基于最大似然差的智能恒虚警检测器

摘要

为了增强恒虚警(constant false alarm rate,CFAR)检测器在杂波边缘环境中的鲁棒性,结合无偏非均匀杂波估计CFAR(HCE-CFAR)检测器检测概率高和可变性指示CFAR(VI-CFAR)检测器虚警控制能力强的优点,提出一种基于最大似然差(maximum likelihood difference,MLD)的智能CFAR检测器MLD-CFAR.MLDCFAR通过计算MLD和均值比判断前、后沿滑窗的杂波环境,进而选择相应参考单元和均值类CFAR算法计算检测门限.仿真结果表明,同HCE-CFAR、VI-CFAR和自动删除单元平均CFAR (ACCA-CFAR)检测器等相比,MLD-CFAR在杂波边缘环境中能够保持恒虚警率,提高目标检测概率,减少目标遮蔽现象,在均匀环境中检测损失很小.论文还在一部线性调频连续波雷达上用实测数据验证了算法的正确性和有效性.%An intelligent constant false alarm rate (CFAR) detector with better robustness performance is proposed. This CFAR detector is based on maximum likelihood difference (MLD) and is named a MLD-CFAR detector. In this detector, the heterogeneous clutter estimation CFAR (HCE-CFAR) with a high detection probability and the variability index CFAR (VI-CFAR) with a low false alarm rate are combined. The MLD and mean ratio are computed to distinguish the clutter environments of the leading and lagging windows, and then the corresponding reference cells and the mean level CFAR algorithm are selected adaptively to calculate the detection threshold. The performance of the proposed MLD-CFAR detector is evaluated and compared with other CFAR detectors by numerical simulation. It shows that the MLD-CFAR keeps a constant false alarm rate with higher probability of detection and lower probability of target masking in clutter edge environments, and is with a lower CFAR loss in homogeneous clutter environments. Experimental results from a linear frequency modulated continuous wave radar system verify the correction and efficiency of the proposed MLD-CFAR detector.