Distribution Independent CFAR Detector Using Extreme Value Theory

Abstract

According to the extreme value theory, the tail distribution of virtually any statistical distribution can be uniquely modeled by the generalized Pareto distribution. Based on this property, a CFAR detector making no assumption about the distribution of interference is proposed. The parameters of the generalized Pareto distribution can be estimated using the method of L-moments. The performance of the new detector is compared to CA- and Optimized Weibull CFAR detector in the Rayleigh, Weibull and lognormally distributed clutter amplitude.