ВИКОРИСТАННЯ СТІЙКИХ АЛГОРИТМІВ В ЗАДАЧІ ВИЯВЛЕННЯ РУХОМИХ ЦІЛЕЙ НА ТЛІ НЕГАУСІВСЬКИХ ЗАВАД.

Radar detection of signals containing useful information about surveillance objects is a complex and multifunctional process that combines the solution of various problems, one of which is the detection of moving targets against the background of passive interference. Its solution is based on a fundamental idea based on the application of the Doppler effect, according to which, methods of selecting moving targets were developed, which consist in changing the frequency of the signal reflected from the moving object. Traditional sweep-to-sweep subtraction algorithm of passive interference of MTI systems, are effective when the interference is stationary, but have disadvantages caused by interference with a complex frequency spectrum. As a result, to improve the efficiency of radar detection systems, adaptive algorithms were proposed and investigated, and their advantages were shown. However, the need for further processing and improvement of existing systems and methods remains relevant and is caused by many aspects, one of which is the fact that radar detection systems are forced to function in conditions of interference of various natures, which cannot always be described by a Gaussian model. Thus, the paper focuses on the problem of detecting radar signals reflected from moving targets against the background of obstacles described by a non-Gaussian distribution, and also investigates ensuring the stability of the detection algorithm. The random occurrence of disturbances in the observation process allows for their mathematical representation to use an autoregression model generated by a disturbance described by the Laplace model. As a result, a locally optimal decision rule is synthesized for detecting a signal of a known form against the background of an autoregressive disturbance, which consists in finding the maximum following the signal parameter and the vector of the disturbance parameters of the obtained likelihood ratio, which is the ratio of the hypotheses put forward concerning the obtained sample. The robustness of the detection algorithm is ensured by estimating the unknown parameters of the noise process using an empirical Bayesian approach. In the course of work, the performance of the synthesized algorithm under the influence of non-Gaussian interference, which is described by the K-distribution model, was also investigated. To understand the effectiveness of the proposed detection algorithm, statistical modeling is performed. The simulation results confirm the effectiveness of the synthesized robust algorithm over the non-robust one, the effectiveness of which is manifested in the presence of impulse interference with an increase in the probability of their occurrence, accordingly, it is more resistant to the occurrence of chaotic impulse interference, while the non-robust algorithm gives significant errors, which leads to the deterioration of the detection of the useful signal. [ABSTRACT FROM AUTHOR]