Performance analysis of ordered-statistic greatest of-constant false alarm rate with binary integration for M-sweeps

The performance of the ordered-statistic greatest of (OSGO) constant false alarm rate (CFAR) scheme with binary integration for M non-coherent sweeps in Weibull background was investigated for homogeneous and non-homogeneous backgrounds, with an assumption of known shape parameter. This kind of processing is based on the fact that the clutter can be segmented in regions in many real radar scenarios where a timevarying two-parameter distribution function family can be fitted, but the clutter power may vary locally inside the region. Under the assumption of known shape parameter, the authors examined the changes of the false alarm rate and detection probability of the OSGO-CFAR with binary integration when the shape parameter differs from the nominal one, and compared them to those of the OSGO-CFAR with single pulse processing. The authors have derived analytic expressions of the detection probability and false alarm rate during clutter power transitions for the OSGO-CFAR with binary integration in Weibull background. It is shown that the OSGO-CFAR with binary integration can not only improve the detection performance significantly, but it also control the rise of the false alarm rate at clutter edges more effectively compared to the OSGO-CFAR with single pulse processing. Moreover, it exhibits a good immunity to the variation of the shape parameter.