Object-Centric Kinodynamic Planning for Nonprehensile Robot Rearrangement Manipulation

2024-09-30

arXiv.org

AI Summary

Existing non-grasping robotic rearrangement methods adopt a robot-centric paradigm, relying on post-hoc evaluation of outcomes—resulting in low efficiency and misalignment with human intuition. This paper introduces the first object-centric planning paradigm for multi-object non-grasping rearrangement, proposing a general-purpose object-centric planner. Our key contributions are: (1) dynamics modeling grounded in object motion intent, coupled with online closed-loop pushing control; (2) a kinodynamic path-planning–informed mapping from object trajectories to robot actions; and (3) a standardized benchmark suite and evaluation protocol. Extensive simulation and real-robot experiments demonstrate substantial improvements in task success rate and action intuitiveness, outperforming state-of-the-art approaches. The framework establishes a reproducible, scalable, and standardized foundation for research in non-grasping manipulation.

Technology Category

• Intelligent Robots: Manipulation
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: The sharing economy

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.

Problem

• Addressing inefficiency in robot-centric nonprehensile rearrangement planning
• Proposing object-centric planning for intuitive task-effective robot actions
• Developing benchmarking protocol for nonprehensile rearrangement research

Innovation

• Object-centric planning for nonprehensile rearrangement
• Closed-loop pushing strategy for robot actions
• Benchmarking protocol for future research standardization