A random matrix theory model for the dominant mode rejection beamformer notch depth

The Dominant Mode Rejection (DMR) adaptive beamformer (ABF) computes its array weights by substituting a modified sample covariance matrix into the expression for the Capon beamformer array weights. Incorporating recent random matrix theory results on sample eigenvector fidelity for spiked covariance models into the DMR beampattern expression results in a model for the notch depth in the direction of a single loud interferer. The model predicts the mean DMR notch depth as a function of the number of snapshots, the interferer-to-noise ratio (INR), the array size, and the interferer location relative to the look direction. The model predictions agree closely with simulations over a wide range of INRs and snapshots.