UCLA's ChairNav teaches autonomous wheelchairs to learn each rider's preferences
zhoubolei · x · 2026-10-31
Zhou Bolei's team and UCLA researchers introduced ChairNav, an autonomous wheelchair navigation system built on vision-based navigation foundation models for long-horizon sidewalk navigation. Training runs in three stages: pre-training adapts camera and motion data from delivery robots; mid-training learns wheelchair navigation from human driving demonstrations; post-training converts rider interventions into preference feedback that improves driving behavior. It handles obstacles, narrow passages, and outdoor routes, embodying the idea that assistive robotics must adapt to the person using it.
Related event: UCLA's ChairNav Brings Long-Horizon Autonomous Navigation to Wheelchairs(3 posts)→
More from Embodied
- Real2Sim2Real is hard: even plugging in a network cable demands high sim fidelity — jonstephens85 · 2026-10-31
- Genspark unveils VoicePad, a $99 five-key AI voice gadget now open for pre-orders — Scobleizer · 2026-10-31
- World's first humanoid robot factory for clothing launches, mass-producing Annie robots — CyberRobooo · 2026-10-31
- Robotics benchmarks feel impossible; Drift builds evals for real deployment sites — jonstephens85 · 2026-10-31
- Tesla's We, Robot promises materialize: Cybercab in production, Robotaxi in six cities — XFreeze · 2026-10-31
- SO-101, AI's 'most famous and infamous' robot, gets official support — jonstephens85 · 2026-10-31