基于多维变换的无线传感器网络定位算法分析-analysis of localization algorithm in wireless sensor network based on multidimensional transformation.docxVIP

以降为O(n2)，更适用于对实时性要求比较高的应用场合。此外，本文还提出了新颖的拼图分割算法对传感器网络进行有效的分割，再结合中心式定位算法实现分布式的传感器网络节点定位。MATLAB的仿真结果表明，这两种改进的中心式定位算法以及提出的分布式定位算法的性能符合理论推导，它们能够在不同的应用场合实现较高精度的节点定位。关键词：无线传感器网络，定位，多维变换，中心式，分布式RESEARCHONMDS-BASEDLOCALIZATIONALGORITHMSFORWIRELESSSENSORNETWORKSABSTRACTWithitsuniqueadvantagesandpotential,thetechnologyofwirelesssensornetworkshasbeenidentifiedasoneofthemostimportanttechnologiesinthe21stcentury.Asitssupportingtechnologiessuchaswirelesscommunication,micro-fabricationandintegration,andembeddedmicroprocessorhavegreatlyadvanced,theapplicationofwirelesssensornetworkshaspenetratedintoeveryaspectofhumanlives.Amongalltherelatedtechnologies,nodelocalizationisthemostfundamentalandkeyissueintheimplementationandmanagementofsensornetworks.Thedetectedinformationbyanodeisoftenmeaninglesswithoutitsknowledgeofposition.Therefore,todeterminethelocationofaneventoranodeisoneoftheunderlyingfunctionsofwirelesssensornetworks,whichplaysanessentialroleintheeffectivenessofthewholenetwork.Thispaperfirstbrieflyintroducesthefeaturesofwirelesssensornetworks,suchasitslargescale,one-timedeploymentandlimitedpower,whichmustbetakenintoaccountwhendoingrelevantresearchordesign.Then,thepaperposestheproblemofnodelocalization,carefullystudiescurrentlocalizationalgorithmsandsummarizesexistingrange-basedandrange-freelocalizationalgorithms.Moreover,thispaperanalyzesthedrawbacksofthetraditionalmultidimensionalscalinglocalizationalgorithmandproposestwoimprovedMDS-basedcentralizedlocalizationalgorithms.Thefirstimprovedalgorithmreplacesthepreviouseigen-decompositionwithcongruent-decomposition,simplifyingthestepsofmatrixdecompositionandenhancinglocalizationaccuracywhilemaintainingacomputationalcomplexityofO(n3).Thisalgorithmissuitableinscenarioswherehighlocalizationaccuracyisrequired.Thesecondimprovedalgorithm,usingforcedpositive-definiteCholeskydecomposition,canreducethecomputationalcomplexitytoO(n2)withlimitederror,whichfitswellwherereal-timeresponseistheprimaryconcern.Thispaperalsoproposesadistributedlocali