移动机器人的同时定位和地图创建方法

36720047JOURNALOFHARBININSTITUTEOFTECHNOLOGYVol136No17July,2004厉茂海,洪炳,罗荣华(,150001,E-mail:[e-mail]):通过声纳传感器实时地创建和修正地图,局部地图是基于栅格的概率模型:工作环境被划分成栅格,每一个栅格赋一个值,标识栅格中有障碍物的概率.地图应用哈夫变换创建,定位通过对超声波数据应用滤波算法并结合地图创建实现.使用Pioneer2移动机器人在实际的室内环境中的实验,验证了该方法的可行性.:同时定位和地图创建;声纳传感器;基于栅格的概率模型;哈夫变换:TP24:A:0367-6234(2004)07-0874-03SimultaneouslocalizationandmapbuildingformobilerobotLIMao-hai,HONGBing-rong,LUORong-hua(SchoolofComputerScienceandTechnology,HarbinInstituteofTechnology,Harbin150001,China,E-mail:[e-mail])Abstract:Amapisbuiltandupdatedimmediatelythroughsonarsensor.Thelocalmapisagrid-basedprobabilisticmodel:theworkenvironmentisdecomposedintogridsandeverygridtakesupavalue,whichindicatestheproba-bilityofanobstacleingrids.ThemapisbuiltthroughHoughtransform.Themapbuildingcombineswiththesonardataobtainedfromafilteralgorithmtorealizethelocalization.Experimentalresultsshowthefeasibilityofthismethod.Keywords:simultaneouslocalizationandmapbuilding;sonarsensor;grid-basedprobabilisticmodel;Houghtransform:2004-04-26.:(2002AA735041).:(1980-),,;(1937-),,,,.(S-imultaneousLocalizationandMapBuilding:SLAM):,,,,.SLAM,,[1][2],[3][4].,Thrun[5],.Ya-mauch[6].,,.1建模111Pioneer2.,,,Pan-Tilt-Zoom(PTZ),,(ax,ay,ath)T.Ca=(ax,ay,ath)TCr=(x,y,H)T.(0,0,0),,X=(ax,ay,ath)T.x,y.:Xnext=axayath=X+cosH0sinH001#vw#t+X=f(X,u,t,X).:u=[vw]t,vt;X;t;X;H.,,X=10-3.112,,.i/0,wi=[aTaxcaycal]T,,Ca=(ax,ay,ath)T,av,axc,ayc,a1.:W=[w1Tw2T,]T.2地图创建算法.,,.Step1:X0=(ax,ay,ath)T=(0,0,0)T,,,.Step2:,,Step3,Step4.Step3:,,,,,Step1.Step4:,,,,,Step1.1,123,,,,4.1211:.i,(axr,ayr,athr)T,Crsjr,j=0,,,15,xjr=(sjr+r)cos(ajr),yjr=(sjr+r)sin(ajr),ajr=2215#j.,r,ajrj.,(aRx,aRy,aRth)T,(xjr,yjr,ajr)T,,(ax,ay,ath)Taxay=cosaRth-sinaRthsinaRthcosaRth#xrjyrj+aRxaRy,ath=aRth+ajr.212i,,,.i,,i+1,,i,.,,1,,.,,,.4m,,315m.#875#7,:213(HoughTransform:HT).x-y()Q=xcosH+ysinH,H-Q(),.,.(xi,yi)(H,Q):xisinH+yicosH=Q.(1)(HT).HQ,A(H,Q).HHk,k=1,,,n,Hk-Hk-1=$H.(xi,yi)(1)nHkQ.QmQk,k=1,,,mQk-Qk-1=$Q,A(Hk,Qk)..(HT):1)HQ,,,,$H$Q,.2),,,,.3定位算法Step1:tmt,.Step2:t-1,.„pt-1,mt„pt-1,Step3,Step4.Step3:t-1„pt-1tmt„pt,Step1.Step4:,,Step1.Pioneer2,,.4结论1)..,,.2).,.参考文献:[1]SCHULTZAC,ADAMSW.Continuouslocalizationusingevidencegrids[J].Proceedingsofthe1998IEEEInterna-tionalConferenceonRoboticsandAutomation.1998,4:2833-2839.[2]AYACHEN,FAUGERASOD.Maintainingrepresentationsoftheenvironmentofamobilerobot[J].IEEETransRobotAutomat,1989,5(6):804-819.[3]KYUOERSBJ.Thespatialsemantichierarchy[J].ArtificialIntelligence,2000,119:191-233.[4]DOUCETA,FREITASN,GORDANN.SequentialMonteCarloMethodsinPractice[M].Springer-Verlag,2001.[5]THRUNS,FOXD,BURGARDW.Probabilisticmappingofanenvironmentbyamobilerobot[J].Proceedingsofthe1998IEEEInternationalConferenceonRoboticsandAu-tomation,1998,4:1546-1551.[6]YAMAUCHIB,SCHULTZA,ADAMSW.Mobilerobotex-plorationandmap-buildingwithcontinuouslocalization[J].Proceedingsofthe1998IEEEInternationalConferenceonRoboticsandAutomation.1998,4:3715-3720.(王小唯)#876#36