A Robust and Efficient
|S.U. Pillai, Y.L. Kim, and J.R. Guerci
Dept. of Electrical Engineering
Six Metrotech Center
Brooklyn, New York 11201
tel: (718) 260-3732
fax: (718) 260-3906
Abstract A new technique for removing dominant interference eigenvectors is presented which does not require that an eigendecomposition be performed-as is the case with the eigencanceler method. Similar to eigen-based cancellation techniques, the resulting space-time pattern is formed by subtracting from the quiescent steering vector, a weighted sum of interference eigenvectors. However, the subtraction is accomplished implicitly, without having to directly estimate the interference eigenvectors. Instead, a subspace projection is formed in the data domain which typically has a dimension an order of magnitude smaller than the total available number of degrees-of-freedom for the space-time array. This new method is ideally suited for nonstationary clutter environments, where only a small sample support is available for the adaptive training set, and where a premium is placed on computational overhead. A detailed analytical comparison between the proposed technique and existing eigen and non-eigen based constrained interference removal techniques is presented. Also, experimental results are presented based on both synthetic and recorded data sets originating from the Mountaintop radar.
Direct comments and questions to: email@example.com
© MIT Lincoln Laboratory. All rights reserved.