O.N. Chirkov1, S.Yu. Beletskaya2

1,2 FSBEI of HE “Voronezh State Technical University” (Voronezh, Russia)

1 chir_oleg@mail.ru; 2 omrt0705@mail.ru

Problem statement. The work is devoted to the development of a signal receiver with orthogonal frequency multiplexing (OFDM), based on convolutional neural networks (CNN), in order to improve the evaluation of the communication channel and increase the efficiency of data transmission. A CNN receiver architecture is proposed, where the input signal is a three-dimensional array containing received data, pilot symbol positions, and a preliminary channel estimate. The receiver evaluates the time offset, demodulates the signal, and evaluates the channel using a CNN. A 5G uplink channel with various propagation parameters (delay, Doppler shifts) is used as a model to test the performance of the proposed solution. The proposed Rx_CNN receiver is compared with the classical channel estimation methods (LS and LMMSE).

Purpose. Improving the accuracy of the evaluation of the communication channel and the performance of the OFDM receiver based on convolutional neural networks.

Results. The proposed CNN architecture significantly improves the accuracy of channel estimation and reduces the bit error rate (BER) compared to traditional methods. The use of all available data, including unknown information symbols, significantly improves the quality of reception of OFDM signals. The proposed OFDM receiver based on convolutional neural networks outperforms the LMMSE receiver with full channel knowledge by 2 dB.

Practical significance. The use of deep neural networks opens up new prospects for further improvement of fifth-generation (5G) wireless communication systems, allowing for better reliability and throughput. The study highlights the potential of neural networks in managing complex communication systems, paving the way for the creation of intelligent receivers and algorithms for adapting to changing environmental conditions.

Chirkov O.N., Beletskaya S.Yu. Improving the performance of the OFDM receiver based on convolutional neural networks. Radiotekhnika. 2025. V. 89. № 7. P. 73−77. DOI: https://doi.org/10.18127/j00338486-202507-14 (In Russian)

1. Zhang C., Patras P., Haddadi H. Deep learning in mobile and wireless networking: a survey. IEEE Communications Surveys Tutorials. 2019. V. 21. № 3. Р. 2224–2287.
2. Chirkov O.N., Antilikatorov A.B., Shkarovskij K.M., Tambovcev M.N. Optimizacii ocenki kachestva poluprovodnikovyh plastin s pomoshh'ju nejrotehnologij CNN. Vestnik Voronezhskogo gos. tehnich. un-ta. 2025. T. 21. № 1. S. 81-87. DOI: 10.36622/1729-6501.2025.21.1.012 (in Russian).
3. Svidetel'stvo o gosudarstvennoj registracii programmy dlja JeVM № 2024661883 (RF). Primenenie nejronnyh setej glubokogo obuchenija dlja povyshenija kachestva svjazi besprovodnyh kanalov svjazi s OFDM moduljaciej: № 2024661190. Chirkov O.N., Tambovcev M. N. Zajavl. 21.05.2024; opubl. 22.05.2024.
4. Chirkov O.N., Pirogov A.A. Primenenie algoritmov mashinnogo obuchenija v zadache ocenki besprovodnogo kanala svjazi s OFDM. Vestnik Voronezhskogo gos. tehnich. un-ta. 2023. T. 19. № 6. S. 164-169. DOI: 10.36622/VSTU.2023.19.6.025 (in Russian).
5. Sviridova I.V., Ostroumov I.V., Chirkov O.N. LDPC-dekoder na baze PLIS so sverhdlinnymi kodami. Vestnik Voronezhskogo gos. tehnich. un-ta. 2025. T. 21. № 1. S. 121-126. DOI: 10.36622/1729-6501.2025.21.1.018 (in Russian).
6. 3rd Generation Partnership Project; Technical Specification Group Radio Access Network; Study on channel model for frequencies from 0.5 to 100 GHz (3GPP TR 38.901 version 16.0.0 Release 16). ETSI, Sophia Antipolis Cedex. France. Oct. 2019.
7. You Y., Li J., Reddi S., Hseu J., Kumar S., Bhojanapalli S., Song X., Demmel J., Keutzer K., Hsieh C.-J. Large batch optimization for deep learning: Training BERT in 76 minutes. in International Conference on Learning Representations. 2020.
8. Chirkov O.N., Bashkirov A.V. Povyshenie jeffektivnosti ocenki kanala pri vysokoskorostnoj peredache dannyh v podvodnoj akusticheskoj svjazi s OFDM. Radiotehnika. 2024. T. 88. № 7. S.45-49. DOI: 10.18127/j00338486-202407-09 (in Russian).