A Bayesian-Based CFAR Detector for Pareto Type II Clutter

The Pareto class of distributions has been shown to be suitable intensity models for X-band maritime surveillance radar clutter. Hence there has been much interest in the construction of non-coherent detection processes with the constant false alarm rate (CFAR) property. As a result of the validity of the Pareto Type I model, as an approximation for the Pareto Type II, it has been possible to derive a large number of relevant detection processes. However, recent work has investigated whether it is possible to achieve CFAR with respect to the Pareto Type II class of models directly. It will be shown that the solution to this can be achieved via a Bayesian predictive inference approach. A detector, which is completely CFAR in Pareto Type II clutter, will be derived and its performance in radar clutter examined.