Radar Signal Recognition Based on Deep Neural Networks

The article presents the problem of recognizing radar sources using multilayer artificial neural networks on the basis of radar parameters such as pulse repetition interval, pulse duration, radio frequency, and inter-pulse temporal structure of signals generated by these sources. Parameters of real radar signals were used as a set of training data. The correctness of the operation was tested using a simulation environment that verified the effectiveness of operation in the conditions of the presence of many radar signals that interfere with each other and in the environment of noise and interferences.