Robust STAP Algorithms for Target Detection and Tracking Using Alpha-Stable Statistical Models of Non-Gaussian Clutter

This project will test and evaluate the performance of newly designed Space-Time Adaptive Processing (STAP) algorithms for improved detection and joint azimuth/Doppler estimation using alpha-stable statistical models. These algorithms are able to support airborne surveillance radar systems in a variety of non-Gaussian clutter environments. MultiSpec's advanced signal processing algorithms based on alpha-stable models hold great promise in this regard, because alpha-stable amplitude statistics were shown to be suitable for characterization of the natural clutter data collected in the DARPA Mountaintop program. Performance improvements over existing methods will be examined and qualified. Three tasks are proposed for the Phase I project: (1) statistical characterization of non-Guassian clutter data with emphasis on multivariate alpha-stable models; (2) design and performance assessment of adaptive CFAR detection algorithms based on alphastable models and associated; and (3) design and performance assessment of STAP algorithms for joint estimation of azimuth/Doppler based on covariations. In all three tasks emphasis will be placed on algorithmic developments that are cost effective in order to assure quick transition to field applications. The efficacy of MultiSpec's will be demonstrated using generic clutter simulations and samples of the Mountaintop clutter data.