Non-Gaussian clutter modeling with generalized spherically invariant random vectors

This paper describes the modeling of non-Gaussian clutter with a set of generalized spherically invariant random vectors (SIRV´s). The generalization extends the traditional model to account for dependence between successive SIRV realizations. Significant properties of generalized SIRV´s are derived, as well as a closed-form expression for a family of generalized SIRV density functions. The density underlying recorded sonar reverberation is approximated with this function through appropriate choice of a shape parameter. Given this reverberation model, the optimum detector is derived from the generalized SIRV density likelihood ratio. This paper concludes by showing how applying this optimum detector to non-Gaussian data leads to a reduction in the false alarm rate when compared to processing with a matched filter alone