A Robust STAP Algorithm Using Two Prior Knowledge

‍ ‍The Space-Time Adaptive Signal Processing technique adaptively suppresses the clutter in the joint space-time domain to realize the detection of moving target detection. Sparse Recovery based STAP （SR-STAP） exploits the sparsity of the clutter spectrum， which can effectively alleviate the performance degradation when the number of training data is insufficient in heterogeneous environment. Although SR-STAP only requires a small number of samples to recover the clutter spectrum and reconstruct the Clutter Covariance Matrix （CCM）， the clutter spectrum recovery accuracy has the potential to be further improved when the number of samples is increased. On the other hand， when the radar receiver consists of the uniformly spaced linear array and the repetition frequency of transmitted pulses is constant in a coherent processing interval， the CCM has the structure of persymmetric. If this prior knowledge is fully utilized， the number of equivalent training samples will be doubled. In this paper， by combining the persymmetric property of CCM into the framework of SR-STAP， we proposed a novel robust SR-STAP algorithm. The proposed algorithm could further improve the estimation accuracy of CCM and the reconstruction accuracy of clutter spectrum under the same number of training data by exploiting the two prior knowledge including the persymmetric property of CCM and the sparsity of clutter spectrum. Firstly， we preprocessed the data in the cell under test and training data by introducing the persymmetry transformation matrix to increase the number of equivalent training data as twice as the original. Then， the CCM was reconstructed and the corresponding STAP filter is devised by combining the preprocessed training data and a sparse iterative covariance-based estimation algorithm. The proposed algorithm did not require any hyperparameter， which made it is easy for operating in practical applications. Finally， numerical examples showed that the proposed algorithm could significantly improve the clutter spectrum recovery accuracy and outperforms its competitor in terms of clutter suppression performance.