Yu.E. Korchagin1, K.D. Titov2, O.N. Titova3

1-3 Voronezh State University (Voronezh, Russia)

2 MESC of the Air Force Academy named after Professor N.E. Zhukovsky and Yu.A. Gagarin (Voronezh, Russia)

1 korchagin@phys.vsu.ru; 2 titovkd@gmail.com; 3 onzavalishina@mail.ru

Designing optimal detectors requires a considerable amount of a priori information, the amount of which depends on the chosen optimality criterion. However, the shape of the signal to be detected must always be known. However, in radio monitoring applications, where a signal of unknown structure must be detected, it is not always possible to write its analytical expression for subsequent optimal processing. Similarly, in practical applications of communication, navigation, and control theory, where received signals must be completely known, propagation conditions, high subscriber density, and complex interference environments lead to distortion of their shape. Thus, when developing optimal detectors, one must either account for or predict possible distortions, or synthesize receiving devices for certain expected signals without taking actual propagation conditions into account. In practice, any optimal processing algorithm will be quasi-optimal. Methods for overcoming a priori uncertainty of a nonparametric type have been little studied to date, necessitating the study of neural network methods for signal processing. However, a detailed and accurate comparison of the performance characteristics of the problems solved using the proposed neural network algorithms with similar characteristics obtained using known optimal algorithms has either not been conducted, or the comparison involves algorithms that are not equivalent in meaning.

The aim of this work is to investigate neural network and hybrid algorithms for detecting ultra-wideband signals with an unknown modulation function shape and compare their effectiveness with the characteristics of known statistical (optimal and quasi-optimal) and energy detection algorithms.

A synthesis and analysis of neural network and hybrid devices supplemented with artificial neural networks has been performed, allowing for improved detector performance under conditions of high a priori nonparametric uncertainty or complex interference conditions. A comparison of the effectiveness of statistical (optimal and quasi-optimal), energy, neural network, and hybrid detection algorithms is performed, and recommendations for the use of one or another type of signal processing are formulated.

For practical applications, a hybrid detector can be recommended, and in order to train the neural network contained in it, it is necessary to use as large a training set as possible, including a wide variety of received signals with different signal-to-noise ratios.

Korchagin Yu.E., Titov K.D., Titova O.N. A hybrid algorithm for detecting ultra-wideband quasi-radio signals with unknown modulation function shape. Radiotekhnika. 2026. V. 90. № 2. P. 98−116. DOI: https://doi.org/10.18127/j00338486-202602-12 (In Russian)

