A two-stage deinterleaving technique for clustering of radar pulses.

Pulse deinterleaving is the process of identifying which received pulses belong to which emitter sources. Pulse deinterleaving techniques generally consist of multi-parameter algorithms that utilize pulse parameters such as angle of arrival, frequency, pulse width and pulse amplitude and interval-only algorithms. In conventional multi-parameter deinterleaving pulse width and pulse amplitude parameters are generally assumed to be unreliable parameters due to distortions caused by the multipath phenomenon in the channel. In this study, a two-stage deinterleaving technique is proposed for such deinterleaving problem. In the first stage radar pulses are deinterleaved according to their frequency and angle of arrival parameters. In the second stage antenna scan rates are firstly estimated by the Levenberg-Marquardt optimization algorithm and then this information is fed back to the first stage to decrease clustering errors. Simulation results show that this goal is accomplished.