矿井无线传感器网络三阶段定位方法

391Vol.39No.120101JournalofChinaUniversityofMining&TechnologyJan.2010:2008211219:(863)(2006AA04Z208):(19752),,,,.E2mail:[e-mail]:13141201636,,,(1.(),,100084):,LQI,Takane(MDS),Anchor2free.,MDS3,3,.ZigBee.:MDSMDS,.:;;;ZigBee:TP393:A:100021964(2010)0120087206ANovelThree2PhaseLocalizationMethodinCoalMineWirelessSensorNetworksPEIZhong2min,DENGZhi2dong,WUTian2hua,XUXiao(TsinghuaNationalLaboratoryofInformationScienceandTechnology,DepartmentofComputerScienceandTechnology,TsinghuaUniversity,Beijing100084,China)Abstract:Anovelthree2phaselocalizationmethodforcoalminewirelesssensornetworksispresentedbasedonLQIandTakanesnon2metricMDS.Toaddresstheproblemsbroughtbyseverenaturalconditionsofcoalmine,thewholelocalizationprocessisdividedintothreepha2ses:regionallocalization,multi2dimensionalscalingandpreciselocalization.Byfusingthelo2calizationdata,anaccuratepositioncoordinateofmovingtargetsundercoalmineisobtainede2ventually.WesetupaZigBeenetworkandconductourexperimentsinarealair2raidsheltertunneltosimulatecoalmineenvironment.TheexperimentalresultsshowthatthelocalizationerrorofthemetricMDSalgorithmislessthanthatofmetricMDSalgorithms.Thethree2phaselocalizationmethodsolvesthelowlocalizationaccuracyproblemswhichcausedbyharshvaporsandthecomplexgeologicalstructureincoalmine.Keywords:wirelesssensornetworks;multi2dimensionalscaling;linkqualityindicator;ZigBeeprotocol,,.GPS,RFID.(WSN),WSN.[123]WSNWSN39WSN.,(C2WSN).WSN,TOA,TDOA,AOA,RSSI[4],(Range2free),APITDV2Hop[526].Anchor2free.Anchor2free.,(MDS)WSNAnchor2free.MDS,WSN,;[5]MDSWSN;[7]MDS;[8]MDSWSN..,WSN:1),.[7],,RSSICC243125m.2),.1a,A,B,AB4,AB.(a)(b)1Fig.1Obstaclecaseandtheshortestpath3)MDSWSN,MDS,.,,RSSI,MDS,RSSI.,,LQI(linkquali2tyindicator),TakaneMDS,C2WSN(MDS23P).3:1).,LQI.RSSI,LQI,,;2)MDS.MDS,DijkstraLQI,3,;3):CC2431,MDS,.1MDS111C2WSNG=(V,E),V,E,u,vV,,(u,v)E.WLQI,|LQI(u,v)|.uv,(u,v)|E,Dijkstra.1b21,w1,2.uLQI0,LQI(u,u)=0.W.,nC2WSNp(p=2p=3),i(i=1,2,,n)Ri=(Ri1,Ri2,,Rip),Ai=(Ai1,Ai2,,Aip).p=2.112MDSMDSTorgersonShepard2kruskalTakane,Torgerson,.,MDS881:,,.W,MDS().MDS:Step1(R0i1,R0i2,,R0ip)(i=1,2,,n),:MDS(),k0.Step2,dki,jpt=1(Rki,t-Rkj,t)2.(1)Step3W(dki,j)nn(^dki,j)nn,Pi,j,u,v,^dki,jdki,j+dku,v2(wi,jwu,vdki,jdku,v),dki,j(wi,jwu,vdki,jdku,v),(2)^dku,vdki,j+dku,v2(wi,jwu,vdki,jdku,v),dku,v(wi,jwu,vdki,jdku,v).(3)Step4S,S(10-4),;Step5.TakaneS=n-1i=1nj=i+1(d2i,j-^d2i,j)2n-1i=1nj=i+1d4i,j.