Slip classification for dynamic tactile array sensors

摘要: The manipulation of objects held in a robotic hand or gripper is accompanied by events such as making and breaking contact slippage, between the fingertips grasped object external surfaces. Humans can distinguish among events, part, because they excite various mechanoreceptors hands differently. As part an effort to provide robots with similar capability, we propose two features that be extracted from dynamic tactile array data used discriminate hand/object object/world slips. Both rely on examining how slippage affects sensors compared way it individual elements array. In comparison approaches require extensive training particular combinations skin, work for wide range frequencies grasp conditions. performance generalizability are verified testing three different kinds textures, forces slip demonstrate greater than 85% accuracy identifying location slip.