基于动态窗口法的移动机器人动态避障研究.pdf

ResearchonObstacleAvoidanceofMobileRobot

BasedonDynamicWindowPrediction

Abstract

Withtherapiddevelopmentofroboticstechnology,mobilerobotsarebeing

increasinglyusedinvariousservicesectors,suchaslogisticsdelivery,cateringservices,

andepidemicpreventionwork.Inthetaskofitemdelivery,pathplanningisobviouslya

keytechnologyfortheautonomousnavigationofmobilerobots.Staticenvironmentpath

planningisnolongersufficienttomeetthepracticalrequirements,anddynamicobstacle

avoidancehasbecomearesearchhotspot.Dynamicenvironmentsaremorecomplex,

involvingbothstaticanddynamicobstacles,requiringaccurateobstacleavoidance

methodstoensuresafety.Traditionaldynamicobstacleavoidancemethodsstillhave

certainlimitations.Therefore,thispaperproposedadynamicobstacleavoidancemethod

formobilerobotsbasedonthedynamicwindowapproach.Themainresearchcontentis

asfollows:

First,addressingtheissuesofthedynamicwindowapproachlackingglobalmap

guidance,gettingtrappedinlocaloptimalsolutions,andlocalpathsbeingsuboptimal,

thispaperproposedusingaglobalpathplanningmethodtoprovideaglobalmapforthe

dynamicwindowapproach.Afteranalyzingandcomparingseveralglobalpathplanning

methods,theJumpPointSearch(JPS)algorithmisselected.Toaddresstheproblemof

multiplepathturningpointsandredundantpathsintheJPSalgorithm,animprovement

methodisproposedtooptimizetheplannedpathbyremovingredundantjumppoints.

Second,addressingtheproblemoffixedevaluationfunctionweightsinthedynamic

windowapproachleadingtopoorenvironmentaladaptability,amethodfordynamically

adjustingtheweightsoftheevaluationfunctionisproposed.Thismethoddynamically

adjuststheweightsbyco