A selective attention-based contextual perception approach for a humanoid robot-一种针对类人机器人的选择性注意力情境感知方法.pdfVIP

2007 JournalefControlTheory 5(3)244—252 DOI andApplications attention-basedcontextual A，selective for ahumanoidrobot approach XIAO YanrongJIANG，Nanfeng Science (SchoolofComputer 510640 andEngineering，soutllChinaUniversity ofTechnology,GuangzhouGuangdongChina) Abstract：Ahumanoidrobotis hlooded sensedinformationwhen the always by environment，andit sensing usually needs timeto and the sensedinformationInthis selective aignihlcantcomputeprocess par)er,a attention—basedcontextual perceptionwas forhumanoidrobotstosensetheenvironment the approachproposed wi山high connotation efficiency．First ofattentionwindowrAWlis andabstractdefinition extendedtomakeamoregeneral andstate咖nsformationsarealsodiscussed Second，theattentioncontrol aredescrlbed．whichintension— policies integrate selectionanddistractor guidedperceptual inhibition，andcandealwith issuesDistractorinhibitionis objects emergent usedtofilterunrelated areview