视觉导引agv的路径规划与避障分析-path planning and obstacle avoidance analysis of vision guided agv.docxVIP

VISUALGUIDEDAGVPATHPLANNINGANDOBSTACLEAVOIDANCEABSTRACTWiththedevelopmentofmodernscienceandtechnology,AutomaticGuidedVehicle(AGV)istobemoreandmorewidelyusedinmanyareas.TheAGVisakindofunmanned-vehicle.Accordingtotheinstructionsofthecontrolsystem,theprogramsdesignedinadvance,andthepositioninformationoftheAGV,whichisdeterminedbythesensorsontheAGV,itcancompleteaseriesoftheworks,drivingalongtheprovidingrouteandstopsinworkstations.ThisarticleindetailsurveystheChineseandEnglishliteratures,summarizesthetechnicalapplicationsoftheAGVsysteminthedomesticandoverseasandcurrentsituationsofthedevelopmentsandresearches,systemicallyanalyzetheAGV’sguidancemethodsandworkingprinciples,characteristic,applicationofindustryprospectsanddevelopmenttrends,andcarriesonthedetailedexplanationtothemachinevisionprincipleandcontents,andapplicationsofthemachinevisionontheAGV.ThepapermainlyfocusesonvisualguidedAGVpathplanningandobstacleavoidance,includingtheglobalpathplanninginthestaticknownenvironment,andthelocalpathplanninginthedynamicunknownenvironment.Forthestaticknownenvironment,geneticalgorithm-basedAGVpathplanning,proposedthemethodforfixedlengthencodingsymbolsofpath.TheobstacleinAGVsenvironmentbeexpressedaspolygon,drawnthenodesoftheAGVpassingwithimprovedroadmap,andarbitrarycodingeachnodewithintegerfrom1ton,codingthepathintoafixed-lengthchromosomestringofnaturalnumbers,thechromosomelengthisn.Algorithmissimpleandeffectiveandhasgoodversatility.Simulationresultsshowthefeasibilityofthealgorithm.AimingatthedynamicunknownenvironmentdefinedtheenvironmentoftheAGVstraveling;proposedthefuzzyneuralnetworkcontrolmethodforthepathplanningofAGVavoidanceobstacle,designedthefuzzyneuralnetworkcontroller.Thecontrollersinputaretheenvironmentalcategory,distancefromAGVtoobstaclesontheleft,frontandrightside,theangle?θoftargetpointandtheAGVscurrentdirection;theoutputarethecontrolparameterαofsteeringdevice,andexpectationsspeedv.AndcarriedoutthesimulationanalysiswithMATLAB,thesimulationresultindicatedthatthefuzzyneuraln