欢迎来到天天文库
浏览记录
ID:39402351
大小:1.05 MB
页数:65页
时间:2019-07-02
《基于多传感器信息的移动机器人定位研究》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库。
1、浙江大学机械工程学系硕士学位论文基于多传感器信息的移动机器人定位研究姓名:王晓娟申请学位级别:硕士专业:机械电子工程指导教师:王宣银20100101浙江大学硕士学位论文AbstraetAbstractMobilerobotlocalizationiSthebasicandmostimportantfunctionofautonomousnavigation,andalsoachallengingmainresearchsubjectwhichgetsmostoftheattentioninthefieldoftherobotics.
2、Thedissertationcentersaroundposetrackinginknownstructuredenvironment,Multi—sensorinformationfusionalgorithmisappliedtotheintegratedpositioningsystembasedonodometerandmulti-sonar.Thepaperincludescontentsasfollows:Firstly,consideringstructuralfeaturesandfunctionalrequire
3、mentsoftheself-designedmobilerobot,anideaofdoubleCPUbasedonDSPTMS320F2808andATmega32isadopted.Thedissertationgivesdetailintroductiononhardwaredesignofthesystem.Secondly,workprincipleofodometerandsonarisintroducedanderrorresourceisanalyzed,onthebasisofwhichthearckinemat
4、icmodelofodometerandthemeasurementmodelofsonarforlinefeaturearemade.Thirdly,toaddresstheaccumulativeerrorproblemcausedbyodometer,alocalizationalgorithm,whichemploysextendedkalmanfilter(EKf)tofuseodometerreadingsandsonardata,isproposed.Itfirstemploysodometerandgivenmapt
5、oobtainpredictedvalueofthefeatures,thenextractsthelinefeaturesinenvkonmentfromsonardataandperiodicallycorrectsthecumulativeerrorofodometerviamatchingandupdatingprocessbetweenrealobservationandpredictiontoimprovepositionprecision.Fourthly,throughtheanalysisoftheeffectso
6、fobservationnoise’SstatisticcharacteristicsontheperformanceofEKF,afuzzylogicbasedadaptiontuningalgorithmispresented.ThismethodCanon-lineadjusttheobservationnoisecovariancebymonitoringthecoincidenceoftheactualcovariance、^ritllthetheoreticalcovarianeeofresiduals,SOastoov
7、ercomethebadeffectsofincompleteknowledgeaboutitandon-lineimprovetheperformanceoflocalizationmethod.Keywords:mobilerobotlocalization,sonar,odometer,extendedkalmanfilter,fuzzylogicIII浙江大学研究生学位论文独创性声明本人声明所呈交的学位论文是本人在导师指导下进行的研究工作及取得的研究成果。除了文中特别加以标注和致谢的地方外,论文中不包含其他人已经发表或撰写过
8、的研究成果,也不包含为获得浙婆太堂或其他教育机构的学位或证书而使用过的材料.与我一同工作的同志对本研究所做的任何贡献均已在论文中作了明确的说明并表示谢意。学位论文作者签名:签字日期:印/D年≥月I)-日学位论文版权使用授权书本学位论文
此文档下载收益归作者所有