基于稀疏贝叶斯学习的稳健STAP算法

Sparse Bayesian learning-based robust STAP algorithm

为了提高阵列幅相误差和格点失配情况下稀疏恢复空时自适应处理(sparse recovery space-time adaptive processing,SR-STAP)算法的性能,提出一种基于稀疏贝叶斯学习的稳健SR-STAP算法。首先,利用空时导向矢量的Kronecker结构构建SR-STAP误差信号模型;然后,利用贝叶斯推断和最大期望算法迭代求取角度多普勒像和误差参数;最后,利用求解参数估计精确的杂波加噪声协方差矩阵并计算权矢量。仿真实验表明,所提算法能够显著提高稀疏信号模型失配时的目标检测性能。

To improve the performance of sparse recovery space-time adaptive processing(SR-STAP)algorithms with both array gain/phase errors and grid mismatches,a sparse Bayesian learning-based robust SR-STAP approach is proposed in this paper.Firstly,the SR-STAP signal model with mismatched errors is constructed using the Kronecker structure of the space-time steering vector.Secondly,the angle-Doppler profile and mismatched parameters are alternatively achieved by utilizing the Bayesian inference and expectation-maximization algorithm.Finally,the precise clutter-plus-noise covariance matrix is estimated and the corresponding weight vector is calculated with the above obtained parameters.Simulation results verify that the proposed algorithm can significantly improve the target detection performance with mismatched sparse signal model.