Track-before-detect processing for an airborne type radar

A real-time feasible track-before-detect process for a scanning pulsed Doppler airborne-type radar is described. Robust distribution-free based ranking is applied in range separately for each Doppler. A dynamic programming process is used to select a best tentative track passing through each range, azimuth, and Doppler cell. The process sequentially optimizes smoothed rank as the field of view is scanned. Results of a simulation are presented for a system with multiple bursts at a given pulse repetition frequency. Track detections are obtained with probability 0.5 at a signal-to-noise ratio of about 5 dB for a Rician signal pulse noise model as measured at the Doppler outputs. The corresponding false track generation rate is estimated to be 5*10/sup -5/ per processed resolution cell.<>