{"ID":2875631,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2509.02815","arxiv_id":"2509.02815","title":"Multi-Embodiment Locomotion at Scale with extreme Embodiment Randomization","abstract":"We present a single, general locomotion policy trained on a diverse collection of 50 legged robots. By combining an improved embodiment-aware architecture (URMAv2) with a performance-based curriculum for extreme Embodiment Randomization, our policy learns to control millions of morphological variations. Our policy achieves zero-shot transfer to unseen real-world humanoid and quadruped robots.","short_abstract":"We present a single, general locomotion policy trained on a diverse collection of 50 legged robots. By combining an improved embodiment-aware architecture (URMAv2) with a performance-based curriculum for extreme Embodiment Randomization, our policy learns to control millions of morphological variations. Our policy achi...","url_abs":"https://arxiv.org/abs/2509.02815","url_pdf":"https://arxiv.org/pdf/2509.02815v1","authors":"[\"Nico Bohlinger\",\"Jan Peters\"]","published":"2025-09-02T20:32:02Z","proceeding":"cs.RO","tasks":"[\"cs.RO\",\"cs.LG\"]","methods":"[]","has_code":false}