The Surprising Effectiveness of Linear Models for Whole-Body Model-Predictive Control
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911168578191360 |
|---|---|
| author | Bishop, Arun L. Alvarez-Padilla, Juan Schoedel, Sam Sow, Ibrahima Sory Chandrachud, Juee Sharma, Sheitej Kraus, Will Park, Beomyeong Griffin, Robert J. Dolan, John M. Manchester, Zachary |
| author_facet | Bishop, Arun L. Alvarez-Padilla, Juan Schoedel, Sam Sow, Ibrahima Sory Chandrachud, Juee Sharma, Sheitej Kraus, Will Park, Beomyeong Griffin, Robert J. Dolan, John M. Manchester, Zachary |
| contents | When do locomotion controllers require reasoning about nonlinearities? In this work, we show that a whole-body model-predictive controller using a simple linear time-invariant approximation of the whole-body dynamics is able to execute basic locomotion tasks on complex legged robots. The formulation requires no online nonlinear dynamics evaluations or matrix inversions. We demonstrate walking, disturbance rejection, and even navigation to a goal position without a separate footstep planner on a quadrupedal robot. In addition, we demonstrate dynamic walking on a hydraulic humanoid, a robot with significant limb inertia, complex actuator dynamics, and large sim-to-real gap. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_17884 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | The Surprising Effectiveness of Linear Models for Whole-Body Model-Predictive Control Bishop, Arun L. Alvarez-Padilla, Juan Schoedel, Sam Sow, Ibrahima Sory Chandrachud, Juee Sharma, Sheitej Kraus, Will Park, Beomyeong Griffin, Robert J. Dolan, John M. Manchester, Zachary Robotics When do locomotion controllers require reasoning about nonlinearities? In this work, we show that a whole-body model-predictive controller using a simple linear time-invariant approximation of the whole-body dynamics is able to execute basic locomotion tasks on complex legged robots. The formulation requires no online nonlinear dynamics evaluations or matrix inversions. We demonstrate walking, disturbance rejection, and even navigation to a goal position without a separate footstep planner on a quadrupedal robot. In addition, we demonstrate dynamic walking on a hydraulic humanoid, a robot with significant limb inertia, complex actuator dynamics, and large sim-to-real gap. |
| title | The Surprising Effectiveness of Linear Models for Whole-Body Model-Predictive Control |
| topic | Robotics |
| url | https://arxiv.org/abs/2509.17884 |