Chrysostomos L. Nikias
Signal and Image Processing Inst., Dept. of Elec. Eng.-Systems, Univ. of Southern California, 3740 McClintock Ave., Rm. 400B, Los Angeles, CA 90089-2564
The importance of extending the statistical signal processing methodology to the so-called alpha-stable farmework is apparent. First, scientists and engineers have started to appreciate alpha-spectra and the elegant scaling and self-similarity properties of stable distributions. Additionally, real life sonar applications exist in which impulsive ocean channels tend to produce large-amplitude, short-duration interferences more frequently than Gaussian channels do. The stable law has been shown to successfully model noise over certain impulsive channels. In this lecture, new robust techniques are proposed for source detection and localization in the presence of noise modeled as a complex isotropic stable process. First, optimal, in the maximum likelihood sense, approaches are presented and the Cauchy beamformer is introduced. Also, subspace methods based on fractional lower-order statistics are developed for sonar applications where reduced computational cost is a crucial design parameter. Finally, simulation experiments demonstrating the performance of the proposed methods are presented.