Learning to Race in Minutes: Infoprop Dyna on the Mini Wheelbot
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_ | 1866913083034697728 |
|---|---|
| author | Subhasish, Devdutt Hose, Henrik Trimpe, Sebastian |
| author_facet | Subhasish, Devdutt Hose, Henrik Trimpe, Sebastian |
| contents | Reinforcement Learning (RL) has the potential to enable robots with fast, nonlinear, and unstable dynamics to reach the limits of their performance. However, most recent advances rely on carefully designed physics-based simulators and domain randomization to achieve successful sim-to-real transfer within reasonable wall-clock time. In this work, we bypass the need for such simulators and demonstrate that Infoprop Dyna, a state-of-the-art uncertainty-aware model-based reinforcement learning (MBRL) framework, can enable robots to learn directly from real-world interactions. Using Infoprop Dyna, the Mini Wheelbot, an underactuated unicycle robot, learns to race around a track within 11 minutes of real-world experience. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_01096 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Learning to Race in Minutes: Infoprop Dyna on the Mini Wheelbot Subhasish, Devdutt Hose, Henrik Trimpe, Sebastian Machine Learning Robotics Reinforcement Learning (RL) has the potential to enable robots with fast, nonlinear, and unstable dynamics to reach the limits of their performance. However, most recent advances rely on carefully designed physics-based simulators and domain randomization to achieve successful sim-to-real transfer within reasonable wall-clock time. In this work, we bypass the need for such simulators and demonstrate that Infoprop Dyna, a state-of-the-art uncertainty-aware model-based reinforcement learning (MBRL) framework, can enable robots to learn directly from real-world interactions. Using Infoprop Dyna, the Mini Wheelbot, an underactuated unicycle robot, learns to race around a track within 11 minutes of real-world experience. |
| title | Learning to Race in Minutes: Infoprop Dyna on the Mini Wheelbot |
| topic | Machine Learning Robotics |
| url | https://arxiv.org/abs/2605.01096 |