sprint-gait1
sprint-gait1
A perpetual policy for the microduck (61-D observation, 14 actions, 50 Hz). Runs until told otherwise — a gait, loaded into a policy slot.
Run it on a robot
sudo robotctl policy load <slot> SccHy/microduck-sprint-gait1
The observation normalizer is baked into policy.onnx; feed raw observations.
manifest.json follows schema 2 of the microduck policy manifest (docs/policy-manifest.md in the daemon repo).
Curriculum learning
Trained via a stage-by-stage curriculum: gait shaping and terrain difficulty ramped up progressively. Stages 1→4 of the wall-school curriculum, from flat-ground learning to the final full-height behavior:
| Stage 1 | Stage 2 |
|---|---|
 |
 |
| Stage 3 | Stage 4 |
 |
 |
Arena result: 1.777s (5/5) — local eval best ck1000 = 1.654s.
Training
- repo:
pollen-robotics/microduck_rl - branch:
challenge-recipe-publishing - commit:
1efd853be
- Downloads last month
- 38