基于自适应参数估计的微动时频表征重构方法  

A Reconstruction Method of Micro-Motion TFR Based on Adaptive Parameter Estimation

作　　者：李开明[1,2] 王欢 解岩[4] 陈卓 高泽岳[5] LI Kaiming;WANG Huan;XIE Yan;CHEN Zhuo;GAO Zeyue(Information and Navigation School,Air Force Engineering University,Xi’an 710077,China;Collaborative Innovation Center of Information Sensing and Understanding,Xi’an 710077,China;The Second System Design Department,Xi’an Electronic Engineering Research Institute,Xi’an 710100,China;Sichuan Provincial Government Affairs Service and Public Resources Exchange Service Center,Chengdu 610000,China;Unit 95894,Beijing 112211,China)

机构地区：[1]空军工程大学信息与导航学院,西安710077 [2]信息感知技术协同创新中心,西安710077 [3]西安电子工程研究所总体二部,西安710100 [4]四川省政府服务和公共资源交易服务中心,成都610000 [5]95894部队,北京112211

出　　处：《空军工程大学学报》2024年第5期107-114,共8页Journal of Air Force Engineering University

基　　金：国家自然科学基金面上项目(62371468,62301599,62271500,62131020)。

摘　　要：针对数据缺失条件下的目标微动回波时频表征重构问题,提出了一种基于自适应参数估计的微动时频表征重构方法。首先,将缺失的微动时频表征重构问题建模为基于L_P范数最小化的稀疏重构问题,其次,引入哈达玛积参数将L_P范数最小化稀疏重构问题转化为多个L_2范数联合最小化问题,并采用迭代吉洪诺夫正则化求解,同时在每次迭代过程中根据重构结果自适应估计正则化参数,最后,采用除偏处理减小了重构时频表征的振幅衰减。与传统微动回波时频表征重构方法相比,所提方法避免了需要人工设置正则化参数不足的问题,并且重构的时频表征更加完整。仿真实验和实测数据处理结果验证了所提方法的有效性和稳健性。In view of the time-frequency representation (TFR) reconstruction of the micro-motion signals under conditions of incomplete data, a reconstruction method of micro-motion is proposed based on the adaptive parameter estimation. Firstly, the missing micro-motion TFR reconstruction problem is modeled on the L P norm minimization sparse reconstruction problem, and by introducing Hadamard product parameter (HPP), the L P norm minimization sparse reconstruction problem is transferred to a joint minimization problem with multiple L 2 norm, and solved by using iterative Tikhonov regularization. Simultaneously, the regularization parameter is estimated adaptively in each iteration based on reconstruction results. Finally, the amplitude decay of the reconstructed TFR is reduced by the de-biasing process. Compared with the traditional micro-motion echo time-frequency representation reconstruction method, the proposed method avoids the disadvantage of setting the regularization parameter manually, and the reconstructed TFR is more complete. The effectiveness and robustness of the proposed method is verified by simulation and measured data processing.

关 键 词：微多普勒效应 信号重构 稀疏优化 参数估计 时频分析 

分 类 号：TN95[电子电信—信号与信息处理]