A Novel Clutter Covariance Matrix Estimation Method Based on Feature Subspace for Space-Based Early Warning Radar
Tianfu Zhang,Zhihao Wang,Ning Qiao,Shuangxi Zhang,Mengdao Xing,Yongliang Wang
DOI: https://doi.org/10.1109/jstars.2021.3123648
IF: 4.715
2021-01-01
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Abstract:Accurate estimation of the clutter covariance matrix for the cell under test (CUT) is a committed step in the spatial-temporal adaptive processing (STAP) algorithm. The unique nonstationary characteristic of signal for space-based early warning radar (SBEWR) leads to the spatial variation of training sample and the insufficient number of optional independent identically distributed (i.i.d.) training samples, which brings difficulties to training sample selection and covariance matrix estimation. To improve the estimation accuracy of clutter covariance matrix and the performance of STAP for SBEWR in a heterogeneous environment, a novel training sample selection and clutter covariance matrix estimation method is proposed. The method based on clutter subspace reconstruction and spectrum correction technology can improve the estimation accuracy of clutter covariance matrix in the case of nonstationary signals and heterogeneous environments. The clutter covariance matrix estimated by the proposed method is similar to the clutter covariance matrix of the CUT, and the performance of STAP is improved. The experimental results confirm the performance of the proposed method.
imaging science & photographic technology,remote sensing,engineering, electrical & electronic,geography, physical