ParkourFormer: Integrating Predictive Supervision and Sequence Modeling into Parkour Locomotion
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914601848799232 |
|---|---|
| author | Mai, Yanheng Xu, Wenhao Huang, Zirui Fu, Yifei Dong, Shengwei Wang, Xinjue Huang, Kailun Xie, Yanzhe Xu, Renjing |
| author_facet | Mai, Yanheng Xu, Wenhao Huang, Zirui Fu, Yifei Dong, Shengwei Wang, Xinjue Huang, Kailun Xie, Yanzhe Xu, Renjing |
| contents | Humanoid parkour requires locomotion policies to coordinate whole-body dynamics across rapidly changing terrains such as stairs, gaps, slopes, and obstacles. Existing reinforcement learning policies are largely reactive, mapping observations directly to actions without explicitly modeling future body states. Such modeling becomes critical in agile locomotion tasks where successful motion execution depends strongly on anticipating upcoming contact transitions and body dynamics. We present ParkourFormer, a Transformer-based sequence modeling framework that reformulates humanoid locomotion as a future-conditioned decision-making problem. The current robot state queries historical sensorimotor trajectories through cross-attention, while a lightweight prediction head forecasts short-horizon future proprioceptive states. The predicted future states, trained with supervised signals, are fused with temporal features to generate actions, enabling the policy to jointly reason over motion history and anticipated future dynamics. We evaluate ParkourFormer on a diverse multi-terrain humanoid parkour benchmark including stairs, gaps, slopes, rough terrain, and obstacle traversal. Experiments in simulation and on a real humanoid robot show that ParkourFormer achieves a 93.85% average traversal success rate on highly challenging terrains, with improvements of up to 42.73% over strong MLP, MoE-based MLP, and vanilla Transformer baselines, while maintaining a single unified policy across all terrain types. These results demonstrate that explicit future-state modeling significantly improves robustness and generalization for agile whole-body locomotion. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_25782 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ParkourFormer: Integrating Predictive Supervision and Sequence Modeling into Parkour Locomotion Mai, Yanheng Xu, Wenhao Huang, Zirui Fu, Yifei Dong, Shengwei Wang, Xinjue Huang, Kailun Xie, Yanzhe Xu, Renjing Robotics Humanoid parkour requires locomotion policies to coordinate whole-body dynamics across rapidly changing terrains such as stairs, gaps, slopes, and obstacles. Existing reinforcement learning policies are largely reactive, mapping observations directly to actions without explicitly modeling future body states. Such modeling becomes critical in agile locomotion tasks where successful motion execution depends strongly on anticipating upcoming contact transitions and body dynamics. We present ParkourFormer, a Transformer-based sequence modeling framework that reformulates humanoid locomotion as a future-conditioned decision-making problem. The current robot state queries historical sensorimotor trajectories through cross-attention, while a lightweight prediction head forecasts short-horizon future proprioceptive states. The predicted future states, trained with supervised signals, are fused with temporal features to generate actions, enabling the policy to jointly reason over motion history and anticipated future dynamics. We evaluate ParkourFormer on a diverse multi-terrain humanoid parkour benchmark including stairs, gaps, slopes, rough terrain, and obstacle traversal. Experiments in simulation and on a real humanoid robot show that ParkourFormer achieves a 93.85% average traversal success rate on highly challenging terrains, with improvements of up to 42.73% over strong MLP, MoE-based MLP, and vanilla Transformer baselines, while maintaining a single unified policy across all terrain types. These results demonstrate that explicit future-state modeling significantly improves robustness and generalization for agile whole-body locomotion. |
| title | ParkourFormer: Integrating Predictive Supervision and Sequence Modeling into Parkour Locomotion |
| topic | Robotics |
| url | https://arxiv.org/abs/2605.25782 |