Berkeley's Morphometric Imitation Hits 89.3% Zero-Shot Real-World Success Across 3 Robot Hands
Berkeley · hf · 2026-09-28
Berkeley researchers present Morphometric Imitation, a three-stage pipeline that converts reconstructed human hand-object interactions (HOIs) into zero-shot sim-to-real visuomotor policies.
- Morphometric optimization (MMO) retargets human motion across hand morphologies while preserving demonstrated contacts
- Residual RL refines the kinematic reference using object pose and contact info to yield dynamically feasible robot demonstrations
- Demonstrations are distilled into visuomotor policies
Results: across three robot hands and ten HOIs, MMO improves contact F1 by at least 8 points over the strongest of five baselines and boosts downstream dynamic retargeting success by up to 35 points; policies achieve 89.3% zero-shot success over 300 real-world trials on 30 objects.
More from Embodied
- IMLE-VLA replaces flow matching with single-step cIMLE, 3.67x faster robot actions at 55Hz — petitegeek · 2026-09-28
- Yacine demos general sim2real pipeline: policy trained in just two minutes — yacineMTB · 2026-09-28
- Only 7 Years: From Kicked Robot Dogs to a Robotic Leap — alvelda · 2026-09-28
- InternW0-Δ: A World Action Model Trained on 20K+ Hours of Open Robot Data, Fully Open-Sourced — Xingyu Miao · 2026-09-28
- SolidBot moves real steel: post-processed robot programs now heading into TCP and accuracy tests — MatthewChang · 2026-09-28
- Owner demos Grok agent running natively in real time on a 2022 Tesla HW3 — Scobleizer · 2026-09-28