基于窗口统计量的水下分布式目标检测算法.pdf

《基于窗口统计量的水下分布式目标检测算法.pdf》由会员上传分享，免费在线阅读，更多相关内容在应用文档-天天文库。

1、计算机研究与发展DOI：10．7544／issn1000—1239．20l4．20121084JournalofComputerResearchandDevelopment51(8)：1880—1887，2014基于窗口统计量的水下分布式目标检测算法刘义海张效民沈殷隽于洋(西北工业大学航海学院西安710072)(1iuyihai@126．com)ADistributedUnderwaterTargetDetectionAlgorithmBasedonWindowStatisticsLiuYihai，ZhangXiaomin，ShenYinj

2、un，andYuYang(CollegeofMarine，NorthwesternPolytechnicalUniversity，Xi’an710072)AbstractForanunderwatersensornetwork(USN)withrandomlydeployedsensors，localsensors’time—varyingdetectionperformanceandshortageoftheintrusiontargetpriorimodel，adistributedtargetfusiondetectionalgori

3、thmisproposedbasedonoptimalwindowstatistics．Thisalgorithmfocusesonthepracticalpassiveunderwatertargetdetectionthatonlythesensorssurroundingthevesseltargetinasmallzonecouldprovidestablelocaldetectionresults．Thusthedetectionprocessiscarriedoutwithamoveablevirtualwindowwhichf

4、usesthebinarydecisionsreportedbylocalsensorsinsideitscoveragewiththecountingfusionrule．Finallythedetectionofthisparticularsubareawiththelargestnumberoffusionsensorreport“1”isequivalenttothatofthepointtarget．Comparedwiththepointtargetdetectionproblem，theextendedareadetectio

5、nismorerobustandreliable．Theapproximatedetectionprobabilityofthesystemlevelisderivedanalytically．Simulationmethodsarealsoemployedtocomparetheapplicationperformancebetweentheproposedalgorithmandtheexistingnonparametricvotingorcountingfusionrulesunderpracticalscenarios．Resul

6、tsillustratethatthisfusionrulecanperformbetterinsystem—leveldetectioncomparedwiththeexistingone，aslongasthescanningwindowsizeapproximatlymatchesthetargetradiationsignalregion．Keywordsunderwatersensornetwork(USN)；targetdetection；informationfusion；optimalwindowstatistics；hyp

7、othesistest摘要针对水声传感器网络对水下目标检测时面临的节点数目、布放位置随机、检测性能时变、缺乏入侵目标先验模型的问题，将对点目标的假设检测推广到对最优海域窗口的假设检测，提出了一种基于最优窗口统计量的融合检测规则，近似推导出了算法系统级的检测性能，并给出了仿真对比实验．结果表明：在满足滑动窗口同目标辐射信号区域近似匹配的条件下，基于最优窗口统计量的融合检验规则可以获得良好的系统级检测性能，与已有的非参数类投票计数融合规则相比，相同信噪比下，基于最优窗口统计量的融合规则目标检测性能更好．关键词水声网络；目标探测；信息融合；最优窗

8、口统计量；假设检验中图法分类号TP393；TJ67近年，水下声学网络作为空中无线传感器网络的一项扩展得到广大研究者的关注．相对于传统(wirelesssensornetwork，