基于稀疏贝叶斯学习的无源雷达高分辨成像

针对无源雷达压缩感知成像，该文提出一种基于稀疏贝叶斯学习的高分辨成像算法。基于一次快拍模式下的无源雷达回波模型，文中首先考虑目标散射系数的统计特性及其对微波频率的依赖关系，将无源雷达成像转化为MMV(Multiple Measurement Vector)联合稀疏优化问题；然后对目标建立了级联形式的稀疏先验模型，并利用稀疏贝叶斯学习技术进行求解。相比之前基于目标确定性假设的稀疏恢复方法，所提算法更好地利用了目标的统计先验信息，具有能够自适应调整参数(目标模型参数和未知噪声功率)和高分辨反演目标等优点。仿真结果验证了该算法的有效性。

High-resolution Imaging of Passive Radar Based on Sparse Bayesian Learning

This paper presents a high-resolution imaging method based on Sparse Bayesian Learning (SBL) for passive radar compressed sensing imaging. Under the one-snapshot echo model, the proposed method firstly takes account of the frequency-dependent statistics of the target scattering centers, and changes passive radar imaging into a joint Multiple Measurement Vector (MMV) sparse optimization problem. Further, a hierarchical Bayesian framework for sparsity-inducing priori of the target is established, then the MMV problem is efficiently solved by utilizing the SBL theory. Unlike the previous sparse recovery algorithms relying on the deterministic assumption of the target, the proposed method makes a better use of the target prior information, and has the advantages of adaptively estimating parameters (including the parameters in the priori model of the target, and the unknown noise power) as well as the high-resolution imaging, etc.. Simulation results show the effectiveness of the proposed method.