Beamforming With Dominant Mode Rejection
Douglas A. Abraham, N.L. Owsley · 2005
High resolution adaptive beamforming of a spatial array of sensors has traditionally been formulated in terms of a minimum noise variance, distor tionless response (MVDR) optimality criterion . This approach results in a constrained spatial Wiener filter structure which requires the estimation and inversion of either the array space-time correlation matrix for time domain implementation or the array cross spectral density matrix (CSDM) for a frequency domain implementation 1. In the situation where the array acoustic environment consists of a limited number of dominant modes the full matrix inversion procedure is both numerically and statistically inadvisable. In this case, the technique of point source interference null steering can be generalized to the nulling of dominant modes without prior knowledge of the mode structure . In effect, a reduced rank version of the spatial Wiener filter is produced which is realized as a generalized mode nulling beamformer .