子空间不确定下多重假设AMF、Rao与Wald检测方法

Multiple Hypotheses AMF,Rao and Wald Detection Method Under Subspace Uncertainty

由于目标多秩子空间大小的不确定性会导致多重假设检测发生,传统目标自适应二元检测方法不再适用。针对此问题,文章提出了子空间不确定下多重假设AMF、Rao与Wald检测方法。首先,基于Kullback-Leibler信息准则,建立了目标多秩子空间存在多种假设下的目标检测模型;然后,基于AMF、Rao和Wald检测准则,设计多重假设检测器,并优化估计未知参数与计算惩罚项。最后,通过仿真实验验证了所提检测器的性能,并分析了惩罚项对各检测器性能的影响。实验结果表明,相比传统检测器,所提检测器在一定情况下具有更优的检测性能。

Since the uncertainty in the size of the target multi-rank subspace will lead to multiple hypotheses detection,the traditional target adaptive binary detection method is no longer applicable.To address this issue,a multiple hypothesis AMF,Rao and Wald detection method under subspace uncertainty is proposed.Firstly,based on the Kullback-Leibler information criterion,a target detection model under multiple assumptions in the target multi-rank subspace is established.Then,based on the AMF,Rao and Wald detection criteria,multiple hypotheses detectors are designed,and the unknown parameters are optimized and the penalty term is calculated.Finally,the performance of the proposed detectors are verified by simulation experiments,and the influence of the penalty term on the performance of each detector is analyzed.The experimental results show that compared with the traditional detector,the proposed detectors have better detection performance under certain conditions.