Probabilistic adaptive CFAR

False alarm regulation is a critical requirement in modern surveillance radars. Present techniques are susceptible to poor false alarm control when operated in non-Gaussian backgrounds representative of real world situations. The conventional log-t CFAR is shown to be highly sensitive to the inclusion of small amounts of receiver noise in addition to the Weibull clutter for which it is tuned. An adaptive log-t CFAR is described which, under certain conditions, provides an order of magnitude better false alarm control than the conventional log-t CFAR. This adaptive processor maintains false alarm regulation by estimating the probability distribution of the interference and hence, in concept, is independent of the interference probability distribution.