基于最大和最小选择的两种新的恒虚警检测器

Two New CFAR Detectors Based on Greatest of and Smallest of Selection

本文提出两种基于有序统计（ＯＳ）和单元平均（ＣＡ）产生局部估计，并分别应用最大选择（ＧＯ）和最小选择（ＳＯ）产生检测单元干扰功率水平估计Ｚ的新的恒虚警检测器（ＯＳＣＡＧＯ）和（ＯＳＣＡＳＯ）。我们推导出了它们在ＳｗｅｒｌｉｎｇⅡ型目标假设下的Ｐ＿（ｆａ），Ｐ＿（ｃｌ）和度量ＡＤＴ的解析表达式。分析了它们在均匀背景和强干扰目标环境之中的检测性能，并把它们与ＯＳ－、ＧＯＳＧＯ－、ＧＯＳＳＯ－ＣＦＡＲ进行了比较。结果表明，ＯＳＣＧＯ在均匀干扰背景和多目标情况下的检测性能与ＯＳ和ＧＯＳＧＯ相比，都有很明显的提高。在干扰目标数为某些值时，ＯＳＣＡＧＯ的ＣＦＡＲ损失比ＧＯＳＧＯ小近３ｂＢ。在均匀背景中，ＯＳＣＡＳＯ的检测性能与ＧＯＳＳＯ相近。但是，在多目标环境中，一定条件下的ＯＳＣＡＳＯ的检测性能甚至超过了ＯＳ和ＧＯＳＧＯ。

This paper presents two new CFAR detectors based on Ordered Statistions(OS)and Cell-Averaging(CA)forming local estimates,using Greatest-of(GO)and Smallest-Of(SO) selection respectively to form interference power level estimate Z in test cell(OSCAGO.OSCASO).Under the Swerling Ⅱ assumption.we derive the analytic expressions of P_(fa).P_(cl) and ADT of the two detector,analyze its detection performance in homogeneous background and in strong interfering targets envi-ronment.and compare them with OS- GOSGO- GOSSO-CFARs. Results show that the detection performance of OSC AGO in homogeneous background and multiple target situations are all obvi-ously better than that of OS and GOSGO When the number of interfering targets is equal to certain value the CFAR losses of OSCAGO are about 3bB less those of GOSGO In homogeneous background.OSCASO possesses the similar performance to GOSSO But in multiple target situa-tions,the performance of OSCASO under certain conditions even exceeds that of OS and GOSGO