Journal of Radio Electronics. eISSN 1684-1719. 2026. №3

Contents

Full text in Russian (pdf)

Russian page

DOI: https://doi.org/10.30898/1684-1719.2026.3.14

COMPARATIVE ANALYSIS OF THE EFFECTIVENESS
OF SIGNAL DETECTION AND PARAMETER ESTIMATION
USING CLASSICAL METHODS AND NEURAL NETWORKS

L.G. Dorosinskiy

Ural Federal University named after the first President of Russia B.N. Yeltsin,

620002, Russia, Yekaterinburg, Mira Str., 19

The paper was received February 26, 2026.

Abstract. The problem of detecting a random Gaussian signal in additive white Gaussian noise with unknown SNR and unknown signal correlation is studied. Energy detection, nonparametric methods (kNN and Parzen windows), and a neural-network-based detector are compared. It is shown that noise whitening followed by principal component analysis significantly improves the performance of nonparametric detectors. For strongly correlated signals, these methods achieve performance close to the optimal Bayesian detector and may outperform neural networks. The detector efficiency is shown to depend on the signal structure and its effective rank. The second part of the work is devoted to solving a similar problem in the task of estimating a signal parameter, for example, its delay, which is linearly related to the distance to the detected observation object.

Key words: detection, parameter estimation, detection characteristics, signal-to-noise ratio, neural network.

Corresponding author: Dorosinskiy Leonid L.Dorosinsky@mail.ru

References

1. Г. В. Трис, Теория обнаружения, оценок и модуляции [G. V. Tris, Theory of Detection, Estimation, and Modulation, Vol. 1], т. 1, Москва: Советское радио, 1972, p. 744.

2. Справочник по радиолокации / Под ред. М.И. Сколника. Пер. с англ. под общей ред. В.С. Вербы. [Handbook of Radar / Ed. M.I. Skolnik. Per. from English. Under the general ed. V.S. Verba]. В 2 книгах. Книга 2. - Москва: Техносфера, 2014. - 680 с.

3. Доросинский Л.Г., Оптимальная обработка радиолокационных изображений, формируемых в РСА, [Dorosinsky L.G. Optimal processing of radar images generated in the RSA] Москва: Издательский дом Академии Естествознания, 2017, p. 212.

4. Тархов Д. А. Нейронные сети. Модели и алгоритмы [Tarkhov D. A. Neural Networks. Models and Algorithms]/Д. А. Тархов – М.: Радиотехника, 2005. – 256 с.

5. Хайкин С. Нейронные сети [Haikin S. Neural Networks]/ С. Хайкин. – М.: Издательский дом «Вильямс», 2006. – 1103 с.

6. Рутковская Д. Нейронные сети, генетические алгоритмы и нечеткие системы [Rutkovskaya D. Neural Networks, Genetic Algorithms, and Fuzzy Systems]/ Д. Рутковская, М. Пилиньский, Л. Рутковский. – М.: Горячая линия – Телеком, 2006. – 452 с.

7. Бишоп К.М. Распознавание образов и машинное обучение. [Bishop C.M. Pattern Recognition and Machine Learning] Диалектика, 2020.- 962с.

8. Мередит Бруссард. Искусственный интеллект. Пределы возможного. [Meredith Broussard. Artificial Intelligence. The Limits of Possibility]-М.: Альпина, 2020.-258с.

For citation:

Dorosinsky L.G. Comparative analysis of the efficiency of signal detection and estimation of their parameters by classical methods and neural networks. // Journal of Radio Electronics. – 2026. – №. 3. https://doi.org/10.30898/1684-1719.2026.3.14 (In Russian)