Power Spectrum Estimation via Modified Indirect Dominant Mode Rejection

It is well known that under low sample support of the correlation matrix, low rank solutions of the Minimum Variance Distortionless Response (MVDR) problem can yield better performance than the full rank solution. A new adaptive approach, Indirect Dominant Mode Rejection (IDMR), was introduced in previous work to solve the MVDR optimization problem. In this paper we extend IDMR beamformers from the signal arrival directions to the entire spatial spectrum. We also introduce several modifications, which bring more robustness to basic IDMR. These include root-music, forward/backward averaging of the sample correlation matrix and power window. The MVDR beamformer is scaled to satisfy the unity gain constraint, while the power window technique determines the scaling factor based on the fact that it is the power arriving from the look-direction. MUSIC is used to provide the power window technique an initial estimate of these powers.