Budapest University of Technology and Economics Department of Broadband Infocommunications and
Electromagnetic Theory
Passive Radar Filtering, Detection and Classication Algorithms
Tamás Pet®
Supervisor József Pávó, Rudolf Seller PhD Thesis Book
2020
2
Background an objectives of the research 1
Background an objectives of the research
Passive radar technology is a remote sensing method that is receiving sig- nicant attention in our time. Unique among radar detection methods, it is able to detect moving targets without its own radiation, using radio frequency sources already present in the environment. It success can be basically att- ributed to its combined civil and military application potentials. In terms of military use, we need to highlight its Low Probability of Intercept capability, which is provided due to the passive operation. It is also important that the utilized illuminator of opportunities (IoO) can be low frequency transmitters operating in the VHF (FM radio, DAB) and UHF (DVB-T, GSM, LTE ..) bands. This makes possible to detect targets with small radar cross section.
In civil applications, the most important advantages are the low production and operating costs, which are due to facts that there is no need for trans- mission and hence, there are no frequency license requirements. The principle of operation has been known for a long time, but the widespread use of prac- tical implementations has so far been expected. The main reasons for this were, that from the measurement point of view no illuminating sources with favorable properties were available, and also there were no available such a low-cost, high-performance signal processing systems that can handle the ext- reme computational needs required by technology. The ambiguity function of the broadcast transmissions now used as an illuminating source has a high level of side lobes due to the nature of the modulation, which makes it impos- sible to detect distant reections. Some of the ltering algorithms presented so far are tested exclusively in a simulated environment. Thus, the assessment of the applicability of the so far proposed ltering procedures remains far below what is expected, compared to a real eld measurement. The main reason of this is that correct modeling the bistatic clutter is still an extremely complex and dicult task. The performance of the proposed algorithms are often eva- luated in a less objective environment and using deceptive metrics. In these evaluations, researchers are use they own demonstrator systems with having dierent radio frequency parameters and they even use dierent illuminating
2 Background an objectives of the research
sources as well. The eligibility of existing results is further reduced by the fact that for demonstrations in most cases there are very few measurements available.
One of the main objectives of my research was to explore a system of necessary requirements that are essential for an objective assessment of the performance of the stages performing ltering and detection. To do this, it is essential to nd well-known and manageable metrics, and to identify the metric-relevant parameters of the used reference system. By applying the re- search results in practice, it becomes possible to study the so far proposed ltering procedures in a higher level, and also the authentic performance eva- luation of new methods.
The ltering methods can be divided into three fundamentally dierent categories according to their operating domain. Into the Doppler frequency, time and space domain based algorithms. The currently available methods, without exception, are only able operate in up to two domains, the domain spanned by the Doppler frequency and time. Filtering, performed by beam- forming in the space-domain is handled independently in all case. Therefore I have carried out further researches focusing on a ltering algorithm capable of exploiting information available simultaneously in space and time domains.
Another very promising application of the passive radar technology is the detection of small sized unmanned aerial vehicles. There is a growing demand for applicable solutions, as the rapid spread of devices increases the risk of criminal applications. Given that the electromagnetic spectrum is extremely saturated in densely populated areas, the use of passive radars may be an obvious choice to avoid interferences and frequency utilization restrictions.
Preliminary experiments have also revealed that the rotors of these small aircrafts produce a unique reection during rotation near their resonant fre- quency. Since the resonant frequencies due to the physical size of the rotor blades fall in the UHF band, where many broadcast transmissions operate, an additional need is induced to develop a passive radar detection. In modern
Experimental methodology applied for the research 3
radar systems, in addition to detection, there is an increasing demand on the implementation of target classication as well. This is especially true in sys- tems used to detect unmanned aerial vehicles, as birds moving in airspace can be easily confused with these ying vehicles due to their physical size. Previous experimental measurements show that the examination of the micro-Doppler spectrum of the reected signal allows the identication of detected objects.
Although the perception of the phenomenon has been reported by several re- searchers, the exact description and model of the reection generated during rotation is not yet known. One of the objectives of my research was to crea- te a model that properly describes the waveform of a reected signal in the resonant region. Using this model it becomes possible to detect, identify and estimate the condition of the rotating propeller with passive radar.
Experimental methodology applied for the research
In radar technology to produce reliable results from which long-term conclu- sions can be drawn, the carry out of real-world experiments is indispensable.
