Rapid adaptation in subspace leakage environments via covariance matrix tapering

In this paper, a new approach to adaptive interference mitigation is presented which combines the principal components (PC) method with the concept of a covariance matrix taper (CMT). The PC-CMT method is shown to preserve the minimal sample support property of the PC method even when there is significant subspace leakage of the dominant interference subspace. In addition to demonstrating its utility for STAP applications, a more computationally attractive variant based on a combined diagonally loaded SMI and CMT (SMI-CMT) is introduced. The Mountain Top Radar is used as an example.