Recognition of radar stations by radio technical signal parameters using machine-learning methods

In the course of solving the problem of electronic surveillance, the issue of classification of detected radio signals arises. Classical classification methods based on comparison of the obtained data on detected stations with reference a priori known parameters of stations do not allow obtaining a result with a high degree of reliability, therefore, to solve this issue, we turned to machine learning methods, which are currently used everywhere. In this paper, such machine learning methods as random forest, multilayer perceptron, gradient boosting, convolutional NN, local emission level method were considered. In addition, sequential binary classification method for determining emissions was developed in the test sample.