【摘要】 本文提出了一种新的Robust序贯检测方法——无记忆批处理Robust序贯方法(RMLGS)。导出了它的一般结构。推证了它在虚警概率、检测概率和平均抽样个数等方面的坚韧性,并讨论了它的渐近性能。最后对p-点分布族给出了检测器的具体结构和有限样本性能。这种检测器结构简单,样本利用率高且对环境要求宽。更多还原

【Abstract】 A memoryless grouped-data robust sequential detection procedure is proposed. Its robustness including false alarm probability, detection probability and average sample number is demonstrated and its asymptotic performance is discussed. Finally, the performance curves in small sample sizes are given for the p-point distribution class. The detector exhibits simplicity in structure and retains sample-efficiency of sequential detector.更多还原