基于自动删除算法的广义变化性指标恒虚警检测器

Generalized Variability Index CFAR Detector Based on Automatic Censored Algorithm

针对自动删除均值检测器存在的排序时间过长的问题,先提出了一种修正自动删除均值检测器.计算机仿真显示,与自动删除均值检测器相比,该检测器的性能或与其相同,或稍有改善,但排序时间却减少了一半.推导了该检测器的平均虚警概率、平均检测概率的解析表达式.比较了该检测器与基于自动删除算法的最小选择恒虚警检测器的性能,分析了这2种检测器各自适用的场合.为了得到稳定的检测性能,提出了一种广义变化性指标恒虚警检测器,该检测器根据变化性指标可以自动在修正自动删除均值检测器和基于自动删除算法的最小选择恒虚警检测器之间转换.最后通过计算机仿真验证了该检测器的有效性.

In order to reduce the sorting time of automatic censored mean level detector, a modified automatic censored mean level detector is proposed first. Computer simulations indicate that compared to automatic censored mean level detector,its performance is either same or slightly improved, but the sorting time is reduced to half. The analytic expressions of average false alarm and average detection probabilities are derived for the detector. The performance of the detector is compared with the smallest of constant false alarm rate（CFAR） detector based on automatic censored algorithm. Application field of two detectors are analyzed. To obtain robust performance,a generalized variability index CFAR detector is proposed. It dynamically switches to the modified automatic censored mean level detector or the smallest of CFAR detector based on automatic censored algorithm depending on the variability index. Validity of the detector is tested by computer simulation.