Unification of Descriptive Experiment Design and Worst-Case Performance Optimization-Adapted Regularization Paradigms for High-Resolution Reconstruction of Radar Imagery

Abstract

We address a new approach to solving radar imaging problems stated and treated as uncertain ill-conditioned inverse problems of nonparametric spatial power spectrum estimation via processing the finite number of independent observations of the degraded array data signals (one realization of the trajectory signal in the case of SAR). The idea is to adapt a statistically optimal minimum risk nonparametric power spectrum estimation approach to the radar imaging scenarios with model-level and system-level uncertainties. The proposed incorporation of the worst-case performance optimization-adapted robust regularization aggregated with the descriptive experiment design paradigm into the minimum risk nonparametric estimation strategy leads to a new unified doubly regularized minimum risk approach for robust adaptive high-resolution reconstructive imaging in the uncertain remote sensing scenarios.