A Unified Approach for Virtual Fixtures and Goal-Driven Variable Admittance Control in Manual Guidance Applications

Abstract

In applications of human-robot physical interaction, it is important to guarantee high accuracy not only in terms of the position of the end-effector, but also in terms of its orientation. While translational movements are easily understandable for a human, rotational ones do not have an immediate interpretation. This does not allow the operator to achieve an adequate positioning accuracy when a target orientation is set. In this work, a suitable strategy that facilitates the human in accomplishing rotational motion is proposed through the formulation of rotational virtual fixtures. This is obtained thanks to a clever description of the orientation that allows to draw an equivalence between the formulation of translational virtual fixtures and their rotational counterparts. To further improve the human comprehension of the best motion direction in the rotational world, a variable admittance control is also designed. We finally propose to jointly exploit these techniques to achieve the best possible performance in terms of positioning accuracy. The performance are evaluated by means of rotational point to point cooperative motions involving several subjects and an ABB IRB-140 robot.