基于最优最小二乘的红外目标单站定位算法  

Optimal Total Least Squares Algorithm for Infrared Target Localization with Single Station

作　　者：吴鑫辉[1] 黄高明[1] 高俊[1] 

机构地区：[1]海军工程大学电子工程学院,湖北武汉430033

出　　处：《南京理工大学学报》2012年第4期674-678,共5页Journal of Nanjing University of Science and Technology

基　　金：国家'863'计划资助项目(2010AA7010422);国家自然科学基金(60901069);中国博士后科学基金(20080431379;200902671)

摘　　要：为消除红外目标单站定位算法估计有偏性,解决现有估计算法中有可能出现多个解向量的问题,提出了一种基于最优最小二乘的目标估计算法。对总体最小二乘定位算法进行了研究,针对扩展测量矩阵出现最小奇异值多重的问题,构造了一个由多重奇异特征向量组成的测度函数,将该函数进行偏微分得到最优的最小二乘解;通过理论分析,证明了最优最小二乘解向量是渐进无偏的;最后,将该算法与总体最小二乘法和修正辅助变量法进行仿真对比。仿真实验结果表明,该方法在不同测量环境噪声下所得到的位置、速度误差曲线能快速地逼近克拉美罗下限,且具有更高的估计精度和稳定性,说明了算法的有效性。In order to eliminate the biased errors from the bearings-only infrared target localization estimators and solve the problem of the existence of multiple solutions in the available algorithms, a bearings-only infrared target localization algorithm based on the optimal total least squares is proposed. With the study of the total least squares theory, a cost function is constructed to solve the problem for identical least singular values with extended measurement equations. A theoretical analysis of the optimum total least squares algorithm proves the simulated results are unbiased. Compared with the total least squares algorithm and modified instrumental variable algorithm, the proposed algorithm can reach the Cramer-Rao lower bound （ CRLB ） quickly under the Gaussian noise, proving better estimation accuracy and stability.

关 键 词：红外目标定位 总体最小二乘法 最优最小二乘法 克拉美罗下限 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]