Navid Daryasafar - Faculty of Electrical Engineering Dashtestan Branch, Islamic Azad University Borazjan, iran
Hamid Dehghani - Faculty of Electrical Engineering Malek Ashtar University Tehran, Iran
Azim Rezaei Motlagh - Faculty of Electrical Engineering Dashtestan Branch, Islamic Azad University Borazjan, Iran

Deinterleaving of radar pulses is performed based on features of pulse trains, including pulse arrival time, pulse width, radio frequency, and among others. Radar pulse deinterleaving is one of the most important parts of radar identification system in which identification/detection process of radar transmitters is done in simultaneous fashion. In this paper, a novel algorithm is proposed by the combination of Imperialist Competitive Algorithm, ICA, and k-means algorithm based on Rough set theory. The identification accuracy on radar data by this method was 78%, emphasizing the improved performance over the conventional methods. It will be shown that this new method can be successfully utilized in busy radar environments with various signals.

Pulse Detection, Pulse Deinterleaving, Radar identification, Threshold Level