A.V. Zyuzin1, E.I. Minakov2, M.S. Basherov3, A.V. Filonovich4

1 Yaroslavl Higher Military School of Air Defense (Yaroslavl, Russia)

2,3 Tula State University (Tula, Russia)

4 South-West State University (Kursk, Russia)

1 Aleksey.zyuzin@mail.ru, 2 EMinakov@bk.ru, 3 basherov@icloud.com, 4 filon8@yandex.ru

The paper discusses the problem of detecting small targets in radar systems using modern deep learning methods. The main attention is paid to the software implementation of algorithms in the MATLAB R2023 environment, which provides modeling of radar signals, the formation of microDoppler spectrograms and the construction of classifiers based on artificial neural networks. Based on the time-frequency representations, several convolutional neural network (CNN) architectures were trained and tested, including ResNet-18, ResNet-50, ResNet-101 and NASNet-Mobile. vectors (SVM). A combined architecture of a neural network classifier for recognizing small-sized targets from radar data is proposed. The architecture is based on the concept of convolutional networks with subsequent SVM classification and is designed to work with both one-dimensional (temporary) signal features and two-dimensional radio portraits (spectrograms). The simulation results showed that the best balance between accuracy and processing speed is achieved when using lightweight architectures, in particular NASNet-Mobile and ResNet-18, which provide accuracy up to 98,33% at moderate computing costs. The presented approach demonstrates the possibility of building highly efficient modules for detecting and classifying targets in real time, using adaptive neural network methods and optimizing for specific hardware limitations. The work is of an applied nature and the results can be used in the development of software systems for intelligent processing of radar data.

Zyuzin A.V., Minakov E.I., Basherov M.S., Filonovich A.V. Detection of small-sized objects using artificial neural net-works in radar applications. Electromagnetic waves and electronic systems. 2026. V. 31. № 1. P. 14−27. DOI: https://doi.org/10.18127/j15604128-202601-02 (in Russian)

