Abstract
It is shown that the dominant eigenvectors of the space-time correlation matrix contain all the information about the space-time distribution of the interferences. The eigencanceler is a new approach to adaptive radar beamforming in which the weight vector is constrained to be in the noise subspace, the subspace orthogonal to the dominant eigenvectors. Two types of eigencancelers are suggested: the minimum power eigencanceler (MPE) and the minimum norm eigencanceler (MNE). It is shown that while the MPE is implemented as a linear combination of noise eigenvectors, the MNE can be formed using dominant eigenvectors only. Particularly for short data records, the MNE provides superior clutter and jammers cancellation, as well as lower variations in the pattern and lower distortion of the mainbeam, and can be carried out at a smaller computational cost than other known beamformers, such as the minimum variance beamformer.

This publication has 22 references indexed in Scilit: