两种基于局部广义符号统计量的融合准则  

Two types of fusion rules based on a local generalized sign test statistic

作　　者：刘向阳[1] 彭应宁[1] 关键[2] 

机构地区：[1]清华大学电子工程系 [2]海军航空工程学院电子工程系,烟台264001

出　　处：《清华大学学报（自然科学版）》2006年第7期1223-1226,共4页Journal of Tsinghua University(Science and Technology)

基　　金：全国博士学位论文作者专项资金资助项目(200237)

摘　　要：为提高基于二元局部判决的分布式广义符号检测器的检测性能,研究了基于广义符号统计量的分布式检测,提出了直接求和与删除求和两种融合准则。在W e ibu ll杂波和Sw erlingII目标下,直接求和融合具有比某参数化的分布式恒虚警检测更好的检测性能,并能达到集中式广义符号检测相同的性能。某些传感器信杂比极低时,采用删除求和融合能获得比直接求和融合更好的检测性能。给出了在多雷达分布式检测中应用各种融合准则的方法。Two types of local generalized sign statistics-based distributed detection fusion schemes, direct summation and censored summation, are presented to improve the detection power of binary local decision-based distributed generalized sign detection in multiradar distributed detection systems. Numerical tests indicate that the direct summation fusion rule provides better performance than the binary local decision-based rule or previous parametric distributed CFAR schemes with Weibull clutter. The direct summation fusion scheme has the same detection power as the centralized generalized sign detector. If some sensors＇ signal-to-clutter ratios are very low, the censored summation fusion has better performance than direct summation. Applications in multiradar distributed detection are also presented.

关 键 词：多雷达分布式检测 分布式检测 广义符号检测 局部检测统计量 

分 类 号：TP212.9[自动化与计算机技术—检测技术与自动化装置] TN957.51[自动化与计算机技术—控制科学与工程]