Microduck robot runs onboard SLAM with just 64 pixels of depth, 4.8 cm median error
chris_j_paxton · x · 2026-10-15
Antoine Pirrone demonstrated Microduck's fully onboard SLAM using only leg odometry and an 8×8 ToF sensor in its head (64 pixels of depth).
- The duck relocalizes after being carried across a room; ArUco tags serve only as ground truth tracked by a fixed camera
- Median position error is 4.8 cm, and 94% of confident localizations fall within 15 cm
- It relocalized after a 1.5 m carry (2 cm off), a seated carry (11 cm) and a fall (3 cm), never locking onto the wrong place across benchmark sessions
More from Embodied
- AlphaNIST: one RL policy solves all 20 NIST assembly tasks with 98.2% success in simulation — hbouammar · 2026-10-15
- Neural field planning for hybrid manufacturing cuts estimated build time 68.3% — jiqizhixin · 2026-10-15
- Work photos are AI's first mile: if capture is wrong, no model can recover — arnabdotorg · 2026-10-15
- LEGO builds an omnidirectional vehicle for automated warehouses — glenbeer · 2026-10-15
- SlipSense: multimodal tactile slip detection hits 96.7% F1 at 23.1ms with zero-shot robot transfer — chris_j_paxton · 2026-10-15
- BlackBerry launches QNX Launchpad to court early-stage Physical AI startups — pdamodaran · 2026-10-15