Object-Centric Kinodynamic Planning for Nonprehensile Robot Rearrangement Manipulation

Kejia Ren,Gaotian Wang,Andrew S. Morgan,Lydia E. Kavraki,Kaiyu Hang

2024-10-01

Abstract:Nonprehensile actions such as pushing are crucial for addressing multi-object rearrangement problems. To date, existing nonprehensile solutions are all robot-centric, i.e., the manipulation actions are generated with robot-relevant intent and their outcomes are passively evaluated afterwards. Such pipelines are very different from human strategies and are typically inefficient. To this end, this work proposes a novel object-centric planning paradigm and develops the first object-centric planner for general nonprehensile rearrangement problems. By assuming that each object can actively move without being driven by robot interactions, the object-centric planner focuses on planning desired object motions, which are realized via robot actions generated online via a closed-loop pushing strategy. Through extensive experiments and in comparison with state-of-the-art baselines in both simulation and on a physical robot, we show that our object-centric paradigm can generate more intuitive and task-effective robot actions with significantly improved efficiency. In addition, we propose a benchmarking protocol to standardize and facilitate future research in nonprehensile rearrangement.

Robotics

What problem does this paper attempt to address?

The problem this paper attempts to address is the inefficiency and ineffectiveness of non-grasping operations in multi-object rearrangement tasks. Specifically, existing non-grasping solutions (such as pushing objects) are robot-centric, meaning that the actions are generated based on robot-related intentions, and their outcomes are passively evaluated after the operations. This strategy differs from human strategies and is generally less efficient.

To improve efficiency and effectiveness, this paper proposes a new object-centric planning paradigm and develops the first object-centric planner for general non-grasping rearrangement problems. By assuming that each object can move actively without relying on robot interaction, this planner focuses on planning the desired object trajectories, which are realized through real-time generated closed-loop pushing strategies. Through extensive experiments and comparisons with existing methods, this paper demonstrates that the object-centric planning paradigm can generate more intuitive and effective robot actions, significantly improving efficiency. Additionally, this paper proposes a benchmarking protocol to standardize and promote future research in non-grasping rearrangement.