第30卷第12期.PDF

第30卷第12期 哈尔滨工程大学学报 V01．30No．12 doi：10．3969／j．issn．1006—7043．2009．12．013 消 张晓平1，刘桂雄1，何学文12 (1．华南理工大学机械与汽车工程学院，广东广州510640；2．江西理工大学，江西赣州341000) 摘要：针对目标发射功率变化下的无线传感器网络(WSN)目标定位问题，分析了无线信道衰减特性，探讨目标功率无 关的信号强度差特征提取方法，结合WSN信息交换与处理过程，提出能消除WSN目标功率变化影响的信号强度差LSS- 特征向量与目标坐标映射关系的LSSVR模型，将各节点目标信号强度测量值的差值所构造特征向量输入LSSVR模型可 LMSD方法能显著减小信号强度值变化对目标定位结果的影响，提高目标定位准确度，并具有较好的实时性能． 关键词：无线传感器网络；目标定位；信号强度；最dx-乘支持向量回归机 中图分类号：TP301文献标识码：A文章编号：1006-7043(2009)12-1414-06 I_SSVRlocalizationmethodinWSN signalstrength target moc哑on differencetoeliminatethe influence targetpower ZHANG Xue．wenl，2 Xiao—pin91，LIUGui—xion91，HE ofMechanicalandAutomotive China of (1．School Engineering，SouthUniversityTechnology，Guangzhou of Scienceand 341000，China) University Technology，Ganzhou betterlocate whenthe ofwirelesschannel Abstract：To targets targettransmittingpowerchanges，characteristics were offeaturesof differencesunrelatedwiththe attenuation analyzed．Extractionsignalstrength was forwirelesssensor wasalsoexamined．Anovel localizationmethod power target proposed informationand basedonleast vector combining exchangeprocessing，and squaressupport of differences．Withthis received differencesbetweendiffer— method，mean modelingsignalstrength signalstrength sensors theinitialfeaturevector．ThenLSSVRmodelswereobtainedLSSVRmodel． ent wereusedtocompose