The authors present a means of exploiting information from other time and Doppler bins in the vicinity of the test cell to ascertain the background level in the test cell, in radars employing Doppler filter banks and multiple bursts per dwell. Current radars often comprise single-Doppler CFAR processors, with OR selection logic between Dopplers on each burst, and M/N binary integration over the N bursts. A variation of this architecture is presented. They take as a point of departure that in order to reduce susceptibility of the processor to spiky clutter, the test statistic must be derived primarily from samples in the same range cell as the test sample. Several Doppler bins are assumed to be available in each range cell, and on each burst. The type of filter implemented is not of importance provided several essentially nonoverlapping Doppler bins are created.
展开▼