结构扰动与粗差鲁棒的l_p正交匹配追踪算法

发布时间：2018-03-06 11:41

本文选题：压缩感知　切入点：结构扰动　出处：《系统工程与电子技术》2017年01期 　论文类型：期刊论文

【摘要】：针对压缩感知稀疏信号恢复,提出了一种对测量结构扰动和粗差同时鲁棒的l1正交匹配追踪(structure perturbation and outlier robust l1-orthogonal matching pursuit,l1-SPOR-OMP)算法。该算法利用l1范数对粗差的鲁棒性,分别约束信号和扰动的拟合误差,进而通过交替迭代使它们同时到达最小,从而同时实现对信号和扰动的有效鲁棒恢复;由于算法仅需求解支撑位置的有效扰动,因此极大地降低了算法的计算复杂度。此外,为了进一步提高算法的恢复效率,也提出了使用鲁棒OMP方法来估计有效扰动的l1,2-SPOR-OMP算法。在证明了所提出算法收敛性的同时,分析了运算复杂度。数值仿真的结果验证了所提算法的有效性和优越性。
[Abstract]:For the recovery of compressed perceptual sparse signals, a l1-orthogonal matching structure perturbation and outlier robust l1-orthogonal matching demanding l1-SPOR-OMP-based algorithm is proposed, which is robust to both structural disturbances and gross errors. The algorithm utilizes the robustness of l 1-norm to gross errors. The fitting error of the signal and the disturbance is constrained respectively, and then they reach the minimum simultaneously by alternating iteration, so that the effective robust recovery of the signal and the disturbance can be realized simultaneously, because the algorithm only needs to solve the effective disturbance of the support position. Therefore, the computational complexity of the algorithm is greatly reduced. In addition, in order to further improve the recovery efficiency of the algorithm, a robust OMP method is proposed to estimate the effective perturbation. The convergence of the proposed algorithm is proved. The computational complexity is analyzed and the effectiveness and superiority of the proposed algorithm are verified by numerical simulation.
【作者单位】：复旦大学信息科学与工程学院电子工程系智慧网络与系统研究中心;
【分类号】：TN911.72

【相似文献】