1. Tihonov V.I. Optimal'nyj priem signalov. M.: Radio i svjaz'. 1983. 220 s. (in Russian).
2. Van Tris G.L. Teorija obnaruzhenija, ocenok i moduljacii. T. 1. Teorija obnaruzhenija, ocenok i linejnaja moduljacija: Per. s angl. Pod red. prof. V.I. Tihonova. M.: Sovetskoe radio. 1972. 744 s. (in Russian).
3. Akimov P.S., Bakut P.A., Bogdanovich V.A. i dr. Teoriya obnaruzheniya signalov. M.: Radio i svyaz'. 1984. 440 s. (in Russian).
4. Helstrom K. Statisticheskaja teorija obnaruzhenija signalov. M.: Izdatel'stvo inostrannoj literatury. 1963. 432 s. (in Russian).
5. Borisov V.I. i dr. Pomekhozashchishchennost' sistem radiosvyazi s rasshireniyem spektra signalov metodom psevdosluchaynoy perestroyki rabochey chastoty. M.: Radio i svyaz'. 2000. 384 s. (in Russian).
6. Trifonov A.P., Korchagin YU.E. Kvazipravdopodobnoye obnaruzheniye signalov s neizvestnymi formoy i momentami poyavleniya i ischeznoveniya. Vestnik Voronezhskogo gosudarstvennogo universiteta. Ser. Fizika. Matematika. 2015. № 4. S. 43-53. (in Russian).
7. Korchagin Ju.Je., Titov K.D. Sintez i analiz algoritmov obrabotki sverhshirokopolosnyh kvaziradiosignalov. Voronezh: Izdatel'skij dom VGU. 2019. 128 s. (in Russian).
8. Korchagin YU.E., Titov K.D., Petrov YU.G., Kosmodem'yanskiy S.A. Obzor primeneniya neyrosetevykh tekhnologiy v radiofi-zicheksikh prilozheniyakh. Sistemnaya inzheneriya i informatsionnyye tekhnologii. 2021. T. 3. № 1(5). S. 81-95. (in Russian).
9. Kolokol'tsev Ye.A., Myakin'kov A.V., Andriyanov A.V. Ispol'zovaniye sverkhshirokopolosnogo signala s povyshennoy chastotoy povtoreniya v prosvetnoy RLS dlya perimetrovoy okhrany. Sbornik trudov XXIII Mezhdunar. Nauch.tekhnich. konf. «Radiolokatsiya, navigatsiya, svyaz'». 2017. T. 3. S. 883-893 (in Russian).
10. Trifonov A.P., Bespalova M.B., Trifonov P.A., Gushchin I.V. Kvazipravdopodobnaya otsenka vremennykh parametrov posledovatel'nosti sverkhshirokopolosnykh signalov neizvestnoy formy pri vozdeystvii uzkopolosnykh pomekh. Izvestiya vuov. Radioelektronika. 2016. №3. T. 59. S. 30-39. (in Russian).
11. Razin'kov S.N. Tochnost' mestoopredeleniya izluchateley posledovatel'nostey sverkhshirokopolosnykh impul'sov v triangulyatsionnykh sistemakh. Fizika volnovykh protsessov i radiotekhnicheskiye sistemy. 2015. № 1. T. 18. S. 55-59 (in Russian).
12. Radchenko Yu.S., Sokhnyshev S.V. Prostranstvenno-vremennaya obrabotka sverkhshirokopolosnykh impul'snykh posledovatel'nostey. Izvestiya vysshikh uchebnykh zavedeniy Rossii. Radioelektronika. 2003. № 2. S. 19-28 (in Russian).
13. Afanas'yev O.V., Razin'kov S.N. Srednekvadraticheskaya oshibka pelengovaniya istochnikov sverkhshirokopolosnykh signalov pri prostranstvennoy rezhektsii uzkopolosnykh pomekh. Antenny. 2014. № 5. S. 14-18 (in Russian).
14. Korchagin Yu.E., Titov K.D., Golovatskaya Ye.E. Otsenka vliyaniya parametrov sverkhshirokopolosnykh radiosignalov na kharakteristiki ikh priyema i peredachi. Zhurnal radioelektroniki. 2023. № 2 (in Russian).
15. Kulikov E.I., Trifonov A.P. Ocenka parametrov signalov na fone pomeh. M.: Sovetskoe radio. 1978. 295 s. (in Russian).
16. Gonorovskiy I.S. Radiotekhnicheskiye tsepi i signaly: Uchebnik dlya vuzov. M.: Radio i svyaz'. 1986. 512 s. (in Russian).
17. Trifonov A.P., Korchagin Yu.E., Kondratovich P.A. Detection of a signal with unknown amplitude and duration. Radiophysics and Quantum Electronics. 2011. V. 54. № 5. P. 391-401.
18. Osovskiy S. Neyronnyye seti dlya obrabotki informatsii. Per. s pol'skogo I.D. Rudinskogo. M.: Finansy i statistika. 2002. 344 s. (in Russian).
19. Khaykin S. Neyronnyye seti i obucheniye mashin: Per. s angl. M.: ID “Vil'yams”. 2016. 688 s. (in Russian).
20. Wasserman P.D. Neural computing: theory and practice. Van Nostrand Reinhold Co.115 Fifth Ave. New York. NYUnited States. 1989. 230 p.
21. Goodfellow I., Bengio Y., Courville A. Deep Learning. The MIT Press. 2016. 800 p.
22. Raschka S., Mirjalili V. Python Machine Learning: Machine Learning and Deep Learning with Python, scikit-learn, and TensorFlow 2, 3rd Edition. Packt Publishing. 2019. 772 p.
23. Nikolenko S.I., Kadurin A.A., Arkhangel'skaya Ye.O. Glubokoye obucheniye. Pogruzheniye v mir neyronnykh setey. SPb: Piter. 2018. 480 s. (in Russian).
24. Kohonen T. Self-Organization and Associative Memory. 2nd edition. Springer Series in Information Sciences. 1988. 312 p.
25. Kohonen T. Self-Organizing Maps. 3rd ed. Berlin. Heidelberg: Springer-Verlag. 2001. 501 p.
26. Neyrokomp'yutery v obrabotke signalov. Рod red. A. I. Galushkina. M.: Radio i svyaz'. 2003. 228 s. (in Russian).
27. Carpenter G.A., Grossberg S. A massively parallel architecture for a self-organizing neural pattern recognition machine. Computer Vision, Graphics, and Image Processing. 1987. V. 37. P. 54-115.
28. Rumelhart D.E., Hinton G.E., Williams R.J. Learning representations by back-propagating errors. Nature. 1986. V. 323. P. 533-536.
29. Haykin S.O. Neural Networks: A Comprehensive Foundation. 2nd Edition. Pearson. 1998. 842 p.
30. Kingma D.P., Ba J. Adam: A Method for Stochastic Optimization. arXiv preprint arXiv:1412.6980. 2014.