基于激光数据的移动机器人地图创建问题研究

《基于激光数据的移动机器人地图创建问题研究》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、国防科学技术大学硕士学位论文基于激光数据的移动机器人地图创建问题研究姓名：蔡向东申请学位级别：硕士专业：控制科学与工程指导教师：张辉2010-11೑䰆⾥ᄺᡔᴃ໻ᄺⷨお⫳䰶⸩຿ᄺԡ䆎᭛ᨬ㽕ᴀ᭛Џ㽕ⷨおњ೼ᅸݙ⦃๗ϟ⿏ࡼᴎ఼Ҏ෎Ѣ▔ܝᠿᦣ᭄᥂ⱘഄ೒߯ᓎ䯂乬DŽ⿏ࡼᴎ఼Ҏ߽⫼▔ܝ⌟䎱Ҿᇍᅸݙ⦃๗䖯㸠ᠿᦣˈ㦋প▔ܝᠿᦣ᭄᥂⚍䲚DŽഄ೒߯ᓎ䖛⿟ߚЎᇍ▔ܝᠿᦣ᭄᥂⚍䲚䖯㸠乘໘⧚੠▔ܝ᭄᥂⚍䲚ऍ䜡ϸ໻ℹ偸ˈ䗮䖛ᕾ⦃ᠻ㸠ˈᅠ៤ഄ೒ⱘ߯ᓎDŽ佪ܜˈ䩜ᇍ▔ܝ᭄᥂⚍䲚乘໘⧚䖛⿟ЁᇍѢ㊒ᑺ੠ᅲᯊᗻⱘ㽕∖ˈᴀ᭛䞛⫼෎Ѣ᭄᥂⚍䯈ⱘ⃻ᓣ䎱⾏䖯㸠ऎඳߚࡆˈ䞛⫼ࡼᗕ䯜ؐⱘLine-Tra

2、ckingⳈ㒓ᦤপㅫ⊩䖯㸠Ⳉ㒓ᦤপˈޣᇣњాໄᇍⳈ㒓ᦤপⱘᕅડˈ䙓ܡњ೎ᅮ䯜ؐ䗝পϡᔧ㗠䗴៤ⱘⳈ㒓ᦤপ㊒ᑺ䕗Ꮒ੠Ⳉ㒓ᦤপϡᅠᭈⱘᚙމDŽ݊⃵ˈ䩜ᇍ䗁ҷㅫ⊩Ёᄬ೼ⱘ䅵ㅫ䞣䖛໻ˈऍ䜡ᬜ⥛Ԣϟⱘ䯂乬ˈᴀ᭛ᇚ▔ܝ᭄᥂⚍䲚ⱘऍ䜡ߚЎ㉫ऍ䜡ǃ㊒ऍ䜡ϸϾ䰊↉DŽ㉫ऍ䜡䰊↉੠㊒ऍ䜡䰊↉ߚ߿䞛⫼⚍-⚍䴲䗁ҷऍ䜡ㅫ⊩੠ICP㊒ऍ䜡ㅫ⊩ˈᑊϨᇍⳌᑨⱘㅫ⊩䖯㸠њᬍ䖯DŽ㒧ড়ϸ⾡ㅫ⊩ᅠ៤ᇍ▔ܝ᭄᥂⚍䲚ⱘऍ䜡ˈᦤ催њㅫ⊩ⱘৃ᥻ᗻ੠ऍ䜡䗳ᑺDŽ㉫ऍ䜡䰊↉ˈ䞛⫼ᬍ䖯ⱘ䴲䗁ҷ⚍-⚍ऍ䜡ㅫ⊩ˈ߽⫼ᦤপߎⱘ㒓↉ッ⚍԰Ўᇍᑨ⚍ᇍ䖯㸠ऍ䜡ˈ೼䆕ऍ䜡㊒ᑺⱘࠡᦤϟˈ᳝ᬜഄᦤ催њऍ䜡ⱘ䗳ᑺDŽ㊒ऍ䜡䰊↉ˈ

3、䞛⫼ᬍ䖯ⱘICPㅫ⊩DŽ佪ܜˈḍ᥂㉫ऍ䜡∖ߎⱘবᤶখ᭄∖পϸᐙ᭄᥂⚍䲚ⱘ݀݅ऎඳˈޣᇥњᇍᑨ⚍᧰㋶ⱘⳆⳂᗻˈᦤ催њᇍᑨ⚍ⱘ᧰㋶䗳ᑺDŽ݊⃵ˈᇍICPㅫ⊩䖯㸠њᬍ䖯ˈܜᇍᦤপߎⱘⳈ㒓↉䖯㸠ऍ䜡ˈݡ೼ᇍᑨ㒓↉ᇍЁ᧰㋶ᇍᑨ⚍ᇍˈᦤ催њऍ䜡䗳ᑺˈ䗁ҷ∖পবᤶখ᭄䖛⿟Ё䞛⫼ߚℹ䗁ҷⱘᮍ⊩ˈՓ䗁ҷ䖛⿟݋᳝䕗催ⱘৃ᥻ᗻ੠剕ẦᗻDŽ᳔ৢˈҟ㒡њNuBotᬥᧈᴎ఼ҎⱘᴎẄ㒧ᵘҹঞ᥻ࠊ㋏㒳ˈḍ᥂ᦤߎⱘ෎Ѣ▔ܝᠿᦣ᭄᥂ⱘഄ೒߯ᓎᮍ⊩ˈ߽⫼ᴀᅲ偠ᅸᓔথⱘ⿏ࡼᴎ఼Ҏ೼ᅸݙᷛޚᬥᧈ⦃๗Ё䖯㸠ഄ೒ⱘ߯ᓎˈᅲ偠㒧ᵰ㸼ᯢˈᴀ᭛ᦤߎⱘㅫ⊩㛑໳᳝ᬜഄᅠ៤⦃๗ЁѠ㓈ഄ೒ⱘ߯ᓎDŽ݇䬂䆡˖▔ܝ⌟䎱

4、Ҿ˗Line-Tracking˗޴ԩ⡍ᕕ˗ഄ೒߯ᓎ˗ߚℹ䗁ҷ㄀i义೑䰆⾥ᄺᡔᴃ໻ᄺⷨお⫳䰶⸩຿ᄺԡ䆎᭛ABSTRACTTheproblemofthemapbuildingbasedonlaserscanningdataismainlyresearchedinthispaper.Byusinglaserrangefindertoscantheindoorenvironment,mobilerobotobtainthelaserscanningdata,wedividedmapbuildingintotwoparts:thepretreatmentof

5、thelaserscanningdatapointsetandtheregistrationofthelaserscanningdatapointset.First,theprocessofthepretreatmentofhelaserscanningdata,fortherequirementofprecisionandrealtime.WeusetheEuclideandistancebetweendatapointsforthesegmentation,useLine-Trackingwithdynamicthresholdforlinee

6、xtraction,whicheffectivelyreducestheimpactofthenoiseonthelineextractionandavoidcausingpooraccuracyoflineextractionandpoorintegrityofthelinefromtheimproperselectionoffixedthreshold.Second,fortheproblemofexcessivecalculationandthelowefficiencyofmatching.Addresstheseissues,themat

7、chingstepisdividedtwoparts:preparativematchingstepandprecisematchingstep.Inthepreparativematchingstep,weusethepoint-pointiterativematchingalgorithm.Inprecisematchingstep,weusetheICPiterativematchingalgorithm,andmakecorrespondingimprovement.Inthepreparativematchingstep,weusethe

8、improvedpoint-pointiterativematchingalgorithmtoimprovethespee