基于Wald检验的海面雷达目标贝叶斯检测方法  

Wald-based Bayesian detection for sea marine targets

作　　者：薛健 朱圆玲 潘美艳 孙孟玲 XUE Jian;ZHU Yuanling;PAN Meiyan;SUN Mengling(School of Communications and Information Engineering,XiJan University of Posts and Telecommunications,Xi'an 710121,China;Xi'an Electronic Engineering Research Institute9Xi'an 710100,China)

机构地区：[1]西安邮电大学通信与信息工程学院,陕西西安710121 [2]西安电子工程研究所,陕西西安710100

出　　处：《西安邮电大学学报》2022年第3期30-38,共9页Journal of Xi’an University of Posts and Telecommunications

基　　金：国家自然科学基金项目(62201455);陕西省教育厅科研计划项目(22JK0566)。

摘　　要：为了改善非均匀非高斯海杂波背景下海面雷达目标的检测性能,提出了一种基于Wald检验的海面雷达目标贝叶斯检测方法。首先,利用逆高斯分布和逆复Wishart分布分别建模海杂波的纹理分量和散斑协方差矩阵。其次,在假设散斑协方差矩阵已知的条件下基于Wald检验设计了非自适应的雷达目标检测器。最后,利用辅助数据估计的散斑协方差矩阵代替其真实值,得到了自适应的雷达目标贝叶斯检测器。仿真结果表明,在不同的形状参数和不同数量的参考数据条件下,提出的检测器优于相关经典检测器。提出的自适应贝叶斯雷达目标检测器提高了非均匀逆高斯纹理复合高斯海杂波环境中雷达目标的检测性能。In order to improve the detection performance of marine radar targets in non-homogenous and non-Gaussian sea clutter,a Bayesian detection method for radar targets based on the Wald test is proposed.First,the texture component and the speckle covariance matrix of sea clutter are modeled using the inverse Gaussian distribution and the inverse complex Wishart distribution,respectively.Second,a non-adaptive radar target detector is designed based on the Wald test under the assumption that the speckle covariance matrix is known.Finally,an adaptive Bayesian detector for radar targets is obtained using the speckle covariance matrix estimated from the reference data instead of its true value.The simulation results show that the proposed detector outperforms the related classical detectors for different shape parameters and different number of reference data.The proposed adaptive Bayesian radar target detector improves the detection performance of radar targets in non-homogenous compound-Gaussian sea clutter with inverse Gaussian texture.

关 键 词：海杂波 雷达目标 自适应贝叶斯检测 WALD检验 逆复Wishart分布 

分 类 号：TN959.72[电子电信—信号与信息处理]