Clutter covariance matrix estimation using weight vectors in knowledge-aided STAP

Abstract

A knowledge-aided space-time adaptive processing (STAP) is a quite useful technique to suppress non-stationary and heterogeneous clutter. It estimates a covariance matrix by combining a conventional covariance matrix based on secondary data with a synthesised one by prior information. A new combining method is presented, where weight vectors, rather than constant weights, are used to combine two covariance matrices. In this method, the weight vectors are derived in a way to maximise clutter-to-noise ratio of the combined covariance matrix. A numerical simulation is conducted for a bistatic radar scenario where clutter non-stationarity and heterogeneity can be assumed and the performance of the proposed method is demonstrated in terms of clutter suppression and target detection.