chapter  7
Dual Order Statistic CFAR Detectors
WithGraham V. Weinberg
Pages 26

Dual order statistic detection processes have appeared within the signal processing literature in several different clutter model contexts. The idea of such a detection process has been to extend H. Rohling's orginal OS-constant false alarm rate (CFAR) process to allow two order statistics of the clutter range profile (CRP) to provide a measurement of the clutter level. The application of pairs of order statistics as a measurement of the level of clutter was first examined in, in the context of Weibull distributed clutter. When specialised to the Pareto Type I case it was proven that its probability of detection could be expressed elegantly in terms of beta-type functions, which then could be used to extract its probability of false alarm in terms of its threshold multiplier. The latter was demonstrated to increase with the first index in the dual order statistic detector.