Image while scan (IWS): An adaptive Gabor spectrum estimation approach

We extend the classical ISAR imaging methodology by introducing a rotating transmit-receive antenna structure that allows us to simultaneously perform target imaging and surface surveillance operations. The inherent problem of processing only a limited number of pulses (per range-bin) is overcome by characterizing the received Doppler signals via a Kalmanized sparse Gabor representation that allows us to accurately capture the spectral information contained in the radar signals at different levels of resolution while also affording us the ability to adapt the representation based on the radar signals received in successive frames. We demonstrate the performance of our spectrum estimation methodology by testing it on real radar data, and comparing it with linear prediction (Burg´s method) and sinusoidal Bayesian spectrum estimation algorithms. We find that our approach offers important advantages over previous approaches to Doppler spectrum estimation, while also potentially enabling the incorporation of wavelet thresholding strategies in order to suppress noise present in radar signals due to environmental and systemic factors.