高斯混合分布杂波下ARM的Wald检测方法  

Wald Detection Method of ARM in Gaussian Mixed Clutter Distribution

作　　者：方前学[1,2] 王永良[3] 王首勇[2] 

机构地区：[1]国防科技大学电子科学与工程学院,长沙410073 [2]空军雷达学院重点实验室,武汉430019 [3]空军雷达学院科研部,重点实验室武汉430019

出　　处：《现代雷达》2010年第2期62-65,70,共5页Modern Radar

摘　　要：提出一种在高斯混合分布杂波下检测反辐射导弹(ARM)的Wald检测方法。针对基于期望最大化(EM)算法估计杂波参数时,由于初始化不当使迭代运算落入初值陷阱、导致估计错误的问题,提出基于矩-EM算法估计杂波参数的方法,导出了高斯混合分布杂波下ARM目标的Wald检测统计量。不同参数条件下的仿真表明,矩-EM算法能够更准确地估计杂波参数;基于高斯混合分布杂波假设的Wald检测性能明显优于基于高斯分布杂波假设的Wald检测性能。In this paper, a Wald detection method is proposed to detect the anti-radiation missile （ARM） in Gaussian mixed clutter distribution. When we estimate the distribution parameters with the expectation maximization （EM） iterative algorithm, the iterarive algorithm may end due to initialization trap and lead to error estimation if there is not an appropriate initialization. So a moment-EM algorithm is suggested to estimate the parameters. The Wald decision rule of the detection of ARM target under ganssian mixed clutter distribution is derived. Simulation experiments with different parameters show that the moment-EM algorithm can esti- mate the parameters more accurately. And the detection performance of Wald detection method based on gaussian mixed distribution significantly outperforms Wald method based on Gaussian clutter distribution.

关 键 词：高斯混合分布 矩-EM算法 反辐射导弹检测 Wald检测 

分 类 号：TN957.51[电子电信—信号与信息处理]