Therefore I have prepared a multichannel coherent receiver for my research work, which I could use to collect representative observations, primarily us- ing FM and DVB-T transmissions.[J2]. In order to implement this reference system, I have examined and formulated the requirements of those critical radio frequency parameters of the passive radar receiver stages, which have decisive importance on the performance of the ltering and detection stages.
Using this system, I have performed a number of experimental measurements on which I recorded data sets from targets moving on known paths, under controlled conditions. For my research on ltering algorithms, I used the re- cords from this dataset. The high-level automated evaluation environment I created is freely available from the ocial Python repository [19].
To investigate the reective properties of the propellers used on unmanned aerial vehicles, I have prepared a special measurement test bench. Using this
4 Experimental methodology applied for the research
test bench, I have carried out observations in a anechoic chamber, using a CW illumination, with multiple propellers, and adjusting their bearing angles.
New scientic results - Thesis I. 5
New scientic results
Thesis I.
After model creation, I have proposed the usage of such a performance metric, that is able to objectively characterize the ltering capabilities of dierent clutter ltering algorithms. Using this, I rst conducted an extensive research and quantitative analysis based on real measurement data, that can be utilized to application specically characterize the xed and adaptive clutter ltering algorithms operating in dierent domains.
Sub-thesis:
(i) Using quantitative research methods, I have performed a comparison on clutter lter algorithms operating in the time and space domain, exa- mining their performance and resource requirements. The information gathered about the behavior of the processes can greatly contribute to the proper selection of the clutter ltering algorithm that best ts to the given application.
(ii) I have proposed an ecient acceleration method for the batched clutter ltering algorithms operating in the time domain, for which I have used the method of minimum redundancy estimation. With the introduced modication, in practical cases, in addition to an indierent loss, the computational requirements of the algorithms can be reduced by up to two orders of magnitude.
(iii) I have developed an adaptive path interference suppression method ope- rating in the analog domain, that eectively reduce the critical hardware requirements of the digitizing stage of passive radar systems.
6 New scientic results - Thesis I.
(iv) I have developed an experimental passive radar reference environment, which could be used to objectively measure the performance of ltering algorithms. Using this system, I have performed experimental measure- ments in real environment, which can be used as a reference for the extensive evaluation of the ltering algorithms.
• Publications related to the thesis: [B1],[J2],[C1],[C2],[C3],[C4]
• Known references of the publications related to the thesis: [1], [2], [3], [4], [5], [6], [7], [8], [9]
New scientic results - Thesis II. 7
Thesis II
I have developed an adaptive clutter ltering algorithm which is able to ope- rate simultaneously both in the space and time domains (STAC). Based on real measurements, the proposed algorithms outperforms currently available algorithms that traditionally operate domain-independently.
• Publication related to the thesis: [J1]
• Known reference of the publication related to the thesis: [16]
8 New scientic results - Thesis III.
Thesis III.
I have proposed a new type of measurement method for the detection, classi- cation and state estimation of small rotary-wing unmanned aerial vehicles, which is uniquely based on the resonant reectivity of the propellers. Due to the typical dimensions and applicability, this method closely matched to the operation of UHF band passive radars.
Sub-thesis:
(i) I have developed a model to describe the reected signal of rotating pro- pellers made from conductive materials near the resonant frequency. I have veried the created model with real measurements for special cases.
(ii) I have shown that at the resonant frequency, in the case of several simul- taneously rotating propellers, the characteristic state of the system can be determined inversely by the spectral evaluation of the reected signal.
(iii) I have proposed a signal processing method for passive radar systems that can be used to detect and classify unmanned aerial vehicles with multiple propellers using resonant illumination.
• Publications related to the thesis: [J3],[C5],[C6]
• Known references of the publications related to the thesis: [10], [11], [12], [13], [14], [15], [17]
Utilization of the research results 9
Utilization of the research results
From among the results of the carried out researches, the application of the minimum redundancy estimation for the batched time domain algorithms, has outstanding practical signicance. Thanks to the algorithm, systems with a smaller signal processing capacity will also be able to perform quasi-real-time clutter ltering.
The RTL-SDR is a low-budget broadband radio receiver that is basically de- signed to receive DVB-T signals, but can also function as a general-purpose software radio with minor modications. Thanks to its low cost of produc- tion, this receiver is widespread worldwide [18]. The next generation of this well-known receiver is the Kerberos SDR with four channels, which is already capable of coherent reception. One of the main objectives of implementing multi-channel operation is to support passive radar technology. Since these receiver stages typically operate in an embedded environment, the computa- tional requirements of the algorithms are critical. Therefore, the passive radar signal processing chain implemented on KerberosSDR applies the Wiener SMI MRE clutter ltering method that I have proposed. With the publication of the system, this clutter ltering algorithm has also become open source, in the hope that the methods found as a result of my research will reach a wide range of radar development professionals. The Python implementation of the algorithm is available from the ocial Python repository, from the Python Package Index (PyPI) as pyAPRIL [19].
