CMU's DeformX trains robots to whip ropes in sim — UR5e knocks apple off a head with 0cm error
DJiafei · x · 2026-09-28
CMU researchers unveiled DeformX, a co-simulation framework for deformable linear objects (DLOs), earning an Oral at IROS 2026 and Best Short Paper at a CVPR 2026 workshop.
- It couples a dedicated Cosserat rod engine with NVIDIA Isaac Sim for physically faithful cable bending, twisting and collision, enabling sim-to-real transfer, CAD-grade data generation, and robot learning.
- Ships with the WireSeg-36k synthetic dataset (depth + instance masks); SAM3 + LoRA fine-tuning delivers strong segmentation gains on real held-out tests.
- Highlight demo: a real UR5e whips a rope to knock an apple off a person's head, cutting real-world error to 0 cm vs. a 15.1 cm baseline. Paper, code and dataset are open-sourced.
More from Embodied
- Watch Opus design a 3D asset, print it, and have a robot arm retrieve it — burny_tech · 2026-09-28
- Robot spotted collecting embodied AI data at a night market — xiaosun86 · 2026-09-28
- Figure's Helix 2.5 results make this founder rethink why humanoids look human — yongqianme · 2026-09-28
- Sakana AI unveils SAIL, scaling in-context imitation learning for robots without retraining — SakanaAILabs · 2026-09-28
- Sharpa Robotics wows with dexterous robot hand demo open to public challenge — jonstephens85 · 2026-09-28
- Jev seen as fit for high-level robotics orchestration: fast schema beats LLM lag — ai · 2026-09-28