ReLoco

Unified Humanoid Locomotion and Fall Recovery via Behavior-Conditioned Motion Priors

One proprioceptive actor. Walk, run, recover, and resume commanded motion.

Walk Fall Recover Walk
Explore the project
Lequn Fu1 Yufeng Hu1 Yijun Zhong1 Jie Zhang1 Xiao Li1 Lingxiang Ji1 Zhiyuan Xu2 Jian Tang2 Shiqi Li1,†

1Huazhong University of Science and Technology   2Beijing Innovation Center of Humanoid Robots
†Corresponding author: Shiqi Li

1deployable policy
64.5 kgrobot weight
1.70 mrobot height
zero-shotsim-to-real transfer

Interactive demo

Take the policy for a walk

Command the robot in MuJoCo, apply a disturbance, and let the same actor attempt recovery.

Tips:

1Use 8/2 to change forward speed, and keep commands small to test stable standing.

2Use 4/6 for lateral velocity and 7/9 for yaw commands.

3If the robot loses balance or falls, let the recovery policy try to get back up; press 0 or use Reset only when you want to restart.

Try low-speed commands first, then increase forward speed after the robot is balanced.

How to Play

Controls

8 / 2Forward speed 4 / 6Lateral speed 7 / 9Yaw speed 0Reset robot PApply a push

Abstract

Recover, then continue the task

Humanoid robots must recover from falls and resume commanded motion to sustain locomotion. Learning this behavior in one policy requires references suited to the robot and supervision that accommodates both upright movement and ground-contact recovery. We present ReLoco, a training framework that combines simulation-refined motion references with behavior-conditioned adversarial motion priors. A tracking policy refines retargeted human motions into robot-executed examples. During policy learning, posture selects the expert pool and conditions a shared discriminator with two behavior-specific outputs. Task rewards and shared regularization support command tracking and recovery, while temporary assistance facilitates exploration. Deployment on the 64.5-kg Tiangong 3.0 humanoid demonstrates walking, running, get-up, and locomotion resumption after falls using one proprioceptive actor, without controller switching or on-robot policy fine-tuning.

01

Refine the references

An offline tracking policy executes retargeted human motions in simulation to build locomotion and get-up reference libraries.

02

Condition the supervision

Robot posture selects the expert pool and one of two outputs from a shared discriminator. Task rewards and shared penalties train the same actor.

03

Deploy one actor

The proprioceptive policy uses velocity commands to walk, run, get up, and resume locomotion. The discriminator and supervision router are used only in training.

Method

Better references, posture-appropriate supervision

ReLoco (Recovery and Locomotion) refines retargeted human motions in simulation to form locomotion and get-up reference sets. Robot posture selects the reference set and conditions a shared discriminator with two behavior-specific outputs to train one actor.

ReLoco framework: human motions are retargeted and refined by a tracking policy; posture-conditioned expert sampling, a shared two-output discriminator, and task rewards train one actor; that actor transfers to Tiangong for locomotion and recovery.
Overview of ReLocoTemporary recovery assistance is withdrawn during training. Deployment uses the actor alone, without the reference tracker, discriminator, assistance, or supervision router.

Reference-motion refinement

Raw kinematically retargeted motion (left) and the robot-executed reference produced by a tracking policy in simulation (right). The refined trajectory is used as expert motion data for AMP training.

Raw motion vs. simulation-refined motionLeft: kinematically retargeted motion. Right: robot-executed reference motion.

Real-world demonstrations

The same policy, across the whole cycle

Direct transfer to the 1.70-m, 64.5-kg Tiangong 3.0 humanoid, without on-robot policy fine-tuning or a separate recovery controller.

01

Fall, recover, and resume walking

After a fall, the actor re-establishes support, gets upright, and resumes the commanded locomotion. Standing-kick trials also cover prone, supine, and side-lying recovery.

Recovery after a front-direction kickFall → recover → walk
Recovery after a rear-direction kickFall → recover → walk
02

Disturbance rejection without falling

Representative kicks from the front and rear directions are rejected during walking, so the robot restores balance without a ground-level fall and continues the command.

Representative front-direction kickNo-fall continuation
Representative rear-direction kickNo-fall continuation
03

Locomotion across terrains

The same actor generates standing, walking, and running. The outdoor sequence shows locomotion on brick pavement, stone-tiled pavement, grass, and grassy slopes.

Locomotion on varied surfacesBrick Pavement / Stone-Tiled Pavement / Grass / Downhill Grassy Slope / Uphill Grassy Slope

Citation

BibTeX

@article{fu2026reloco,
  title  = {ReLoco: Unified Humanoid Locomotion and Fall Recovery via Behavior-Conditioned Motion Priors},
  author = {Fu, Lequn and Hu, Yufeng and Zhong, Yijun and Zhang, Jie and Li, Xiao and Ji, Lingxiang and Xu, Zhiyuan and Tang, Jian and Li, Shiqi},
  journal = {arXiv preprint arXiv:2611.23916},
  year   = {2026}
}