1. Seidaliyeva U., Ilipbayeva L., Taissariyeva K., Smailov N., Matson E.T. Advances and Challenges in Drone Detection and Classification Techniques: A State-of-the-Art Review. Sensors. 2024. V. 24. № 1. P. 125. DOI 10.3390/s24010125.
2. Podgornov V.A., Naumenko M.Yu., Podgornov S.V. Aspects of countering a massive attacks by small-sized unmanned aerial vehicles. Issues of defense technology. Series 16: Technical means of countering terrorism. 2022. № 3-4(165-166). P. 62–71. (in Russian)
3. Kostromitsky S.M., Nefedov D.S., Khramenkov A.S., Chygryai V.G. Statistical models of radar cross section fluctuations of small-sized unmanned aerial vehicles. Bulletin of the Almaz – Antey Aerospace Defense Concern. 2023. № 3. P. 24–36. DOI 10.38013/2542-0542-2023-3-24-36. (in Russian)
4. Tryapkin N.S. Detection of small targets in a complex interference environment. [Electronic resource] – Access mode: https://elib.pnzgu.ru/files/eb/doc/cwdI9kavj0I1.pdf, date of reference 20.08.2025. (in Russian)
5. Pimenova M.B. Application of the Kalman filter in tracking problems of air objects. Politechnical student journal. 2019. № 12(41). P. 8. DOI 10.18698/2541-8009-2019-12-557. (in Russian)
6. Boriskevich I.A., Tsvetkov V.Yu. Tracking of low-sized targets from non-stationary camera based on covariance features and predictor. Reports of the Belarusian State University of Informatics and Radioelectronics. 2014. № 3(81). P. 33–39. (in Russian)
7. Puchkov A.Yu., Dli M.I., Lobaneva E.I., Vasilkova M.A. Predicting an object state based on applying the Kalman filter and deep neural networks. Software & Systems. 2019. V. 32. № 3. P. 368–376. DOI 10.15827/0236-235X.127.368-376. (in Russian)
8. Tao Wang Luk. Recognition of low-flying low-speed radar targets by trajectory signs: dis. ... Candidate of Technical Sciences. St. Petersburg: St. Petersburg State Electrotechnical University "LETI" named after V.I. Ulyanov (Lenin). 2024. 112 p. (in Russian)
9. Kazachkov E.A., Matyugin S.N., Popov I.V., Sharonov V.V. Detection and classification of small-scale objects in images obtained by synthetic-aperture radar stations. Bulletin of the Almaz – Antey Aerospace Defense Concern. 2018. № 1. P. 93–99.
10. Ananenkov A.E., Bakalov V.P., Konovaltsev A.V., Nuzhdin V.M., Rastorguev V.V., Skosyrev V.N., Sokolov P.V. Detection of small objects by ultrashort pulse radar. Collection of reports of the III All-Russian Scientific Conference "Ultra-broadband signals in radar, communications and acoustics". 2010. P. 202–207. (in Russian)
11. Jiang W., Wang Y., Li Y., Lin Y., Shen W. Radar Target Characterization and Deep Learning in Radar Automatic Target Recognition: A Review. Remote Sensing. 2023. V. 15. № 15. P. 3742. DOI 10.3390/rs15153742.
12. Molchanov P., Egiazarian K., Astola J., Harmanny R.I.A., de Wit J.J.M. Classification of small UAVs and birds by micro-Doppler signatures. European Radar Conference. Nuremberg, Germany. 2013. P. 172–175. DOI10.1017/S1759078714000282.
13. Wang Y., Feng C., Zhang Y., Ge Q. Classification of Space Targets with Micro-motion Based on Deep CNN. IEEE 2nd International Conference on Electronic Information and Communication Technology (ICEICT). Harbin, China. 2019. P. 557–561. DOI 10.1109/ICEICT.2019.8846441.
14. Brooks D., Schwander O., Barbaresco F., Schneider J.-Y., Cord M. Temporal Deep Learning for Drone Micro-Doppler Classification. 19th International Radar Symposium (IRS). Bonn, Germany. 2018. P. 1–10. DOI 10.23919/IRS.2018.8447963.
15. Belkov S.A., Malygin I.V. Use of the neural network for detection and identification of interference when receiving a spread spectrum signal. Physics of wave processes and radio engineering systems. 2019. V. 22. № 2. P. 37–43. DOI 10.18469/1810-3189.2019.22.2.37-43. (in Russian)
16. Vorobyov E.N. Investigation of distinctive features for recognition of small UAVs in passive radar. Bulletin of the Novgorod State University. 2019. № 4(116). P. 72–77. DOI 10.34680/2076-8052.2019.4(116).72-77. (in Russian)
17. Berdyshev V.P., Pomazuev O.N., Savelyev A.N., Smolkin M.A., Kopylov V.A., Loy V.V. Recognition of Classes and Types of Air Objects on Two-Dimensional Radar its Images in the Surveyed Radar. Journal of Siberian Federal University. Engineering & Technologies. 2019. V. 12. № 1. P. 18–29. DOI 10.17516/1999-494X-0102. (in Russian)
18. Pitaš K., Rejfek L., Nguyen T.N., Beran L., Tran P.T., Fišer O. FMICW Radar Target Classification By Neural Network. 30th International Conference Radioelektronika (RADIOELEKTRONIKA). Bratislava, Slovakia. 2020. P. 1–5. DOI 10.1109/RADIOELEKTRONIKA49387. 2020.9092342.
19. Gurbatov M.E., Litnovsky V.Ya. Analysis of the effectiveness of neural network algorithms for trajectory processing of radar information. Radiotekhnika. 2023. V. 87. № 3. P. 31−39. DOI 10.18127/j00338486-202303-03. (In Russian)
20. Slavutskiy A., Slavutskii L., Slavutskaya E. Neural Network for Real-Time Signal Processing: the Nonlinear Distortions Filtering. International Ural Conference on Electrical Power Engineering (UralCon). Magnitogorsk, Russian Federation. 2021. P. 84–88. DOI 10.1109/UralCon52005.2021.9559619.
21. Buryachenko V.V., Pahirka A.I. Object tracking with deep learning. Siberian Journal of Science and Technology. 2020. V. 21. № 2. P. 150–154. DOI 10.31772/2587-6066-2020-21-2-150-154.
22. Kostromitsky S.M., Nefedov D.S. Threshold selection for detection small-sized targets by "track-before-detect" method. Basis. 2021. № 2(10). P. 24–32. DOI 10.51962/2587-8042_2021_10_24. (in Russian)
23. Kishore M., kh Amitab, Kumar D.V., Debdatta K. Noise Reduction in Synthetic Aperture Radar Images Using Convolutional Neural Network. Journal of Computational and Theoretical Nanoscience. 2018. V. 15. № 8. P. 2572–2576. DOI 10.1166/jctn.2018.7500.
24. Chang Y.L., Anagaw A., Chang L., Wang Y.C., Hsiao C.Y., Lee W.H. Ship detection based on YOLOv2 for SAR imagery. Remote Sensing. 2019. V. 11. № 7. P. 786. DOI 10.3390/rs11070786.
25. Geng Zh., Xu Y., Wang B.N., Yu X., Zhu D.Y., Zhang G. Target Recognition in SAR Images by Deep Learning with Training Data Augmentation. Sensors. 2023. V. 23. № 2. P. 941. DOI 10.3390/s23020941.
26. Mahafza B.R. Radar Systems Analysis and Design Using MATLAB. Taylor and Francis Group. Ed. 4th. New York: Chapman and Hall/CRC. 2022. 690 p. DOI 10.1201/9781003051282.
27. Platt J.C. Probabilistic Outputs for Support Vector Machines and Comparisons to Regularized Likelihood Methods. Advances in Large Margin Classifiers. 1999. P. 61–74.