光学学报, 2017, 37 (4): 0412004, 网络出版: 2017-04-10   

基于均值似然估计的激光探测微动特征提取和分离

Extraction and Separation of Micro-Motion Feature Based on Mean Likelihood Estimation in Laser Detection
作者单位
解放军电子工程学院脉冲功率激光技术国家重点实验室,电子制约技术安徽省重点实验室, 安徽 合肥 230037
摘要
最大似然估计是提取目标微动特征参数的最佳估计方法,但直接用网格法求解计算量巨大,且激光探测微多普勒回波信号对应的代价函数具有高度非线性,存在多个局部最大值。为此,提出均值似然估计与蒙特卡罗结合的估计方法,给出了最大似然参数估计的闭合表达式,再通过设计压缩似然函数获得全局最大值,通过蒙特卡罗法抽样并计算循环均值估计出参数。该方法避免了传统方法中对高精度初始值和复杂迭代算法的依赖,能够实现参数的联合估计。对于多分量微多普勒信号,该方法可在参数估计的同时实现各微动分量分离,且不增加算法的复杂性。对仿真和实验数据进行估计,结果表明,该方法在达到近似于最大似然估计性能的同时可有效降低计算复杂度并确保了全局收敛,实现信号的分离和参数估计。
Abstract
Maximum likelihood estimation (MLE) is the optimal estimator for target micro-motion feature parameter extracting. However, the grid search will cause the enormous computational amount, and the cost function of laser detection of micro-Doppler echo signals has high nonlinearity and exists many local maxima. A new method combining the mean likelihood estimation and the Monte Carlo method is proposed to solve this problem. A closed-form expression of maximum likelihood parameter estimation is derived. Then the compressed likelihood function is designed to obtain the global maximum. The parameters are estimated by Monte Carlo method sampling and calculating the circular mean value. The dependence of hight accurate initial values and the complex iteration algorithms are avoided in this method, and the joint estimation of parameters can be realized. Furthermore, for multi-component micro-Doppler signal, the presented algorithm can separate the micro-motion component signals at the same time with the estimations, which will not add complexity of algorithm. Applied to the simulated and experimental data, the proposed method achieves similar performance as MLE with less computational complexity. Meanwhile, this method guarantees the global convergence and realizes signals separation and parameters estimation.

郭力仁, 胡以华, 王云鹏. 基于均值似然估计的激光探测微动特征提取和分离[J]. 光学学报, 2017, 37(4): 0412004. Guo Liren, Hu Yihua, Wang Yunpeng. Extraction and Separation of Micro-Motion Feature Based on Mean Likelihood Estimation in Laser Detection[J]. Acta Optica Sinica, 2017, 37(4): 0412004.

本文已被 1 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!