On blind channel identification for impulsive signal environments

New methods for parameter estimation and blind system identification for impulsive signal environments are presented. The data are modeled as stable processes. First, methods for estimating the parameters (characteristic exponent and dispersion) of a symmetric stable distribution are presented. The fractional lower-order moments, both positive and negative order, and their applications are introduced. Then a new algorithm for blind channel identification based on fractional lower-order moments is proposed. The alpha-spectrum, a spectral representation for impulsive environments, is developed. Conditions for blind identifiability of nonminimum phase FIR channels are established using the properties of the Alpha-Spectrum.