IROS 2024 paper lifts robot grasp adaptation success by 28.5%
Ed__Johns · x · 2026-07-29
A robotics paper on adapting manipulation skills to novel grasps without object CAD models or camera calibration.
Key points
- The method learns from a short self-supervised data collection phase where a camera watches the robot move an object rigidly grasped in its end effector.
- It works with RGB, depth, or both, and does not require prior object knowledge.
- In 1,360 real-world evaluations, self-supervised RGB data outperformed depth-based alternatives, including several state-of-the-art pose estimation methods.
- The best-performing baseline was beaten by an average of 28.5% higher success rate on everyday manipulation tasks.
The post also notes that the work was presented at IROS 2024 and links to the paper and video explanation.
More from Embodied
- UMD's PAIR framework bridges perception and action in VLA models via bridge tokens — Kaixi Feng · 2026-10-16
- CI-MSE offline metric hits -0.87 Spearman correlation with real robot task success vs weak raw MSE — chris_j_paxton · 2026-10-16
- List of 15 best robotics researchers on X, from Jim Fan to Chelsea Finn and Fei-Fei Li — FinanceYF5 · 2026-10-16
- Fake Apple listing for dual-screen 'iPhone Duo' jokes it's the best phone for watching your AI agents — AIFlow_ML · 2026-10-16
- Inside Auki's Hong Kong office: a fake grocery store powering robotics data collection — LoganGrasby · 2026-10-16
- Physicist calls compressed-air car a scam: one charge would take a week of compressor time — rickasaurus · 2026-10-16