Spherically invariant random processes for modeling non-Gaussian radar clutter

This investigation is motivated by the problem of modeling correlated non-Gaussian radar clutter. Experimental research has confirmed that radar clutter can have an extended tail under certain conditions. Since the Gaussian model fails to predict the extended tail behavior, non-Gaussian probability density functions (PDF) have been used for the first order PDF of the clutter. Usually, radars process N samples at a time. Therefore, a complete statistical characterization would involve the ability to specify the joint PDF of N correlated non-Gaussian random variables. This paper presents mathematically elegant and tractable techniques for specifying the joint PDF of N correlated non-Gaussian random variables. The approach used in this paper is based on the theory of spherically invariant random processes (SIRP). Several important properties of SIRPs are summarized