Search

Recent developments in unmanned aerial vehicle (UAV) technology are spurring new interest in the fields of bistatic and multistatic synthetic aperture radar (SAR). This dissertation develops the theory necessary to extract useful information from bistatic and multistatic SAR phase history data. The approach taken in this work involves generalization of several existing monostatic SAR signal processing algorithms.A model for bistatic SAR data collection is introduced, and based on this model, a maximum likelihood algorithm for nonparametric scene reconstruction, or image formation, is proposed. This method is commonly known as matched filtering. By deriving a bistatic far-field assumption, the matched filtering algorithm may be approximated by a bistatic formulation of the common polar format algorithm. The polar format algorithm has complexity O(N2log2N), versus the O(N4) complexity of the matched filtering algorithm.The effect of platform motion measurement errors on bistatic SAR imagery are examined. Requirements on inertial navigation accuracy are derived, and it is shown that autofocus algorithms for monostatic SAR may applied to properly formatted bistatic SAR data. Design of nonlinear receiver flight paths, with the goal of preserving and reconstructing three-dimensional scene information, is considered. Parametric models for the three-dimensional bistatic responses from canonical scattering centers are proposed and validated against high-frequency simulation results. Lastly, algorithms for three-dimensional surface reconstruction from multistatic SAR data are proposed, based on existing theory for monostatic stereo SAR and interferometric processing. A theoretical framework for predicting the performance of these algorithms is suggested.