← Back to the feed
Robotics & manufacturingResearch findings

Perception-Driven Humanoid Motion from Simulation

Original published 2026.11.05 03:55 KSTAdded to Blackdeck 2026.11.07 21:52 KST

What has changed?

A pipeline generates vision-language-motion data across thousands of procedurally generated homes to train autonomous movement

Key points

  1. Trains humanoid motion using simulation-based vision-language-motion data
  2. Generates data across thousands of procedurally generated home environments
  3. Achieves 91% success rate in zero-shot transfer to six real-world, unmapped environments

Connect it to your work

Achieves zero-shot transfer to real-world environments without prior exposure to actual homes

What remains unproven

This research note is based on an abstract; check the original for test conditions and application scope.

Originals used for this note

Links open the original publisher in a new tab.

  1. arxivPerceptive Humanoid Motion from Simulation