Efficient tapering methods for STAP

Abstract

The sample support problem in space-time adaptive processing (STAP) arises from the requirement to adapt to a changing interference environment where the available wide-sense-stationary sample support is severely limited for direct implementation of adaptive algorithms. In this paper we outline several approaches to address the sample support problem by utilizing efficient covariance matrix tapering (CMT) methods to retain the a-priori known structure of the covariance matrix. By combining efficient tapering approaches along with terrain knowledge based STAP and other preprocessing schemes such as subarray - subpulse, relaxed projection method, it is possible to reduce the data samples required in a nonstationary environment and consequently achieve superior target detection. In addition, the application of Khatri-Rao product to the data domain implementation of CMT is also introduced thus expanding the class of robust algorithms for real-time STAP implementation.