advantages of cognitive radar (or perception-action

cycle (PAC) mechanism) over TAR important for drone
detection, when using the same signal processing technique
(Bayesian filtering), were reported in [44]. The studied
problem was the target reentry in space. It was shown that to
reach a Root-Mean-Square Error (RMSE) of the altitude, CR
took about 0.25 s and TAR 2.5 s (Fig. 5), which is
significantly faster. A similar conclusion can be derived in the
case of RMSE of the velocity (Fig. 6).
Eminent scholars from the Daegu Gyeongbuk Institute of
Science And Technology (South Korea) devised a drone
detection radar system in 2019 which functions within the Ku-
band (12 - 18 GHz), and they implemented high-resolution
Fig. 6. RMSE of velocity for Cognitive Radar and Traditional Active Radar
signal processing technology in order to exactly define the (both implemented with Cubature Kalman Filter) [44]
position of drones by using AESA technology to enhance the
maximum detection distance of the radar. Moreover, the given
system is capable of identifying and detecting drones with REFERENCES
higher precision in real-time by integrating GAN (Generative [1] K. Tretyakov (August 18, 2018). “How can I control a Drone via a
Adversarial Network)-based cognitive technology, which is Computer using Radio Signals?” Quoara. Retrieved August 20, 2020
also known as next generation deep learning algorithm [46]. from: https://www.quora.com/How-can-I-control-a-Drone-via-a-
AESA radar is able to transmit and receive signals Computer-using-Radio-Signals
independently because the modules in the antenna have [2] P. Baranyi, Á. Csapó, “Definition and Synergies of Cognitive
individual amplification and phase shifting. GAN represents Infocommunications”, Acta Polytechnica Hungarica, Vol. 9 No. 1, pp.
67-83, 2012.
the next generation's deep learning AI algorithm that generates
[3] P. Baranyi, A. Csapo, G. Sallai, “Cognitive Infocommunications
new data through the self-learning process with the same (CogInfoCom)”, Springer International Publishing, 208 pages, ISBN:
statistics as the training set, due to its neural network structure. 978-3-319-19607-7, 2015.
[4] J. Katona, A. Kovari, „The evaluation of BCI and pebl-based attention
CONCLUSIONS tests”, Acta Polytechnica Hungarica, vol 15, no 3, pp. 225-249, 2018.
[5] J. Katona, A. Kovari, “Examining the learning efficiency by a brain-
The paper presents the existence of a wide diversity in computer interface system”, Acta Polytechnica Hungarica, 2018, vol 15,
technologies and principles used in the production of drones. no 3, pp. 251-280, 2018.
With that in mind, it can be concluded that defending them is a [6] B. Lampert, A. Pongracz, J. Sipos, A. Vehrer and I. Horvath,
very complex task. It practically means that there is no single “MaxWhere VR-learning improves effectiveness over clasical tools of e-
reliable defense solution, but in order to achieve as much learning”, Acta Polytechnica Hungarica, vol. 15, no. 3, pp. 125-147,
2018.
success in that domain, it is advisable to combine multiple
[7] I. Horvath, A. Sudar, “Factors contributing to the enhanced performance
approaches. The primary role in the successful defense of of the maxwhere 3D VR platform in the distribution of digital
drones is their detection. In that sense, the appropriate radar information”, Acta Polytechnica Hungarica, vol. 15, no. 3, pp. 149-173,
system in combination with other sensors is very effective in 2018.
detection practice. The additional significance of this paper is [8] B. Berki, “Desktop VR as a Virtual Workspace: a Cognitive Aspect”
in the analysis of the technical characteristics of actual anti- Acta Polytechnica Hungarica, vol. 16, no. 2, pp. 219-231, 2019.
drone technologies. Familiarity with them enables the focus on [9] V. Kövecses-Gősi, “Cooperative Learning in VR Environment” Acta
certain solutions, depending on the specific protection Polytechnica Hungarica, vol. 15, no. 3, pp. 205-224, 2018.
requirements. Also, the paper pointed out that the cognitive [10] K. Bubnó and L. V. Takács, ”Cognitive Aspects of Mathematics-aided
approach in drone detection increases accuracy and Computer Science Teaching”, Acta Polytechnica Hungarica, vol. 16 no.
6, pp. 73-93, 2019.
adaptability, thanks to the application of learning algorithms in
[11] F. Erdos and R. Nemeth, “AMT-based Real-Time, Inter-Cognitive
radar technologies. Communication Model” Acta Polytechnica Hungarica, vol. 16, no. 6,
pp. 115-127, 2019.
[12] M. Gósy, V. Krepsz, “Evaluation of Cognitive Processes using
Synthesized Words: Screening of Hearing and Global Speech
Perception”, Acta Polytechnica Hungarica, vol. 15, no. 5, pp. 31-45,
2018.
[13] K. Chmielewska, A. Gilanyi “Educational Context of Mathability”, Acta
Polytechnica Hungarica, vol. 15, no. 5, pp. 223-237, 2019.
[14] E. Gogh, A. Kovari, “Experiences of Self-regulated Learning in a
Vocational Secondary School”, Journal of Applied Technical and
Educational Sciences, vol. 9, no. 2, pp. 72-86, 2019.
[15] A. Kővári, “ Adult education 4.0 and Industry 4.0 challenges in lifelong
learning,” PEDACTA, vol. 9, no. 1, pp. 9-16, 2019.
[16] T. Ujbanyi, G. Sziladi, J. Katona, A. Kovari, “ICT Based Interactive and
Smart Technologies in Education - Teaching Difficulties,” Proceedings
Fig. 5. RMSE of altitude for Cognitive Radar and Traditional Active Radar of the 229th International Conference on Education and E-learning
(both implemented with Cubature Kalman Filter) [44] (ICEEL), 2017, pp. 39–44.

