HomeBody gives VLMs persistent spatial memory: Unitree G1 cleans a kitchen with no task-specific training
rsasaki0109 · x · 2026-10-12
HomeBody equips frontier VLMs (e.g. GPT Astra) with persistent spatial memory and composable humanoid skills for long-horizon tasks. In a previously unseen kitchen, a Unitree G1 guided by GPT Astra cleaned up across the room and retrieved a remembered object from an underspecified request — no environment-specific training data or extra policy learning.
Pipeline (per diagram): exploration driven by Astra collects scene data via iPhone video, SLAM+depth and joint states; a Real2Sim Agent reconstructs the scene in IsaacLab/MuJoCo/Blender. Then the robot executes "clean up the kitchen": pick and stow the coffee bag, navigate from memory to the milk carton, pick it, and drop it in the bin.
Related event: Stanford's HomeBody Gives VLM-Powered Humanoids Spatial Memory(2 posts)→
More from Embodied
- NVIDIA's Seattle Robotics Lab shows robots assembling GB300 tester trays — TinfoilTricorn · 2026-10-12
- Dev builds Claude-coded 'daily wearable' firmware packing 27 games and 16 utilities — ezshine · 2026-10-12
- China's driverless buses now ship without steering wheels — CurieuxExplorer · 2026-10-12
- ENCORE: Coding Agents Write Robot Policies From 5 Demos, Hitting 96.3% on LIBERO-PRO — 2077AIDataFoundation · 2026-10-12
- ARTISAN: Multi-Agent Pipeline Generates Physically Grounded Articulated 3D Assets with New Benchmark — elliottszwu · 2026-10-12
- $84,000 Boston Dynamics robot washes a lioness, sparks animal-intervention debate — CurieuxExplorer · 2026-10-12