Development of non-coherent CFAR detection processes in Weibull background

Abstract

The Weibull distribution is a model for land and sea based radar backscattering, and as such there is continued interest in the development of non-coherent detection processes under such a clutter model assumption. The focus here is on the derivation of such detectors which achieve the constant false alarm rate property in Weibull background. This property is acquired without the requirement of clutter parameter approximation through secondary data. Towards this objective, recent investigations have shown that it is possible to design such detectors for operation in clutter environments which can be modelled by scale and power invariant distributions, of which the Weibull distribution is an example. Based upon a general form of detector a series of specific examples are introduced, and their performance is investigated, in both synthetic and real clutter returns. It will be demonstrated that a non-coherent detection process exists, which achieves the full constant false alarm rate property in homogeneous Weibull clutter, whose performance matches detectors based upon a priori knowledge of the Weibull clutter parameters.