000025

Authorized licensed use limited to: Raytheon Technologies Corporation. Downloaded on December 19,2020 at 07:38:28 UTC from IEEE Xplore. Restrictions apply.
 P. Cisar et al. • Principles of Anti-Drone Defense

[17] Zs. Namestovski, L. Major, Gy. Molnár, Z. Szűts, P. Esztelecki, G. [32] Learning RC (April 13, 2017). “Quadcopter Communication Protocols –
Kőrösi, “External Motivation, the Key to Success in the MOOCs Overview” Retrieved August 20, 2020 from: http://learningrc.com/rc-
Framework”, Acta Polytechnica Hungarica, vol. 15, no. 6 pp. 125-142, communication-protocols/
2018 [33] S. Cole (November 30, 2015). “Securing military GPS from spoofing
[18] B. Orosz, C. Kovács, D. Karuovic, Gy. Molnár, L. Major, V. Vass, Z. and jamming vulnerabilities”. Military Embedded Systems. Retrieved
Szűts, Zs. Námesztovszki, “Digital education in digital cooperative August 20, 2020 from: http://mil-embedded.com/articles/securing-
environments”, Journal of Applied Technical and Educational military-gps-spoofing-jamming-vulnerabilities/
Sciences/Alkalmazott műszaki és pedagógiai tudományos folyóirat, vol. [34] L. Hauzenberger, and E. Holmberg Ohlsson, “Drone detection using
9, no. 4. pp. 55-69, 2019 audio analysis,” M.S. thesis, Department of Electrical and Information
[19] Gy. Molnár, Z. Szűts, K. Biró, “Use of Augmented Reality in Learning”, Technology, Lund University, Sweden, June 2015. Retrieved August 20,
Acta Polytechnica Hungarica, vol. 15, no. 5 pp. 209-222, 2018 2020 from: http://lup.lub.lu.se/student-papers/record/7362609
[20] K. Hercegfi and M. Köles, “Temporal Resolution Capabilities of the [35] L. Dormehl (August 20, 2018). “7 amazing anti-drone technologies
Mid-Frequency Heart Rate Variability-based Human-Computer designed to swat UAVs out of the sky“. Digital Trends. Retrieved
Interaction Evaluation Method”, Acta Polytechnica Hungarica, vol. 16 August 20, 2020 from: https://www.digitaltrends.com/cool-tech/anti-
no. 6, pp. 95-114, 2019. drone-technologies/
[21] I. Poggi, F. D’Errico and L. Vincze, “Uncertain Words, Uncertain Texts. [36] G. Torbet (April 7, 2020). “7 Ways to Prevent Drones Infringing on
Perception and Effects of Uncertainty in Biomedical Communication”, Your Privacy“. MakeUseOf, Retrieved August 20, 2020 from:
Acta Polytechnica Hungarica, vol. 16, no.2, pp. 13-34, 2019. https://www.makeuseof.com/tag/8-ways-prevent-drones-infringing-
[22] Z. Balogh, M. Magdin, Gy. Molnár, “Motion Detection and Face privacy/
Recognition using Raspberry Pi, as a Part of, the Internet of Things”, [37] R. Joglekar (December 23, 2018). “4 strategies for stopping 'rogue'
Acta Polytechnica Hungarica, vol. 16, no.3 pp. 167-185, 2019 drones from flying in illegal airspace”. Retrieved August 20, 2020 from:
[23] I. Horváth, “MaxWhere 3D Capabilities Contributing to the Enhanced https://6abc.com/4-strategies-for-stopping-rogue-drones-from-flying-in-
Efficiency of the Trello 2D Management Software”, Acta Polytechnica illegal-airspace--/4949778/
Hungarica, vol. 16, no. 6, pp. 55-71, 2019. [38] R. Sasi (January 26, 2015). “Maldrone the First Backdoor for drones”.
[24] I. Horvath, A. Sudar, “Factors Contributing to the Enhanced Retrieved August 20, 2020 from:
