Subspace based DOA estimation in the presence of correlated signals and model errors

High-resolution subspace based direction-of-arrival (DOA) estimation requires a number of assumptions about the signal and antenna. In our application automotive radar, among many practical problems, signals are often correlated and the antenna is not calibrated. This difficult combination has been seldom addressed in the literature. In this paper, we study their simultaneous impact on DOA estimation, describe a so called ldquocoherent model error interferencerdquo phenomenon, propose a prewhitening scheme for algebraic subspace based DOA estimation, and show some simulation results.