V.V. Makarenkov1, G.N. Ulyanov2, I.S. Shamsiev3, A.A. Shatalov4, V.A. Shatalova5, N.A. Kupriyanov6, I.V. Kakaev7

1, 7 Military Space Academy named after A.F. Mozhaisky (St. Petersburg, Russia)
2–4 Mikhailovskaya Military Artillery Academy (St. Petersburg, Russia)
5 Bonch-Bruevich Saint Petersburg State University of Telecommunications (St. Petersburg, Russia)
6 Krasnodar Higher Military Aviation School of Pilots n. a. Hero of the Soviet Union A.K. Serov (Krasnodar, Russia)

Digital filters and multi-frequency sampling filter blocks are used in radio communications, antenna systems, image compression, analog voice communication systems, speech processing, and industrial digital audio equipment. Significant progress has been made in the field of multi-speed systems over the past few years. These include the development of decimation and interpolation filters, the analysis and synthesis of filter blocks (quadrature mirror filters), and the formulation of new sampling theorems. The use of parallel processing of multi-speed signals and adaptation in conditions of a priori uncertainty of the statistical characteristics of ultra-wide-band (UWB) signals should simplify data processing, reduce computational complexity in the information systems under consideration, and increase their throughput. One of these approaches is the use of polyphase decomposition of the input random process into frequency subbands in adaptive UWB filtering algorithms. However, the effect of superimposing signal spectra of different subbands, caused by the imperfection of the transfer functions of the filters and their difference from the desired rectangular shape, leads to significant losses in signal processing and limits the scope of the considered approach. This article discusses the use of an adaptive modified version of the polyphase decomposition, which allows taking into account the imperfection of the transmission characteristics of filters when decomposing the input joint venture into frequency subbands.

Goal – to synthesize an adaptive algorithm for multidimensional filtering of UWB signals based on the preliminary decomposition of the input random process into frequency subbands, taking into account the imperfection of the transmission characteristics of the filters.

An adaptive algorithm for multidimensional filtering of UWB signals is synthesized based on the decomposition of the input RP into frequency subbands using a modified version of the polyphase decomposition. This decomposition is based on the unitary transformation of the input random vector into a vector with mutually uncorrelated subvectors. As a result of this transformation, subsequent signal processing is performed in subbands with unknown statistical characteristics of input vector UWB random process. In this case, the covariance matrix of the input RP samples is reduced to a block-diagonal form. It is proposed to use the direct calculation method to adapt the algorithm in conditions of parametric a priori uncertainty.

An implementation of the developed algorithm is proposed in the form of a multidimensional adaptive bank filter with the calculation of unknown statistical characteristics of input vector UWB random process using the direct calculation method. The use of a modified version of the polyphase decomposition makes it possible to ensure the required accuracy of calculations of the considered algorithm by taking into account data processing distortions caused by imperfect transmission characteristics of filters. It is shown that the use of the direct calculation method makes it possible to increase the convergence rate of the adaptive algorithm in comparison with stochastic gradient methods.

Makarenkov V.V., Ulyanov G.N., Shamsiev I.S., Shatalov A.A., Shatalova V.A., Kupriyanov N.A., Kakaev I.V. Adaptive algorithm for filtering ultra-wide-band signals based on the input random process polyphase decomposition into frequency subbands. Science Intensive Technologies. 2026. V. 27. № 2. P. 5−21. DOI: https://doi.org/ 10.18127/j19998465-202602-01 (in Russian)

