Pose-Invariant Object Classification from Haptic Features in Robotic Grasps

Two cost-efficient methods for pose-invariant object identification using a multi-fingered robotic hand with proprioceptive sensing, without requiring knowledge of the relative pose between the object and hand.