hooneyskywalker/g1-motion-tracking-policies

G1 motion tracking policies

Whole-body control policies that make a Unitree G1 humanoid follow human motion capture clips in simulation. One policy per motion.

The pipeline behind them: human mocap → retargeting to the G1 → reference motion → policy training. Stages 1 to 3 have no physics; physics enters at stage 4, where the robot has to hold itself up.

What is here

policies/<sequence>/ holds one trained policy each.

• model_29999.pt — RSL-RL checkpoint at (30,000) iterations
• policy.onnx — the actor network exported for inference, 160 in and 29 out
• run_name.txt — the training run folder the checkpoint came from

eval/<sequence>.json holds the measured numbers for that policy.

How to run a policy

policy.onnx is the actor network on its own. It takes one observation vector and returns one action. It holds no reference motion, so the caller supplies the reference terms.

The observation is 160 values concatenated in this order.

index	term	dim
0:29	reference joint position at the current motion frame	29
29:58	reference joint velocity at the current motion frame	29
58:61	motion anchor position in the robot anchor frame	3
61:67	motion anchor orientation, first two columns of the rotation matrix	6
67:70	base linear velocity in the base frame	3
70:73	base angular velocity in the base frame	3
73:102	joint position relative to the default pose	29
102:131	joint velocity relative to the default pose	29
131:160	previous action	29

The action is an offset on the joint position target: target = default joint position + action, for all (29) joints. Control runs at (50) Hz (sim step (0.005) s, decimation (4)), the same rate as the reference motion.

import numpy as np, onnxruntime as ort

sess = ort.InferenceSession("policies/walk1_subject1/policy.onnx")
obs = np.zeros((1, 160), dtype=np.float32)      # fill as in the table above
action = sess.run(["actions"], {"obs": obs})[0]  # (1, 29)

Joint order is the articulation order of the G1 asset used by BeyondMimic. Read it from the environment rather than assuming one.

How they were trained

BeyondMimic on Isaac Sim (5.1) and Isaac Lab (2.3.2), PPO, (4,096) environments, (30,000) iterations, one policy per motion. The reference motions come from LAFAN1 retargeted to the G1 with GMR.

Reward is tracking accuracy against the reference, not gait plausibility.

Results

Measured over (100) rollouts with domain randomization off. Completion rate is the share of rollouts that reach the end of the clip without the anchor body crossing a height or orientation threshold. The error column is the mean global body position error over all rollouts.

sequence	completion	mean alive	E_g-mpbpe (mm)
aiming1_subject1	100%	100%	89
dance2_subject3	100%	100%	103
obstacles2_subject1	0%	18%	368
walk1_subject1	99%	99%	83
walk2_subject1	100%	100%	114
walk2_subject4	99%	100%	90
walk3_subject2	100%	100%	79
walk3_subject4	0%	88%	116
walk4_subject1	-	-	-

obstacles2_subject1 fails on purpose to be informative: that clip has the actor climbing stairs, with the pelvis above (1.05 m) for (48.7) seconds. The training ground is flat, so the reference is not reachable and the error never drops. It is kept here as a record of what the selection criterion missed, since the criterion only looked at retargeting foot error and never asked whether the target was physically possible on flat ground.

Source data and license

The reference motions derive from LAFAN1 by Ubisoft, released under CC BY-NC-ND 4.0. That license does not permit sharing adapted material, so the retargeted motion data is not included here — only the trained weights and the measured numbers. The onnx export carries the actor network alone and no reference motion.

Use is non-commercial. Credit Ubisoft for LAFAN1.