1. Egorova E.V., Aksyaitov M.H., Rybakov A.N. Cifrovaya obrabotka radiolokacionnyh signalov pri neskol'kih skorostyah peredachi dannyh. Naukoemkie tekhnologii. 2025. T. 26. № 3. S. 32–40. DOI 10.18127/j19998465-202503-04. EDN MOVRLE (in Russian).
2. Peteshov A.V., Pafikov E.A., Pyhtunkin A.V. Funkcional'naya skhema formirovatelya vektornogo zondiruyushchego signala. Izv. Tul'skogo gos. un-ta. Ser.: Tekhnicheskie nauki. 2019. № 10. S. 70–76. EDN KTBJDY (in Russian).
3. Platonov O.I., Lobov E.M. Cifrovye banki fil'trov analiza-sinteza signalov. Telekommunikacii i informacionnye tekhnologii. 2017. T. 4. № 1. S. 24–29. EDN ZBKJCh (in Russian).
4. Ivanov S.A., Starodubcev Yu.I. Teorema o predstavlenii nepreryvnogo mnogoparametricheskogo signala s nenulevoj dispersiej diskretnymi otschetami. Sistemy upravleniya, svyazi i bezopasnosti. 2021. № 2. S. 12–36. DOI 10.24412/2410-9916-2021-2-12-36. EDN KQLIWT (in Russian).
5. Gorbunov Yu.N. Teorema o stohasticheskoj diskretizacii izobrazhenij v radiolokacii i svyazi. Zhurnal radioelektroniki. 2018. № 10. S. 3. DOI 10.30898/1684-1719.2018.10.8. EDN FCIPXB (in Russian).
6. Makarenkov V.V., Shatalov A.A., Shatalova V.A., Kupriyanov N.A., Yakubovskij S.V. Adaptivnyj algoritm raspoznavaniya stacionarnyh i nestacionarnyh sluchajnyh processov v usloviyah parametricheskoj apriornoj neopredelennosti statisticheskih harakteristik signalov i pomekh. Nelinejnyj mir. 2025. T. 23. № 4. S. 12–29. DOI 10.18127/j20700970-202504-02. EDN WYNUPO (in Russian).
7. Vasyukov V.N., Lozovskij I.F., Morozov Yu.V. i dr. Prostranstvenno-vremennaya obrabotka shirokopolosnyh signalov radiolokacionnyh sistemah s adaptivnymi cifrovymi antennymi reshetkami. Novosibirsk: Novosibirskij gosudarstvennyj tekhnicheskij universitet. 2022. 240 s. (Monografii NGTU). ISBN 978-5-7782-4743-7. DOI 10.17212/978-5-7782-4743-7. EDN MLXGRJ (in Russian).
8. Lityuk V.I., Lityuk L.V. Metody cifrovoj mnogoprocessornoj obrabotki ansamblej radiosignalov. M.: Obshchestvo s ogranichennoj otvetstvennost'yu «SOLON-Press». 2019. 590 s. (Biblioteka inzhenera). EDN KRVOKK (in Russian).
9. Kalmykov M.I., Yurdanov D.V., Kononova N.V., Kalmykov I.A., Tyncherov K.T. Razrabotka mnogokanal'nogo sistolicheskogo algoritma vychisleniya teoretiko-chislovyh preobrazovanij signalov v pole Galua GF(M). Sovremennye naukoemkie tekhnologii. 2019. № 1. S. 58–62. EDN POIZUR (in Russian).
10. Oppengejm A., Shafer R. Cifrovaya obrabotka signalov. M.: Tekhnosfera. 2006 856 s. (in Russian).
11. Petrov D.A., Bogolyubov A.N. Primenenie polifaznogo razlozheniya dlya effektivnoj vychislitel'noj realizacii algoritma formirovaniya signala na osnove konechnomernogo obobshchennogo bazisa Vejlya-Gejzenberga. Zhurnal radioelektroniki. 2010. № 3. S. 2. EDN MSPQIZ (in Russian).
12. Vinokurov A.D., Kupriyanov N.A., Makarenkov V.V., Ul'yanov G.N., Shatalov A.A., Shatalova V.A. Osobennosti sozdaniya i primeneniya algoritmov mnogomernoj adaptivnoj fil'tracii vo vremennoj i chastotnoj oblastyah. Chast' 1. Fil'traciya. Naukoemkie tekhnologii. 2024. T. 25. № 3. S. 5–19. DOI 10.18127/j19998465-202403-01. EDN JNRXDY (in Russian).