(4)Step5kk+1,(Rki,1,Rki,2,,Rki,p)Rki,t=Rk-1i,t+n-1nj=1ji1-^dk-1i,jdk-1i,j(Rk-1j,t-Rk-1i,t),(t=1,2,,p),(5),Step2.2MDS23P211C2WSN2,[9].C2WSN3.C2WSN,...2C2WSNFig.2SystemarchitectureofC2WSN:1),WSN,,,(3);2)LQI,LQI.,:3C2WSNFig.3SchematicdiagramofC2WSNnodesdeploymentStep1,LQI,,LQIResponse.LQI,LQI,,,.Step2LQI,,LQI.,Response,LQI.Step3,LQI,.,.Step4LQI,9839,.:,LQI,,,.212MDSMDS,3,,:Step1DijkstraLQIW,MDS.Step2MDS,.Step33,,.m(3mn),,,.,(1),,((A2-A1)T,(A3-A1)T,,(Am-A1)T)T=((R2-R1)T,(R3-R1)T,,(Rm-R1)T)TQ,(6)Q.,M=((A2-A1)T,(A3-A1)T,,(Am-A1)T)T,(7)N=((R2-R1)T,(R3-R1)T,,(Rm-R1)T)T.(8),Q=(NTN)-1NTM.(9)Q,(ATm+1,ATm+2,,ATn)T=((Rm+1-R1)T,(Rm+2-R1)T,,(Rn-R1)T)TQ+(AT1,AT1,,AT1)T.(10)Step4.G.213CC2431TIWSN.RSSI,,RSSI,[8].P2=P1+10n(lgd1-lgd2),(11):d1,d2,m;P1,P2,dBm;n.Step1CC2431.MDS,,CC2431,.Step2LQI,MDS,MDS.Step33.,CC2431MDS,;CC2431MDS,.3311WSNCicada215,TICC2430/CC2431ZigBee,ZigBee,Mesh.5.,150m.2215,7618%,.1051,Zig2Bee.,,.,20m110,.MDS23P,LQI.LQIMDS,1,4,6,8,.,MDS23P.091:3121LQI,5,.110LQI.,LQI0,,LQI0.123,9054.1LQITable1LQIvaluesbetweeneachpairofnodesLQI10905400000002900985600000035498095450000040569509360000050045930750000600060750110000700000110058008000000750903990000005890094100000000399402DijkstraLQI,MDS.,1LQI,2LQI,.LQI0.2DijkstraLQITable2LQImatrixwithDijkstraalgorithmLQI1090541469917428435934239829009856143116226301284340354980954512023030528828441465695093601702452282845991434593075185260243299617411612060750110185168224728422623017018511007558114835930130524526018575090399342284288228243168589009410398340344284299224114399404MDS.4MDSFig.4RealcoordinatesandtheMDSestimatedones4,.,79.,MDS,[5].,67,.,,.313WSN.MDS23P.l=ni=m+1xiest-xireal2(n-m)R,(12):n;m;xiesti;xireali;R,R75m.1,2LQI,3MDS,.MDSW.MDSW(11).53MDS,.3MDS,10.Shepard2kruskalTakaneMDSMDS,.5,MDS,,C2WSN,,,,,.5MDSFig.5LocalizationerrorsfordifferentMDSalgorithms:1)LQI;2).2s,.,19392s,LQI,CC2431.CC2431,.4,LQIMDSC2WSN.:1)LQI,,;2)MDS,MDSC2WSN,C2WSN;3),3,3,.:[1]KENNEDYGA.Highresiliencewirelessmeshnet2workingcharacteristicsandsafetyapplicationswithinundergroundmines[D].Exeter:CambomeSchoolofMinesUKExeterUniversity,2006.[2]MOHANTYPK.Applicationofwirelesssensornetworktechnologyforminertrackingandmonito2ringhazardousconditionsinundergroundmines[EB/OL].UnitedStatesDepartmentofLaborMSHA,2006[2006207222].[3]ECEMultimediaSystemsLab.Real2timewirelesssensornetworkplatform[EB/OL].CarnegieMellonUniversity,2007[2007201212].[4]DANIELM,DAVIDT.Sensornetworksoccer[D].Switzerland:UniversityofAppliedSciencesRapper2swilHSR,2006.[5]JIXiang,ZHAHong2yuan.Sensorpositioninginwirelessad2ho