Advanced SAR Processing Techniques Based on Alpha-Stable Statistical Models

This project will address the validation and testing of advanced processing techniques that are needed to support synthetic aperture radar (SAR) systems, particularly in the low-frequency UWB/UHF region for Navy littoral surveillance missions and for terrain characterization and mapping, especially in rugged forested regions. Underscoring this effort are MultiSpec's advanced signal processing algorithms based on alpha-stable statistical models and their corresponding fractional lower- order statistics (FLOS). Alpha-stable models have shown a great promise in terms of clutter characterization, robustness in target detection and parameter estimation, and performance improvement over existing second-order correlation-based matched filters and related signal analysis methods. Consequently, the exploratory develop- ment, validation and testing of FLOS-based algorithms to the SAR processing problem has become timely. Five directions are proposed during this exploratory study: (1) statistical characterization of targets and clutter in terrain environments with emphasis on alpha-stable models; (2) assessment and demonstration of joint range/Doppler estimation algorithms based on fractional lower-order statistics; (3) assessment and demonstration of FLOS-based matched filter processing system; (4) demonstration of efficacy using Navy-supplied data; and (5) as an optional task, development and testing of FLOS-based motion compensation algorithms.