DefFlow: Goal-Conditioned Deformable Object Manipulation using 3D Point Flows

Key points

• Manipulating deformable objects toward desired 3D configurations is a fundamental capability for robots operating in unstructured physical environments.1
• However, as the range of possible initial and target configurations grows, goal-conditioned deformable manipulation requires a robot to infer increasingly diverse manipulation behaviors in addition to precise continuous actions.1
• We show that this formulation better captures both the manipulation decisions, such as when and where to grasp, release, or re-grasp, and the precise continuous motions required for goal reaching, rather than predicting only the robot flow.1
• Across rope, granular-media, and cloth manipulation settings with diverse initial and target configurations, DefFlow substantially outperforms goal-conditioned action-prediction and robot-flow-only baselines on success rates and goal-distance metrics.1

Sources (1)

• [1]DefFlow: Goal-Conditioned Deformable Object Manipulation using 3D Point Flows
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 16, 02:09 AM

  We introduce DefFlow, a framework for goal-conditioned manipulation of deformable objects toward a broad range of 3D configurations.

  Manipulating deformable objects toward desired 3D configurations is a fundamental capability for robots operating in unstructured physical environments.

Extractive summary: sentences quoted from the sources.