Computing K-clutter-plus-Gauss distributions

K-distributions have been used to model certain non-Gaussian noise processes including the intensity of radar returns from the sea surface. Assuming K-distributed clutter, the noise at the output of the receiver may then be the sum of receiver noise, which is assumed to have a Gaussian distribution, and K-distributed noise. The distribution of the amplitude of this K-plus-Gauss (KPG) noise has been represented as an integral, and moments of the distribution have been obtained. To apply a distribution to data, it should be computable. However, closed form solutions of the KPG integral have not been obtained. For certain values of the parameters of the KPG distribution the integrand is singular, and in these cases numerical integration may be difficult. In this paper the integral is expressed as an infinite series of Tricomi functions. The KPG densities are computed by using this series and various approximations.