Robust Krylov-Subspace Methods for Passive Sonar Adaptive Beamforming

Krylov-subspace methods, such as the multistage Wiener filter and conjugate gradient method, are often used for reduced-dimension adaptive beamforming. These techniques do not, however, allow for steering vector mismatch, which is typically present in many applications of interest, including passive sonar. Here, we discuss recently proposed robust methods that do allow for steering vector mismatch by combining Krylov-subspace methods with robust Capon beamforming techniques. We present experimental data examples that illustrate the benefits of using these algorithms in passive sonar.