Planning for Dexterous Ungrasping: Secure Ungrasping through Dexterous Manipulation

TL;DR

A planning framework for determining a feasible minimum-cost motion path that completes dexterous ungrasping is presented and Digit asymmetry in a gripper, i.e. difference in digit lengths, is discovered as the key to feasible and secure un grasping.

Abstract

This paper presents a robotic manipulation technique for dexterous ungrasping. It refers to the capability of securely transferring a grasped object from the gripper to the robot's environment, i.e. the inverse of grasping or picking, through dexterous manipulation. The game of Go offers an example: consider how the player would typically place an initially pinch-grasped stone onto the board through the dexterous interaction between the fingers, the stone, and the board. Likewise, dexterous ungrasping addresses the necessity of changing the object's configuration relative to the gripper or the environment in order to securely keep hold of the object. In particular, we present a planning framework for determining a feasible minimum-cost motion path that completes dexterous ungrasping. Digit asymmetry in a gripper, i.e. difference in digit lengths, is discovered as the key to feasible and secure ungrasping. A set of experiments show the effectiveness of dexterous ungrasping in practical placement tasks.

Figures (10)

• Figure 1: (a) Go stone placement by human. (b) How can a robot place a Go stone stably?
• Figure 2: (a) Shallow-depth insertion using a conventional parallel-finger gripper. (b) Releasing the linear object from the gripper: the contacts $G$ and $A$ remain fixed while $B$ is sliding and $\psi$ is increased. (c) Bringing the object to the hole: $G$, $A$, and $B$ remain fixed while $\theta$ is decreased.
• Figure 3: (a) Complete insertion path (red arrow) contained in the set of force-closure grasps shaded gray (both light and dark), computed with the unit contact wrenches shown in (b), denoted $\mathbf{w}_{Xi}$ at the contact $X$. $\mu_X$ denotes the friction coefficient at $X$. The colored contours delineate the cross-sections of the set on the $\theta \psi$-planes. (c) Grasp with no force-closure, marked in the lightly shaded volume in (a), due to the flatness at $G$.
• Figure 4: (a) Generalized model of planar ungrasping with a semi-elliptical object and a gripper whose fingertip (currently at $A$) can be placed freely relative to the thumb, as implied by the silhouette of a multi-jointed finger. The arrows are the contact normals. (b) 3D view of the ungrasping setting.
• Figure 5: Motion primitive to decrease $d_A$. Both $A$ and $B$ slide while $G$ remains fixed.