欢迎来到天天文库
浏览记录
ID:34891876
大小:449.56 KB
页数:9页
时间:2019-03-13
《Robot_Motion_Planning.pdf》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、RobotMotionPlanningGaneshSwaminathanSchoolofEngineeringScienceSimonFraserUniversity,Canadagswamina@sfu.caFebruary27,20061IntroductionMotionplanningisafundamentalprobleminrobotics.Theclassicpathplanningproblemisdescribedasthefollowing:givenathreedimensionalrigidbodyandaknownsetofobstacle
2、s,thetaskisto¯ndacollision-freepathfromastartcon¯gurationtoagoalcon¯guration.Additionally,thistaskistobecompletedinareasonableamountoftime.Thisisknownasthepianomover'sproblem[1].Amoregeneralscenario,knownasthegeneralizedmover'sproblemwouldhavetherobotas°exiblepolyhedrawithmovingpolyhedr
3、alobstacles.Thisreportwillreviewclassicalmotionplanningparadigms-potential¯eldmethods,roadmapmethodssuchasvisibilitygraphmethodsandprobabilisticroadmaps.Moreemphasiswillbeplacedonroadmapsasthistechniquehasshownpromisingresultsforcomplicatedrobotplanningproblems.Actionsinthephysicalworld
4、aresubjecttophysicallaws,uncertainityandgeometricconstraints.Thedesignandanalysisofmotionplanningalgorithmsraisesquestionsinanumberof¯elds-mechanics,controltheory,computationalanddi®erentialgeometry,andcomputerscience.Thisreportwillonlyconsiderthesimpli¯edgeometricproblemandassuchignore
5、issuessuchasincompleteinformation,nonholonomicconstraints.movinganddeformableobjects,multiplerobotsandcontrolrelatedissues.1.1Mathematicalde¯nitionsBeforewecangoonwiththereviewonmotionplanning,itisinstructionaltoestablishauniformmathematicnotation.Thenotationfollowedinthisreportisthesam
6、easin[2].Thecon¯gurationspaceQisde¯nedasthesetofallcon¯gurationsarobotcanachieve.Thisspaceisgenerallynon-Euclidean.Thedimensionofthisspaceisequaltothenumberofdegreesoffreedomintherobot.Forexample,arigidbodyinthreedimensionshassixdegreesoffreedom-threefortheposition(x,y,z)andthreefortheo
7、rientation(roll-yaw-pitch).Qfreeisde¯nedasthesetofrobotcon¯gurationsthatdonotinteresectanobstacle.Aparticularcon¯gurationfromthisspaceisde¯nedasq.2PotentialFieldMethodPotential¯eldmethodswere¯rstdescribedbyKhatib[3]foron-linecollisionavoidanceforarobotwithproximitysensors.Thisa
此文档下载收益归作者所有
点击更多查看相关文章~~