Salvato in:
| Autori principali: | , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2605.27724 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866910263991599104 |
|---|---|
| author | Lin, Kevin Mandlekar, Ajay Garrett, Caelan Reed Chernyadev, Nikita Fang, Yu Ding, Runyu Xie, Yuqi Tran, Justin Fan, Linxi Zhu, Yuke |
| author_facet | Lin, Kevin Mandlekar, Ajay Garrett, Caelan Reed Chernyadev, Nikita Fang, Yu Ding, Runyu Xie, Yuqi Tran, Justin Fan, Linxi Zhu, Yuke |
| contents | Imitation learning is a promising approach for training humanoid robots to both walk and manipulate, but it requires a large number of demonstrations, which are time-intensive and difficult to collect via teleoperation. Existing data-generation algorithms can automatically synthesize demonstrations for manipulators, but they are ineffective on humanoids because their high-dimensional composite action spaces involve arms, legs, and torsos. We present HumanoidMimicGen, a method for generating humanoid legged loco-manipulation data. Our method adapts contact-rich whole-body skills from a handful of source demonstrations to new states, generalizing across changes in object pose. By interleaving these single- and dual-arm skills with whole-body locomotion and manipulation planning, the method generates stable, collision-free data across diverse scenes and layouts. To evaluate our approach, we introduce a new simulated loco-manipulation benchmark containing nine diverse tasks that test humanoid loco-manipulation capabilities. There, we demonstrate that HumanoidMimicGen automatically generates large datasets for imitation learning and enables a systematic study of how data generation and policy learning decisions impact model performance. We show that whole-body visuomotor policies co-trained with data generated by HumanoidMimicGen outperform those trained only on real-world data by 20%. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_27724 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | HumanoidMimicGen: Data Generation for Loco-Manipulation via Whole-Body Planning Lin, Kevin Mandlekar, Ajay Garrett, Caelan Reed Chernyadev, Nikita Fang, Yu Ding, Runyu Xie, Yuqi Tran, Justin Fan, Linxi Zhu, Yuke Robotics Artificial Intelligence Imitation learning is a promising approach for training humanoid robots to both walk and manipulate, but it requires a large number of demonstrations, which are time-intensive and difficult to collect via teleoperation. Existing data-generation algorithms can automatically synthesize demonstrations for manipulators, but they are ineffective on humanoids because their high-dimensional composite action spaces involve arms, legs, and torsos. We present HumanoidMimicGen, a method for generating humanoid legged loco-manipulation data. Our method adapts contact-rich whole-body skills from a handful of source demonstrations to new states, generalizing across changes in object pose. By interleaving these single- and dual-arm skills with whole-body locomotion and manipulation planning, the method generates stable, collision-free data across diverse scenes and layouts. To evaluate our approach, we introduce a new simulated loco-manipulation benchmark containing nine diverse tasks that test humanoid loco-manipulation capabilities. There, we demonstrate that HumanoidMimicGen automatically generates large datasets for imitation learning and enables a systematic study of how data generation and policy learning decisions impact model performance. We show that whole-body visuomotor policies co-trained with data generated by HumanoidMimicGen outperform those trained only on real-world data by 20%. |
| title | HumanoidMimicGen: Data Generation for Loco-Manipulation via Whole-Body Planning |
| topic | Robotics Artificial Intelligence |
| url | https://arxiv.org/abs/2605.27724 |