Point It, Strike It: Direction-Conditioned Dynamic Manipulation of Deformable Linear Objects

2026-10-07

AI Summary

This study addresses the challenges of jointly controlling 3D position and end-effector orientation, complex dynamics modeling, and significant sim-to-real gaps in the dynamic manipulation of flexible linear objects. To this end, it proposes the TRACE data generation strategy and the RECAP residual calibration method. Technically, a high-fidelity simulation environment, DeformX2.0, is constructed by integrating a GPU-accelerated Cosserat rod solver, an aerodynamic model, conditional flow matching, and cross-entropy-based adaptive trajectory optimization. Experimental results demonstrate that the proposed approach substantially narrows the sim-to-real gap, achieving 92.1% simulation accuracy. In real-world robotic experiments, the success rate for pure position tasks improves to 87%, while that for tasks involving orientation constraints increases markedly from 50% to 79%.

Abstract

Goal-conditioned dynamic manipulation of deformable linear objects has mainly specified goals as positions for a rope tip to reach. Many tasks, however, depend on how the tip arrives. We therefore study single-swing rope striking with goals that specify the tip's 3D position and arrival direction, across the workspace and on different ropes. This is challenging because rope dynamics are hard to model, no demonstrations exist, distinct swings reach the same goal with different reliability, and the sim-to-real gap extends beyond the rope. To address these challenges, we extend the state-of-the-art DLO simulator DeformX with GPU acceleration, a stable Cosserat rod solver, and a cross-flow aerodynamic model, yielding DeformX2.0, which is more than $20{,}000\times$ faster. We then propose TRACE (Trace-rooted Adaptive Cross-Entropy), which generates striking data by warm-starting each new target from the stored swing whose tip path passes closest to it. Its cost penalizes rope bending and abrupt tip motion to favor repeatable swings. A conditional flow-matching policy trained on this data reaches 92.1% accuracy in simulation. Finally, we propose RECAP (Residual Calibration Policy), which fits the simulator's rope and rig parameters to a few calibration swings and adapts actions with a correction policy trained in simulation. On a real robot, across three ropes, RECAP raises success within 5cm from 72% to 87% for position goals, and within 10cm and 10° from 50% to 79% for goals that also specify the arrival direction.

Problem

• deformable linear objects
• dynamic manipulation
• goal-conditioned control
• sim-to-real gap
• rope dynamics

Innovation

• Deformable Linear Objects
• Dynamic Manipulation
• Simulation Acceleration
• Sim-to-Real Transfer
• Conditional Flow Matching