| 摘要: |
| 在航天领域的许多测量场景中，信号的相位信息容易遭到破坏。本文旨在解决有异常值干扰的相位检索问题。中值可以对抗异常值，故常作为正则化标准截取无相测量值，这会导致可用测量值的减少,从而提高测量复杂度。为克服这一难点，进行加权处理，提出了加权中值截取算法，利用可以表征指向原信号概率的权值进一步挖掘了未被截取测量值的可用信息，与同类算法相比,在测量复杂度和对抗异常值干扰方面更为优越。 |
| 关键词: 相位检索 异常值 中值截取 |
| DOI： |
|
| 基金项目:国家自然科学基金(62071149) ；国家自然科学基金(61671177)；鸿鹊创新中心开放基金 |
|
| A Weighted Median Truncated Algorithm for Phase Retrieval with Outliers |
| LI Xiaodong,LIU Guannan,FU Ning,QIAO Liyan |
| (School of Electronics and Information Engineering, Harbin Institute of Technology, Harbin 150001, China;Beijing Institute of Astronautical Systems Engineering, Beijing 100076, China) |
| Abstract: |
| In numerous measurement scenarios within the aerospace domain, signal phase information often faces susceptibility to degradation. This article endeavors to address the challenge of phase retrieval amidst the presence of outliers. Leveraging the robustness of the median against outliers, it serves as a standard regularization technique to intercept phaseless measurements. However, this interception can potentially reduce available measurements and escalate measurement complexity. To surmount this obstacle, a novel weighted median truncation algorithm is proposed. This algorithm integrates weights that aptly capture the likelihood of referring to the original signal, thereby delving deeper into the available information of unmeasured values. Comparative analysis reveals that this approach offers superior measurement complexity and resilience against outlier interference when juxtaposed with similar algorithms. |
| Key words: Phase retrieval Outliers Median truncated |