As of 2018, the Defense Research Institute of the Hungarian Ministry of De- fense started the development of a DVB-T based passive radar in cooperation with the Microwave Remote Sensing Laboratory of the SzHVT department of BME. The main objectives of the system, designed for nationwide coverage, include long-range, real-time detection of ground and air targets. The support this the hardware capable of the multi-channel coherent operation. Due to the requirements of the system, great emphasis was placed on selecting the app- ropriate clutter ltering algorithms. Therefore, to maximize the performance of the ltering, the STAC algorithm found as a result of my research was
10 Utilization of the research results
implemented in the system. By using this ltering procedure, the 12-element antenna system can be utilized with greater eciency. In addition, to support the rapid update rate, the system utilize the minimum redundancy estimati- on version of the ECA-S algorithm, that I have proposed as a results of my research.
The results of the researches have been utilized in the project of the Euro- pean Defence Agency (EDA), that aimed to developed a passive radar sys- tem, Multichannel passive ISAR for military application (MAPIS) (no. B-1359 IAO2 GP). The research involved5countries9institutes, including BME and SZTAKI from Hungary.
Publications related to the PhD thesis 11
Publications related to the PhD thesis
Book chapter
[B1] Tamás Pet®, Rudolf Seller: Topics in Radar Signal Processing: Adap- tive Clutter Cancellation Techniques for Passive Radars, InTech Open Access Publisher, 2018. pp. 139-166.
Papers published in journals
[J1] Tamás Pet®, Rudolf Seller: Space-Time Adaptive Cancellation in Pas- sive Radar Systems, International Journal of Antennas and Propagation, 2018
[J2] Tamás Pet®, Rudolf Seller: Quad Channel Software Dened Receiver for Passive Radar Application, Archives of Electrical Engineering, 2017, 66, (1), 12p
[J3] Károly Marák, Tamás Pet®, Sándor Bilicz, Szabolcs Gyimóthy, József Pávó: Electromagnetic simulation of rotating propeller blades for radar detection purposes, IEEE Transactions on Magnetics, 54, (3)
Papers published on conferences
[C1] Tamás Pet®, Levente Dudás, Rudolf Seller: DVB-T based passive radar, 24th International Conference Radioelektronika, Bratislava, Slovakia, April, 15-16, 2014
[C2] Tamás Pet®, Rudolf Seller: Time Domain Filter Comparison in Passive Radar Systems, Int. Radar Symposium, Prague, Czech Republic, June 2017, 10p
12 Publications related to the PhD thesis
[C3] Tamás Pet®, Rudolf Seller: Quad Channel DVB-T Based Passive Radar, 17th International Radar Symposium (IRS), Kraków, Poland, May 10- 12, 2016, 4p
[C4] Tamás Pet®, Levente Dudás, Rudolf Seller: Analog Direct Path Inter- ference Suppression for FM Based Passive Radars, 28th International Conference Radioelektronika, Prague, Czech Republic, 19-20 April 2018, 4p
[C5] Tamás Pet®, Sándor Bilicz, László Sz¶cs, Szabolcs Gyimóthy, József Pá- vó: The Radar Cross Section of small propellers on Unmanned Aerial Vehicles, 10th European Conference on Antennas and Propagation, Eu- CAP 2016, Davos, Switzerland, 5-10 April 2016
[C6] Tamás Pet®, Károly Marák, Sándor Bilicz, József Pávó: Experimental and Numerical Studies on Scattering from Multiple Propellers of Small UAVs, Proceedings of the 12th European Conference on Antennas and Propagation (EUCAP), London, UK, 9-13 April 2018
References 13
Hivatkozások
[1] Yuqi Liu, Jianxin Yi, Xianrong Wan, Xun Zhang et al.:Evaluation of Clut- ter Suppression in CP-OFDM-Based Passive Radar IEEE Sensors Jour- nal, 19 (14), 2019, pp. 55725586
[2] Ivan Low, Xavier Chia, Shao Ying Huang: A Low-Cost DVB-T2-based Passive Bistatic Phased Array Radar for Target Detection and Localisati- on, IEEE Asia-Pacic Microwave Conference (APMC) 10-13, December, 2019
