Nonlinear filtration of clutter

The problem of nonlinear filtration according to minimum mean-square error as a criterion is solved for a signal, appearing in an additive mixture with clutter and white noise. For synthesis of the nonlinear optimum filter, its characterizing continuous operator is expanded into a product of sums and a product. Minimization of the corresponding functional reduces the problem to a system of equations which yields the filter parameters and mean-square error. For illustration, the procedure is applied to filtration of a narrow-band signal appearing with a narrow-band non-Gaussian clutter and white noise. The energy advantage of the nonlinear optimum filter over the linear one increases power of the non-Gaussian clutter, but vanishes in the absence of white noise.