Co-carrying an object by robot in cooperation with humans using visual and force sensing.
Journal:
Philosophical transactions. Series A, Mathematical, physical, and engineering sciences
Published Date:
Aug 16, 2021
Abstract
Human-robot collaboration poses many challenges where humans and robots work inside a shared workspace. Robots collaborating with humans indirectly bring difficulties for accomplishing co-carrying tasks. In our work, we focus on co-carrying an object by robots in cooperation with humans using visual and force sensing. A framework using visual and force sensing is proposed for human-robot co-carrying tasks, enabling robots to cooperate with humans and reduce human efforts. Visual sensing for perceiving human motion is involved in admittance-based force control, and a hybrid controller combining visual servoing with force feedback is proposed which generates refined robot motion. The proposed framework is validated by a co-carrying task in experiments. There exist two phases in experimental processes: in , the human hand holds one side of the box object, and the robot gripper of the Baxter robot automatically approaches to the other side of the box object and finally holds it; in , the human and the Baxter robot co-carry the box object over a distance to different target positions. This article is part of the theme issue 'Towards symbiotic autonomous systems'.