Abstract The problem of nonconvex and nonsmooth sparse representation for direction of arrival (DOA) estimation in monostatic multiple-input multiple-output (MIMO) radar is addressed in this paper, which is dealt with by a novel iterative reweighted proximal projection method. The proposed method firstly obtains the array covariance vector by performing the vectorization operation on the reduced dimensional covariance matrix. Then a sparse representation framework is formulated for DOA estimation through minimizing the nonconvex and nonsmooth sparsity promoting function, and the weighted matrix, which is based on the high-order power of the inverse of the reduced dimensional covariance matrix, is designed for reweighting the nonconvex and nonsmooth minimization to enhance the sparsity of the solution. Thereafter, an iterative algorithm using proximal projection approach along with reweighted penalty ideas is developed to recover the sparse solution. Finally, DOA estimation is accomplished by searching the spectrum of the solution. Due to achieving a better approximation to the l0 norm, the proposed method exhibits better DOA estimation accuracy than the reweighted l1-SVD algorithm and reweighted SL0 algorithm. Furthermore, the proposed method can avoid any a-priori information on the number of targets. Simulation results are presented to verify the effectiveness of the proposed method.

中文翻译：

---

基于迭代加权近端投影的单站MIMO雷达DOA估计算法

摘要 本文解决了单站多输入多输出（MIMO）雷达中到达方向（DOA）估计的非凸和非光滑稀疏表示问题，该问题通过一种新颖的迭代重加权近端投影方法进行处理。所提出的方法首先通过对降维协方差矩阵进行向量化操作，得到数组协方差向量。然后通过最小化非凸和非光滑的稀疏促进函数为DOA估计制定稀疏表示框架，并设计基于降维协方差矩阵的逆的高阶幂的加权矩阵对非凸进行重新加权和非光滑最小化以增强解的稀疏性。此后，开发了一种使用近端投影方法以及重新加权惩罚思想的迭代算法来恢复稀疏解。最后，通过搜索解的频谱来完成 DOA 估计。由于实现了对 l0 范数的更好近似，所提出的方法比重加权 l1-SVD 算法和重加权 SL0 算法表现出更好的 DOA 估计精度。此外，所提出的方法可以避免任何关于目标数量的先验信息。仿真结果验证了所提出方法的有效性。所提出的方法比重新加权的l1-SVD算法和重新加权的SL0算法表现出更好的DOA估计精度。此外，所提出的方法可以避免任何关于目标数量的先验信息。仿真结果验证了所提出方法的有效性。所提出的方法比重新加权的l1-SVD算法和重新加权的SL0算法表现出更好的DOA估计精度。此外，所提出的方法可以避免任何关于目标数量的先验信息。仿真结果验证了所提出方法的有效性。