ECHO-N

ECHO’s context-adaptive policy trained with necessary domain randomization. It infers dynamics from interaction history and adjusts ECHO-PRE’s motion commands. The paper reports four G1 tracking tests without added payload.

Papers & reports (1)

ECHO: In-Context Behavior foundation model for Humanoid Whole-Body Control

Paper available

arXiv: 2611.25778DOI: 10.48550/arxiv.2611.25778
Authors & affiliations (10)

Author order: Original order

  1. Xingyi Wangco-first author

    Zhejiang University · Institute of Artificial Intelligence (TeleAI), China Telecom

  2. Bin Lico-first author

    Institute of Artificial Intelligence (TeleAI), China Telecom · ShanghaiTech University

  3. Jiyuan Shi

    Institute of Artificial Intelligence (TeleAI), China Telecom

  4. Chenyun Zhang

    Institute of Artificial Intelligence (TeleAI), China Telecom

  5. Weiji Xie

    Institute of Artificial Intelligence (TeleAI), China Telecom · Shanghai Jiao Tong University

  6. Xiangyu Miao

    Institute of Artificial Intelligence (TeleAI), China Telecom · Shanghai Jiao Tong University

  7. Junsong Wu

    Institute of Artificial Intelligence (TeleAI), China Telecom · Shanghai Jiao Tong University

  8. Wei Song

    Zhejiang University

  9. Shiqiang Zhucorresponding author

    Zhejiang University

First published
2026-11-20
Latest paper revision
v1 · 2026-11-20
Venue
Not provided

Covered model versions

Paper ↗PDF ↗

How to access and use

Training code for this version is unconfirmed. Public code resources do not establish access to the exact training implementation.

Training / fine-tuning code

Release information

SourceVerified on · 2026-11-20

Checks before deployment

Available weights or code do not establish plug-and-play robot compatibility. Check the required configuration, data and interfaces in the source documentation.

Control and inputs / outputs

Model type
Motion & whole-body control
Backbone
ECHO-PRE
Input
Robot state

Scope and training

Embodiment scope

Single embodiment

Functional purpose

action_control

Output

Actions

Robot form

Humanoid

Task scope

Multiple tasks

Training

Reinforcement learning

Submit a source or correction