Performance of the MaxWhere 3D VR Platform in the Distribution of http://garage4hackers.com/entry.php?b=3105
Digital Information”, Acta Polytechnica Hungarica, vol. 15, no. 3, pp. [39] N. Rodday, Hacking a Professional Drone, Black Hat, Asia 2016.
149-173, 2018. Retrieved August 20, 2020 from: https://www.blackhat.com/docs/asia-
[25] B. Berki, “Does Effective Use of MaxWhere VR Relate to the Individual 16/materials/asia-16-Rodday-Hacking-A-Professional-Drone.pdf
Spatial Memory and Mental Rotation Skills?”, Acta Polytechnica [40] P. Paganini (June 4, 2013). “Hacking Drones … Overview of the Main
Hungarica, vol. 16, no. 6, pp. 41-53, 2019. Threats”. Retrieved August 20, 2020 from:
[26] B. Berki, “2D advertising in 3D virtual spaces”, Acta Polytechnica https://resources.infosecinstitute.com/hacking-drones-overview-of-the-
Hungarica, vol. 15, no. 3, pp. 175-190, 2018. main-threats/#gref
[27] M. Kuczmann and T. Budai, “Linear State Space Modeling and Control [41] M.S. Greco, F. Gini, P. Stinco, and K. Bell, “Cognitive Radars: A
Teaching in MaxWhere Virtual Laboratory” Acta Polytechnica Reality?”, Cornel University, 2018, Retrieved August 20, 2020 from:
Hungarica, vol. 16, no. 6, pp. 27-39, 2019. https://arxiv.org/ftp/arxiv/papers/1803/1803.01000.pdf.
[28] K. Hercegfi, A. Komlódi, M. Köles, and S. Tóvölgyi, “Eye-Tracking- [42] S. Haykin, “Cognitive radar: A way of the future”, IEEE Signal
based Wizard-of-Oz Usability Evaluation of an Emotional Display Processing Magazine, vol. 23, no. 1, pp. 30-40, Jan. 2006.
Agent Integrated to a Virtual Environment”, Acta Polytechnica [43] A. F. Martone, “Cognitive radar demystified,” in URSI Radio Science
Hungarica, vol. 16, no. 2, pp. 145-162, 2019. Bulletin, vol. 2014, no. 350, pp. 10-22, Sept. 2014, doi:
[29] T. Budai, M. Kuczmann, “Towards a Modern, Integrated Virtual 10.23919/URSIRSB.2014.7910111.
Laboratory System” Acta Polytechnica Hungarica, vol. 15, no. 3, pp. [44] K. Krishnan, T. Schwering, and S. Sarraf, “Cognitive Dynamic Systems:
191-204, 2018. A Technical Review of Cognitive Radar”, Cornel University, arXiv:
[30] N. Goodman, Chapter 3 - Foundations of cognitive radar for next- 1605.08150, 2016, Retrieved August 20, 2020 from:
generation radar systems, Academic Press Library in Signal Processing, https://arxiv.org/pdf/1605.08150.pdf
Volume 7, Array, Radar and Communications Engineering, pp 153-195, [45] P. Labbé, “Cognitive Radars (CRs) Could Improve Target Engagement
2018, ISBN 978-0-12-811887-0, Retrieved August 20, 2020 from: Success Rate”, SPACOMM 2020: The Twelfth International Conference
https://doi.org/10.1016/C2016-0-00738-6 on Advances in Satellite and Space Communications, 2020.
[31] Electronics Coach, Pulse Position Modulation. Retrieved August 20, [46] S. Chin (July 22, 2019). „Radar system sensitive enough to detect drones
2020 from: https://electronicscoach.com/pulse-position-modulation.html 3 km away“.Fierce Electronics. Retrieved August 20, 2020 from:
https://www.fierceelectronics.com/sensors/radar-system-sensitive-
enough-to-detect-drones-3-km-away

000026

Authorized licensed use limited to: Raytheon Technologies Corporation. Downloaded on December 19,2020 at 07:38:28 UTC from IEEE Xplore. Restrictions apply.