[3] Junichi Honda, Takuya Otsuyama: Positional Estimation of Obstacles and Aircraft by Using ISDB-T Signal Delay, IEEE 30th International Con- ference on Advanced Information Networking and Applications (AINA) 2016, Crans-Montana, Switzerland, 23-25 March 2016, pp. 346351 [4] Junaid Abdullah, Syed Ali Hassan: Geometry optimization for WiFi-based
indoor passive multistatic radars, 13th International Wireless Communi- cations and Mobile Computing Conference (IWCMC), Valencia, Spain, 23-30 June 2017, pp. 2068-2017
[5] Junichi Honda, Takuya Otsuyama: Feasibility Study on Aircraft Position- ing by Using ISDB-T Signal Delay, IEEE Antennas and Wireless Propa- gation Letters, 15, 2016 pp. 17871790
[6] Dinghe Wang, Qinglong Bao, Ruiqi Tian et al.:Bistatic weak target de- tection method using non-cooperative air surveillance radar, Journal of Systems Engineering and Electronic, 26 (5), 2015, pp 954963
[7] Panhe Hu, Qinglong Bao, Zengping Chen: Weak Target Detection Method of Passive Bistatic Radar Based on Probability Histogram, Mathematical Problems in Engineering, 2018 (8243686), 2018, pp. 110
[8] Rafael Gonçalves Licursi de Mello, Fernando Rangel de Sousa, Cynthia Junqueira: SDR-based radar-detectors embedded on tablet devices, 2017
14 References
SBMO/IEEE MTT-S International Microwave and Optoelectronics Con- ference (IMOC), Aguas de Lindoia, Brasil, 27-30 August 2017
[9] Yucheng Yi, Xianrong Wan, Jianxin Yi, Xiaomao Cao: Polarization Diver- sity Technology Research in Passive Radar Based on Subcarrier Processing, IEEE Sensors Journal, 19 (5), 2018, pp. 17101719
[10] Jan Farlik, Miroslav Kratky, Josef Casar, and Vadim Stary: Multispectral Detection of Commercial Unmanned Aerial Vehicles, Sensors (Basel), 19 (7), 2019
[11] Kupryashkin I.F., Sokolik N.V.: Algorithm of Signal Processing in the Radar System with Continuous Frequency Modulated Radiation for Detec- tion of Small-Sized Aerial Objects, Estimation of Their Range and Velo- city, Journal of the Russian Universities. Radioelectronics, 2019,(1), pp.
3955
[12] Arne Schröder, Uwe Aulenbacher, Matthias Renker, Urs Böniger, Roland Oechslin, Axel Murk, Peter Wellig : Numerical RCS and micro-Doppler investigations of a consumer UAV, Target and Background Signatures II.
9997 (999704), 2016
[13] Alexey V. Khristenko, Maxim O. Konovalenko, Mikhail E. Rovkin et al.: Magnitude and Spectrum of Electromagnetic Wave Scattered by Small Quadcopter in X -Band IEEE Transactions on Antennas and Propagation, 66 (4), 2018
[14] A. V. Khristenko, M. O. Konovalenko, M. E. Rovkin et al: A system for measurement of electromagnetic wave scattered by small UAVs 2017 Inter- national Siberian Conference on Control and Communications (SIBCON), Astana, Kazakhstan , 29-30 June 2017
[15] Min Guo, Yi Lin, Zhanshan Sun, Yunqi Fu: Research on Monostatic Radar Cross Section Simulation of Small Unmanned Aerial Vehicles 2018 International Conference on Microwave and Millimeter Wave Technology (ICMMT), Chengdu, China, 7-11 May 2018
References 15
[16] D. I. Lekhovytskiy, V. P. Riabukha, A. V. Semeniaka, D. V. Atamanskiy, Ye. A. Katiushyn: Protection of Coherent Pulse Radars against Combi- ned Interferences. 1. Modications of STSP Systems and their Ultimate Performance Capabilities, Radioelectronics and Communications Systems, July 2019, 62, (7), pp. 311-341
[17] Evgenii Vorobev, Vladimir I. Veremyev, Nikolay I. Tulenkov: Experimen- tal DVB-T2 Passive Radar Signatures of Small UAVs, 2019 Signal Pro- cessing Symposium (SPSympo), Poland, 17-19 September 2019
[18] RTL-SDR ocial website: https://www.rtl-sdr.com/
[19] pyAPRIL in the PyPi repository: (https://pypi.org/project/pyAPRiL/)
