Bilinear compressed sensing for array self-calibration

Benjamin Friedlander, Thomas Strohmer

Abstract: We consider array self-calibration algorithms for direction finding in the presence of unknown complex sensor gains. By exploiting the sparsity in the azimuth domain, we can utilize methods developed in the context of compressive sensing. A self-calibration algorithm is proposed that alternates between solving a joint-sparse optimization problem and estimating the unknown array gains. Furthermore, for the case of fully correlated signals we introduce a convex programming algorithm based on bilinear compressive sensing.