APPLICATION OF QUEUING THEORY TO THE ESTIMATION OF THE FEATURE SPACE FOR RADAR SOURCES IN MACHINE LEARNING

The estimation of the feature space in analysis radar signals (airplanes, ships, navigation stations, etc.), is an important element machine learning. From point view queuing theory, a mathematical model complex detected signal can be represented as ordinary flow events described by Poisson distribution for randomly varying parameters signal. paper demonstrates orthogonality characteristic sources based on entropy estimation. We show that information measure discrepancy (Kullback-Leibler divergence) between intra-pulse modulation and frequency-modulated decreases with increasing pulse duration. Thus, recognition using methods learning, it possible to limit ourselves features without algorithms processing modulation. Key words: source, Kullback-Leibler divergence,