摘要
分析了弹道导弹突防过程中有源假目标航迹的运动学特性,指出和实体目标相比,真假目标运动学特性的差异是一种非显著性差异,这种非显著性差异可以体现在初始参数估计的残差上。对于实体目标而言,其最优估计的归一化残差为高斯分布,而假目标残差为非高斯分布,可以作为鉴别的依据。设计了初始参数的最大似然(maximum likelihood,ML)估计算法,采用Levenberg-Marquardt(L-M)迭代法进行求解,并利用估计残差设计了假目标的鉴别算法。该方法的优点是不仅可以鉴别假目标航迹,而且对真目标而言,该算法同时是最优定轨算法。仿真结果验证了算法的有效性。
The kinematical characteristics of active decoy trajectories used for midcourse penetration are analyzed.It is indicated that the kinematical difference between decoys and physical targets is non-significant and this non-significant difference can be reflected by the estimated residuals of initial parameters.For physical targets,the normalized estimated residual is Gaussian;while for decoys,the residual is non-Gaussian,this can be served as a basis for discrimination.Then,the maximum likelihood（ML） estimate problem of the initial parameter is presented and a Levenberg-Marquardt（L-M） iterative algorithm is employed to obtain the parameter.The discrimination algorithm is designed based on the normalized residual.The advantage of the method is that it can not only discriminate the decoys but also determinate the orbit of true targets precisely.Finally,the simulation verifies the validity of the algorithm.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2010年第6期1195-1200,共6页
Systems Engineering and Electronics
基金
国家自然科学基金(60875019
60672033)资助课题
