A neural approach to tracking radar pulse trains with complex pulse repetition interval modulations

Abstract

The intervals between the pulse time of arrival (TOA) values of a radar pulse train nearly always follow a specific modulation type. Many modern emitters are now capable of producing radar pulse trains with complex pulse interval modulations. We build on earlier work that allows us to automatically and rapidly deduce this modulation type. This information is used to prime the weights of a linear adaptive filter employing a single linear neuron with possibly multiple inputs. As a result, we are rapidly able to track the pulse train and to predict the next pulse event, even in the presence of missing and spurious pulses, as well as complex pulse interval modulations.