DEVICE FOR CLASSIFYING RADIO SIGNALS ACCORDING TO STRUCTURE-TIME PARAMETERS Russian patent published in 2011 - IPC G06N3/00 

FIELD: physics.

SUBSTANCE: existing device is supplemented with units for storing educational data and classification, which enable to calculate reference vectors of structure-time parameters of radio signals of known classes and corresponding values of intra-cluster dispersion using Kohonen self-organising maps. The calculated reference vectors and values of intra-cluster dispersion are used to construct a probabilistic neural network, realised by the classification unit. In contrast to the existing device, the probabilistic neural network constructed that way enables automation of the device operation and increases classification robustness.

EFFECT: automation and increase in robustness of the process of classifying radio signal according to structure-time parameters.

1 dwg