2022 Fiscal Year Annual Research Report
確率モデルベース強化学習を用いた二足ロボット用歩行コントローラーの開発
| Project/Area Number |
22J14202
|
|---|
| Allocation Type | Single-year Grants |
|---|
| Research Institution | Nara Institute of Science and Technology |
|---|
Principal Investigator |
Kuo ChengーYu 奈良先端科学技術大学院大学, 先端科学技術研究科, 特別研究員(DC2)
|
|---|
| Project Period (FY) |
2022-04-22 – 2024-03-31
|
| Keywords | モデルベース強化学習 / 弾性ロボット / 二脚ロボット / 確率手法 |
|---|
| Outline of Annual Research Achievements |
During this year, we utilized model-based reinforcement learning to capture the dynamics of a spring-loaded biped robot. This is significant because the robot's compliancy makes modeling analytic dynamics difficult. Using the learned dynamics, we successfully performed hopping tasks with real-time planning in a simulation environment. These results demonstrate the effectiveness of our approach. Additionally, we have begun hardware implementation, achieving a simple walking task. The performance is expected to significantly improve in the following year.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
In our first year, we made significant progress in researching bipedal robots. We used model-based reinforcement learning to capture their dynamics, which is notable because these robots are difficult to model analytically due to their compliance. Using the learned dynamics, we successfully performed hopping tasks with real-time planning in a simulation environment. We then implemented this approach in hardware and achieved a simple walking task, a crucial step towards testing it in the real world. We are confident and eager to continue making progress in the field of bipedal robots.
|
| Strategy for Future Research Activity |
We will conduct hardware testing and modifications, compare our approach with other locomotion methods, and analyze robot performance metrics like speed, energy efficiency, and stability. We will also explore potential applications beyond bipedal robots. Our ultimate goal is to develop efficient bipedal robots for human environments and make a significant impact in this field.
|
