Semi-blind turbo decoding in impulsive noise channels

In this paper, we deal with the application of turbo codes over impulsive noise channels. We investigate the robustness of the turbo decoding performances in presence of an impulsive noise considered by the Generalized Gaussian (GG) probability density function (PDF) model. This noise model has been proposed for many applications in communication systems. In addition, we propose a semi-blind decoding procedure suitable for impulsive noise environments. This robust scheme is obtained by a simple procedure based on higher order statistics (HOS) which allows to set the best noise distribution estimation into the soft metrics needed for the decoding process. Results show that the proposed approach offers considerable performance improvement.