摘要
目标回波的多普勒和方位参数能体现目标的重要特征.本文从参数的最大似然估计出发,采用非线性来放大似然函数并构造一个新的代价函数,该函数能更明显区分局部和全局最佳值.针对新代价函数的复杂形式,我们提出了基于蒙特卡洛的重要性采样方法来求解最佳值,并给出了重要性函数的一种选取方法.该方法不仅在低信噪比和小样本数条件下性能良好,而且各个参数自动配对并适用于任意的已知阵型.与已有的基于子空间的联合估计算法的仿真比较表明,本文提出的算法在性能上非常接近参数估计的CRLB.
The DOAs and Dopplers of the echoed signal from targets are very important parameters in some applications for radar and sonar.This paper proposes a new cost function which is based on the likelihood function amplified with a nonlinear function,and the global convergence of this function can be obtained more easily than the original cost function.An Importance Sampling（IS） technique based on Monte Carlo method is proposed to find the optimum of this cost function,and the suitable importance function is chosen to estimate these parameters. All parameters are paired automatically and this technique can be applied on the arbitrary known array shape. Computer simulations show that the proposed method is superior to other methods based on subspace decomposition under the condition of low SNR and small snapshots, and the MSEs of Dopplers and DOAs in this method are very close to CRLB.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2009年第9期1965-1970,共6页
Acta Electronica Sinica
基金
国家自然科学基金(No.60702067)
关键词
多普勒方位联合估计
蒙特卡洛方法
重要性采样
最大似然估计
joint estimation of dopplers and DOAs
monte carlo method
importance sampling
maximum likelihood estimation（MLE）
