论文部分内容阅读
本文根据修正的奈曼-皮尔逊准则,提出了一种Robust-M检测器。在噪声分布发生变化或噪声电平增大的情况下,导出的最佳检测器在恒虚警性能方面明显优于经典检测器。此外,对于给定的虚警概率上限,该检测器是最大功效的。所用的检验统计量是一个信号参数θ的Rubost-M估计量,它可用牛顿迭代法求出,用递归结构实现。
In this paper, we propose a Robust-M detector based on the modified Neyman-Pearson criterion. In the case of a change in the noise distribution or an increase in the noise level, the derived optimal detector outperforms the classical detector in terms of CFAR performance. In addition, the detector is the most efficient for a given upper false alarm probability ceiling. The test statistic used is a Rubost-M estimator of the signal parameter θ, which can be found using the Newton’s iterative method, using a recursive construct.