非参量量化秩检测器的性能研究

 许多作者讨论过非参量秩检测器在雷达信号处理中的应用。秩检测器首先把接收波形样本转换为秩。如果检验单元和参考单元的噪声样本独立和分布,则无信号时检验单元的秩具有离散均匀分布,与输入噪声的分布无关。所以秩检测器可能提供分布自由的恒虚警率性能。量化秩检测器（QRD）只对二进量化秩进行积累,所以它实现起来很经济。本文分析QRD的检测性能。证明QRD有一最佳秩量化门限（ORQT）。确定高斯和韦伯噪声中的ORQT。另外,把QRD同高斯噪声中的局部最佳秩检测器和最佳参量检测器进行比较。

STUDY ON PERFORMANCE OF NONPARAMETRIC QUANTIZED RANK DETECTORS

The detection performance of quantized rank detectors （QRD）has been investigatel. These QRD are shown to be in the same class of structure with the locally optimum rank detector, but its implementation is easy. We obtain the efficacy of QRD in Gaussian noisewhere N is the number of reference cell, T, is the rank quantization threshold and σf is the average power of noise. The optimum rank quantization threshold （ORQT） which maximizes the eOR in Gaussian noise is approximatelyThe threshold value should be increased for a finite number of pulses integrated, and reduced in Weibull noise.Also, the performance of the quantized rank detector is compared with that of the locally optimum rank detector and the optimum parametric detector for Gaussian noise. The difference of asymptote performance between the QRD and MSD （Modified Savage Detector） is less than IdB.When Gaussian noise is considered some losses in signal-to-noise ratio are incurred in the QRD relative to the traditional SLD （square law detector） and LD （linear detector）. But the loss decreases and can even become a gain when the noise distribution has a longer tail.