RECURRENT NEURAL NETWORK FOR CONTROLLING THE SPECTRUM WIDTH OF A NON-STATIONARY RANDOM SIGNAL

DOI: 10.47026/1810-1909-2023-2-5-17

УДК 621.316.91

ББК 31.247

Dmitry Yu. ALYUNOV

Key words

artificial neural networks, recurrent neural networks, signal processing in electrical engineering, autocorrelation, non-stationary random process, transients

Abstract

The purpose of the study is to develop a recurrent neural network for detecting the moment of the beginning of the transient process of a random non-stationary signal in a sliding window. The possibility of using the apparatus of artificial neural networks (ANN) in analyzing non-stationary random processes is investigated. Rapid detection of the moment at which a non-stationary random process changes its behavior is an urgent task of electrical engineering.

Materials and methods. A comparison is made between the use of the autocorrelation function and a neural network algorithm based on a recurrent neural network to control non-stationary noise.

Results. The novelty of the study consists in developing a new algorithm for estimating the spectrum width of a non-stationary random signal based on the use of the ANN apparatus. It is shown that recurrent neural networks are capable of processing the original signal. They do not require special pre-processing and data preparation. A study of the quality of ANN operation depending on the parameters of the signals and the size of the sliding window was carried out. The ways to improve the architecture of the neural network and to enrich the data to improve the quality of the classifier are proposed.

Conclusions. It was found that there is an optimal ratio between the time of detecting a change in signal parameters and the size of the sliding window, which imposes restrictions on the choice of the latter and the method of applying the trained neural network model.

