Sparse Source Location for Real Aperture Radar Using Generalized SPICE Approach

Source location for real aperture radar (RAR) has raised many concerns in the fields of ground-based monitoring for aircrafts and vessels. However, the resolution of RAR in azimuth is constrained by the antenna beam width, which results in low degree of location accuracy. In this paper, a superresolution methodology to locate sources is proposed by exploiting the inherent sparseness of the target distribution. Based on the weighted covariance fitting criterion, we first introduce the popular SParse Iterative Covariance-based Estimation (SPICE) approach to improve the angular resolution. Furthermore, to increase the sparseness and location accuracy, a varying penalty on noise term is adopted to extend the SPICE method generally. Numerical results validates that the generalized SPICE method enjoys improved resolution and higher locationthe accuracy when compared with the RAR system and existing superresolution algorithms.