ETH's Automated Curriculum RL Gets ANYmal Robot to 2.5 m/s on Rough Terrain

ETH Zurich researchers published LP-ACRL, a learning-progress-based automated curriculum RL method that trains a single policy for the ANYmal D quadruped to traverse diverse challenging terrain, reaching 3.0 m/s on flat ground and 2.5 m/s on rough terrain after teacher-student distillation.

2026-09-30 ~ 2026-09-30 · 3 related posts

• Single distilled policy drives ANYmal D to 3.0 m/s on flat and 2.5 m/s on rough terrain — breadli428 · 2026-09-30
• ETH Zurich's Auto-Curriculum RL Gets ANYmal to 2.5 m/s on Rough Terrain With One Policy — breadli428 · 2026-09-30

1 near-duplicate retellings: ChongZzZhang