HKUST's HALO collects whole-body mobile manipulation demos up to 2.9x faster than Mobile ALOHA
Robbyant-Research · hf · 2026-11-16
HKUST's Xu Lab and Robbyant present HALO, a human-aligned data collection system for whole-body mobile manipulation.
- Problem: learning whole-body mobile manipulation needs demos coordinating bimanual actions with locomotion across large workspaces, but existing interfaces trade off direct robot control, perception-action alignment, and natural operator mobility.
- Approach: align the operator's perception and action space with the robot—kinesthetically guiding end-effectors while pushing a low-resistance platform (without the deployed mobile base) for bimanual manipulation, preserving direct physical feedback and natural locomotion.
- Sim-to-real gaps: deployment-aligned demonstration processing calibrates platform motion against the deployed base response and masks/inpaints demonstrator hands in recorded observations.
- Results: up to 2.9x faster demo collection vs Mobile ALOHA and VR teleoperation with comparable or higher success rates; trained policies handle precise tasks like zip-tie fastening and hit 73.3% success on a six-stage long-horizon mobile manipulation task.
More from Embodied
- Impulse-bought robot arm at new Austin Micro Center to automate a 3D print farm overnight — Sentdex · 2026-11-16
- Ropedia Xperience 2 releases 1,589 3D scans and 4,230 egocentric recordings for physical AI scaling — liuziwei7 · 2026-11-16
- Joe Rogan praises Tesla FSD one-button drive home, urges Tesla to license it — elonmusk · 2026-11-16
- Figure.AI humanoid robot filmed 'leaving the company' in viral clip — Distinct-Question-16 · 2026-11-16
- A beginner's guide to training VLAs: microVLA dev shares low-compute training intuitions — joemeno · 2026-11-16
- Most VLAs lose their base VLM's vision-language skills during robot pre-training, eval shows — DJiafei · 2026-11-16