Skip navigation
Excited to introduce COLA: Learning Human–Humanoid Coordination for Collaborative Object Carrying 🤝🤖
0:00 / 3:13

Excited to introduce COLA: Learning Human–Humanoid Coordination for Collaborative Object Carrying 🤝🤖

Siyuan Huang

@siyuanhuang955.3K followers

Visit on X
158 likesDownloadOriginal post

22,105 views

Excited to introduce COLA: Learning Human–Humanoid Coordination for Collaborative Object Carrying 🤝🤖 COLA makes humanoids truly helpful in human collaboration — capable of carrying objects, pushing carts, or responding to human push commands. It provides a proprioception-only policy for compliant human–humanoid coordination across diverse movement patterns. The core idea is simple yet effective: 👉 Fine-tune a collaborative policy from a locomotion policy using a residual teacher. 👉 Train in simulation and distill to a real-world student policy for deployment. Paper: Project:

7 Comments

Comments from the original post on X (Twitter)

  • Niraj Pudasaini's profile picture

    @nirajpudasaini

    Great work! Really close to our work :

    3 likes

  • Winston Gao's profile picture

    @WinstonGu_

    Nice work!

    1 like

  • ζ Pedram ζ's profile picture

    @zenstyle

    Love this

  • bob35ooops's profile picture

    @bob35ooops55717

    Amazing! I’m curious—how much data did you use for training, and how was it collected? Do all tasks share a single policy?

  • SoloTech's profile picture

    @Solo__Tech

    COLA is a game-changer for human–humanoid collaboration

  • Tirth Gada's profile picture

    @tirth_gada

    Humanoid clerks soon

  • Sammy's profile picture

    @1samdavid

    This is super helpful!