Learning a Vision-Based Footstep Planner for Hierarchical Walking Control
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866912530003132416 |
|---|---|
| author | Kim, Minku Acosta, Brian Chaudhari, Pratik Posa, Michael |
| author_facet | Kim, Minku Acosta, Brian Chaudhari, Pratik Posa, Michael |
| contents | Bipedal robots demonstrate potential in navigating challenging terrains through dynamic ground contact. However, current frameworks often depend solely on proprioception or use manually designed visual pipelines, which are fragile in real-world settings and complicate real-time footstep planning in unstructured environments. To address this problem, we present a vision-based hierarchical control framework that integrates a reinforcement learning high-level footstep planner, which generates footstep commands based on a local elevation map, with a low-level Operational Space Controller that tracks the generated trajectories. We utilize the Angular Momentum Linear Inverted Pendulum model to construct a low-dimensional state representation to capture an informative encoding of the dynamics while reducing complexity. We evaluate our method across different terrain conditions using the underactuated bipedal robot Cassie and investigate the capabilities and challenges of our approach through simulation and hardware experiments. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_06779 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Learning a Vision-Based Footstep Planner for Hierarchical Walking Control Kim, Minku Acosta, Brian Chaudhari, Pratik Posa, Michael Robotics Bipedal robots demonstrate potential in navigating challenging terrains through dynamic ground contact. However, current frameworks often depend solely on proprioception or use manually designed visual pipelines, which are fragile in real-world settings and complicate real-time footstep planning in unstructured environments. To address this problem, we present a vision-based hierarchical control framework that integrates a reinforcement learning high-level footstep planner, which generates footstep commands based on a local elevation map, with a low-level Operational Space Controller that tracks the generated trajectories. We utilize the Angular Momentum Linear Inverted Pendulum model to construct a low-dimensional state representation to capture an informative encoding of the dynamics while reducing complexity. We evaluate our method across different terrain conditions using the underactuated bipedal robot Cassie and investigate the capabilities and challenges of our approach through simulation and hardware experiments. |
| title | Learning a Vision-Based Footstep Planner for Hierarchical Walking Control |
| topic | Robotics |
| url | https://arxiv.org/abs/2508.06779 |