Matched direction beamforming based on signal subspace

The actual manifold vector (or steering vector) of the signal of interest (SOI) is often imprecise in practical applications. However the true manifold vector can be expressed as a product of a known matrix and an unknown coordinate vector in many cases. This model can accommodate many manifold uncertainties, for instance, the look direction error, local scattering, etc. Matched direction beamformer (MDB) is referred to as the beamformer resolving the signal that is drawn from an unknown direction inside the known subspace. The main contribution of this paper is to propose a new MDB that can estimates the coordinate vector associated with the SOI, without using the knowledge of the interference subspace (IS). Moreover the proposed approach is robust to the dimension overestimation of the signal subspace.