RAI Institute's Koala Gripper pairs two-thumbed capture and robot devices with full parity
lukas_m_ziegler · x · 2026-08-31
Handheld data collection is how many manipulation datasets are built today: a person holds a gripper-shaped device, performs the task by hand, and the recording becomes training data. The catch is that these devices are usually shaped to match an existing robot gripper, so the human works around the robot's morphology — costing both ergonomics and demonstration quality.
The @raiinst team designed the Koala Gripper: two devices with full parity — a capture device a person holds, and a robotic device that executes the learned policy. Anything one can do, the other reproduces, and the shared feature set updates as the design iterates so neither side quietly compromises the other. Its morphology features two independently controlled underactuated fingers ("two thumbs").
More from Embodied
- Peking University's SimForcing distills simulation motion priors into robot world models — PekingUniversity · 2026-10-16
- NVIDIA and Mercedes-Benz Expand Partnership on Next-Gen Automated Driving — nvidia · 2026-10-16
- ~25,000 Humanoid Robots Shipped in H1 2026, Over 95% from Chinese Vendors — ingliguori · 2026-10-16
- openpilot brings stoplight-stopping autonomy to a 2019 Corolla — k7agar · 2026-10-16
- Hironobu Fujiyoshi publishes deck: from AI agents to Agentic Robotics — HirokatuKataoka · 2026-10-16
- CoDA co-evolves simulation and robot policies via shared failure memory, lifting OOD success from 5.3% to 27.6% — PLAN-Lab · 2026-10-16