4 Evolutionary Articial Potential Fields and Their Application in Real Time Robot Path Planning

《4 Evolutionary Articial Potential Fields and Their Application in Real Time Robot Path Planning》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、EvolutionaryArtificialPotentialFieldsandTheirApplicationinRealTimeRobotPathPlanningPrahladVadakkepat,KayChenTanandWangMing-LiangDepartmentofElectricalEngineeringTheNationalUniversityofSingapore10KentRidgeCrescent,Singapore119260felepv,eletankc,engp9577g@nus.edu.sgAbstract-Anewmethodologynam

2、edEvolutionaryAr-tionontheenvironmentisambiguous.TheartificialpotentialtificialPotentialField(EAPF)isproposedforreal-timefieldapproachesaremuchconvenientthanAImethodsforrobotpathplanning.Theartificialpotentialfieldmethodtheirhighefficiencyinpathplanningprovidedthatthework-iscombinedwithgenetical

3、gorithms,toderiveoptimalpo-ingenvironmentisknown.tentialfieldfunctions.TheproposedEvolutionaryAr-Artificialpotentialfieldsareinuse,inthecontextofob-tificialPotentialFieldapproachiscapableofnavigat-stacleavoidance,sinceKhatib[18].Inthisapproachamo-ingrobot(s)situatedamongmovingobstacles.Potenti

4、albilerobotappliesaforcegeneratedbytheartificialpotentialfieldfunctionsforobstaclesandgoalpointsarealsode-fieldascontrolinputtoitsdrivingsystem.Thetraditionalfined.Thepotentialfieldfunctionsforobstaclescontainartificialpotentialfieldapproacheswereoftenpurelyreac-tunableparameters.Multi-objectivee

5、volutionaryalgo-tiveinnatureanddonotoptimizethepatharrivedat[19,20].rithm(MOEA)isutilizedtoidentifytheoptimalpoten-Modifiedpotentialfieldapproacheswithrobustandimprovedtialfieldfunctions.Fitnessfunctionslike,goal-factor,performancehavebeenreportedlately:ChanclouB.[6]sug-obstacle-factor,smooth

6、ness-factorandminimum-path-gestedpathplannerstoimpartrobustfeaturestotheartificiallength-factoraredevelopedfortheMOEAselectioncri-potentialfieldfunctions.Thisapproachreliesontwopath-teria.Analgorithmnamedescape-forceisintroducedplanners,oneforglobalplanningthroughphysicalcompu-toavoidtheloca

7、lminimaassociatedwithEAPF.Movingtations,andtheotherforlocalplanningthoughaphysicalobstaclesandmovinggoalpositionswereconsideredtosimulationofthevehiclewithintheenvironment.Thesetwotesttherobustperformanceoftheproposedmethodology.plannersworktogethertoarriveata