A robust STAP method for airborne radar based on clutter covariance matrix reconstruction and steering vector estimation

In this paper, we consider the problem of space-time adaptive processing (STAP) when inaccurate target (namely the angle and Doppler frequency of target are uncertain) exists in the training samples, which would result in target self-nulling effects and degrade the performance of STAP significantly. A robust STAP method is proposed by reconstructing the clutter-plus-noise covariance matrix (CPNCM) and estimating the target spatial-temporal steering vector. The CPNCM is reconstructed by integrating clutter Capon spectrum over a region which includes the clutter component and is separated from the region of interest (ROI), and we use a discrete sum method to compute the integral approximately for approximate computation. Similarly, the ROI covariance matrix is calculated by integrating signal Capon spectrum over the ROI which includes the signal component. After that, the correlation coefficient between the presumed steering vector and the dominant eigenvectors of ROI covariance matrix is computed. The spatial-temporal steering vector of target is then estimated by searching for the eigenvector corresponding to the maximum correlation coefficient. Finally, a novel robust STAP weight is obtained based on the reconstructed CPNCM and estimated spatial-temporal steering vector of target. Compared to conventional robust STAP methods, simulation examples demonstrate that the proposed method can obtain a more robust performance against target spatial-temporal steering vector mismatch, better output signal-to-clutter-plus-noise (SCNR) and target detection performance in practical clutter environment.