References

1. Averkin A.N., Yarushev S.A. Sravnitel’nyi analiz metodov prognozirovaniya vremennykh ryadov na osnove neironnykh setei i regressionnogo analiza [Comparative analysis of time series forecasting methods based on neural networks and regression analysis]. Sistemnyi analiz v nauke i obrazovanii, 2015, no. 2, pp. 34–49.
2. Andreev O.N., Vasil’eva L.N. Neirosetevaya obrabotka signalov s nelineinymi iskazheniyami v «skol’zyashchem vremennom okne» [Neural Network Processing of Signals with Nonlinear Distortions in a “Sliding Time Window”]. Vestnik Chuvashskogo universiteta, 2022, no. 1. pp. 5–
3. Gazizov D.I. Obzor metodov statisticheskogo analiza vremennykh ryadov i problemy, voznikayushchie pri analize nestatsionarnykh vremennykh ryadov [Overview of methods for statistical analysis of time series and problems arising in the analysis of non-stationary time series]. Nauchnyi zhurnalб 2016, no. 3(4), pp. 9–
4. Getmanov V.G. Tekhnologiya spektral’no-vremennogo analiza nestatsionarnykh kolebatel’nykh signalov mekhanicheskikh sistem [Technology of spectral-temporal analysis of non-stationary oscillatory signals of mechanical systems]. Problemy mashinostroeniya i avtomatizatsii, 2010, no. 2, pp. 121–129.
5. Golovko V.A., Khatskevich M.V., Brich A.L. Metod prognozirovaniya vremennykh ryadov na osnove mnogosloinogo perseptrona [Time series forecasting method based on multilayer perceptron]. Vestnik Brestskogo gosudarstvennogo tekhnicheskogo universiteta. Fizika, matematika, informatika, 2013, vol. 5(83), pp. 2–6.
6. Zhurilova O.E., Bashkirov A.V., Beletskaya S.Yu. Sovremennye metody i zadachi spektral’nogo analiza signalov: kratkii obzor i sravnenie [Modern Methods and Problems of Signal Spectral Analysis: Brief Review and Comparison]. Vestnik Voronezhskogo gosudarstvennogo tekhnicheskogo universiteta,. 2019, vol. 15, no. 2, pp. 128–131.
7. Krivosheev V.I., Lupov S.Yu. O nekotorykh vozmozhnostyakh i problemakh sovremennogo tsifrovogo spektral’nogo analiza [On some possibilities and problems of modern digital spectral analysis]. Vestnik Nizhegorodskogo universiteta im. N.I. Lobachevskogo, 2011, no. 5-3, pp. 109–117.
8. Labintsev E. Metriki v zadachakh mashinnogo obucheniya [Metrics in Machine Learning Problems]. Available at: https://habr.com/ru/company/ods/blog/328372.
9. L’vova K.I. Spektral’noe otsenivanie granits segmentov EEG dlya al’fa-aktivnosti mozga cheloveka [Spectral estimation of the boundaries of EEG segments for alpha activity of the human brain]. In: XII vserossiiskoe soveshchanie po problemam upravleniya VSPU-2014: sbornik trudov konferentsii (16–19 iyulya 2014 g.) [XII Russ. Meeting on Management Problems VSPU-2014: Proceedings of the Conference (July 16–19, 2014)]. Moscow, 2014, pp. 6742–6749.
10. Mokhov V.G., Dem’yanenko T.S. Modelirovanie osnovnykh parametrov rynka na sutki vpered i indeksa balansiruyushchego rynka [Modeling the main parameters of the day-ahead market and the balancing market index]. Vestnik Yuzhno-Ural’skogo gosudarstvennogo universiteta. Ser. Matematicheskoe modelirovanie i programmirovanie, 2015, no. 8.4, pp. 127–130.
11. Proskuryakov A.Yu. Model’ prognozirovaniya vremennykh ryadov na trekhsloinom perseptrone v informatsionnoi sisteme monitoringa zagryaznyayushchikh vybrosov [Time series forecasting model on a three-layer perceptron in an information system for monitoring polluting emissions]. Radiotekhnicheskie i telekommunikatsionnye sistemy, 2014, no. 2(14), pp. 57–63.
12. Romashikhin M.Yu., Gorbokonenko P.A. Realizatsiya okonnogo preobrazovaniya Fur’e na FPGA dlya spektral’nogo analiza nestatsionarnykh signalov [FPGA Windowed Fourier Transform Implementation for Spectral Analysis of Non-Stationary Signals]. In: Innovatsionnye tekhnologii, v elektronike i priborostroenii: sbornik dokladov Rossiiskoi nauchno-tekhnicheskoi konferentsii (05–12 aprelya 2021 goda) [Proc. of Russ. Sci. Conf. «Innovative technologies in electronics and instrument making»]. Moscow, 2021, pp. 177–182.
13. Slavutskaya E.V., Slavutskii L.A. O vybore struktury iskusstvennykh neirosetei i algoritmov analiza psikhodiagnosticheskikh dannykh [On the choice of the structure of artificial neural networks and algorithms for the analysis of psychodiagnostic data]. Kazanskii pedagogicheskii zhurnal, 2020, no. 5(142), pp. 202–
14. Sholtanyuk S.V. Sravnitel’nyi analiz neirosetevoi i regressionnykh modelei prognozirovaniya vremennykh ryadov [Comparative Analysis of Neural Network and Regression Models for Time Series Forecasting]. Tsifrovaya transformatsiya, 2019, no. 2, pp. 60–
15. Andreev O.N., Slavutskii L.A., Slavutskiy A.L. Neural network in a “sliding window” for power grids signals structural analysis. In: IOP Conference Series: Earth and Environmental Science, 2022, vol. 990, 012054. DOI: 10.1088/1755-1315/990/1/012054.
16. Dolphin R. LSTM Networks, a Detailed Explanation. Available at: https://towardsdatascience.com/lstm-networks-a-detailed-explanation-8fae6aefc7f9.
17. Fadıl C.M., Murat E., Mehmet H.Ö. Comparison of big data time series analysis methods. In: Proc. of the V International scientific and vocational studies congress – engineering (Bilmes en 2020). Ankara, 2020, Available at: https://www.researchgate.net/publication/348548541_COMARISON_OF_BIG_DATA_TIME_SERIES_ANALYSIS_METHODS.
18. Slavutsky A.L., Slavutsky L.A., Slavutskaya E.V. Neural Network for Real-Time Signal Processing: The Nonlinear Distortions Filtering. In: Proc. Of the International Ural Conference on Electrical Power Engineering. Magnitogorsk, 2021, pp. 84–88.
19. Yadarova, O.N., Okhotkin G.P., Slavutskii L.A. A control system of a fan drive based on airflow doppler ultrasonic control. Russian Electrical Engineering, 2017, vol. 88, no. 7, pp. 420–422. DOI: 10.3103/S1068371217070173.
20. Zhang X., Liu Y., Yang M. et al. Comparative Study of Four Time Series Methods in Forecasting Typhoid Fever Incidence in China. PLoS ONE, 2013, vol. 8(5). DOI: 10.1371/journal.pone.0063116. Available at: https://www.researchgate.net/publication/236644248_Comparative_Study_of_Four_Time_Series_Methods_in_Forecasting_Typhoid_Fever_Incidence_in_China.

Information about the author

Dmitry Yu. Alyunov – Senior Lecturer, Department of Computer Technologies, Chuvash State University, Russia, Cheboksary (aldmitry89@gmail.com; ORCID: https://orcid.org/0000-0001-8673-3683).

For citations

Alyunov D.Yu. RECURRENT NEURAL NETWORK FOR CONTROLLING THE SPECTRUM WIDTH OF A NON-STATIONARY RANDOM SIGNAL. Vestnik Chuvashskogo universiteta, 2023, no. 2, pp. 5–17. DOI: 10.47026/1810-1909-2023-2-5-17 (in Russian).

Download the full article