Learning Spring Mass Locomotion: Guiding Policies with a Reduced-Order Model
| Autor: | Green, Kevin, Godse, Yesh, Dao, Jeremy, Hatton, Ross L., Fern, Alan, Hurst, Jonathan |
|---|---|
| Rok vydání: | 2020 |
| Předmět: | |
| Druh dokumentu: | Working Paper |
| Popis: | In this paper, we describe an approach to achieve dynamic legged locomotion on physical robots which combines existing methods for control with reinforcement learning. Specifically, our goal is a control hierarchy in which highest-level behaviors are planned through reduced-order models, which describe the fundamental physics of legged locomotion, and lower level controllers utilize a learned policy that can bridge the gap between the idealized, simple model and the complex, full order robot. The high-level planner can use a model of the environment and be task specific, while the low-level learned controller can execute a wide range of motions so that it applies to many different tasks. In this letter we describe this learned dynamic walking controller and show that a range of walking motions from reduced-order models can be used as the command and primary training signal for learned policies. The resulting policies do not attempt to naively track the motion (as a traditional trajectory tracking controller would) but instead balance immediate motion tracking with long term stability. The resulting controller is demonstrated on a human scale, unconstrained, untethered bipedal robot at speeds up to 1.2 m/s. This letter builds the foundation of a generic, dynamic learned walking controller that can be applied to many different tasks. Comment: 7 pages, 8 figures. Accepted to IEEE Robotics and Automation Letters (RA-L) with ICRA 2021 presentation option. Video supplement: https://youtu.be/80oJeaAd8CE Code: https://github.com/osudrl/ASLIP-RL |
| Databáze: | arXiv |
| Externí odkaz: |
|