| Project/Area Number |
17K12759
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Intelligent robotics
|
|---|
| Research Institution | Nara Institute of Science and Technology |
|---|
Principal Investigator |
Kobayashi Taisuke 奈良先端科学技術大学院大学, 先端科学技術研究科, 助教 (10796452)
|
|---|
| Project Period (FY) |
2017-04-01 – 2020-03-31
|
| Project Status |
Completed (Fiscal Year 2019)
|
| Budget Amount *help |
¥4,030,000 (Direct Cost: ¥3,100,000、Indirect Cost: ¥930,000)
Fiscal Year 2019: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2018: ¥1,820,000 (Direct Cost: ¥1,400,000、Indirect Cost: ¥420,000)
Fiscal Year 2017: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
|
|---|
| Keywords | 知能ロボティクス / 強化学習 / 多目的最適化 / 継続学習 / 歩行 / 知能ロボティックス |
|---|
| Outline of Final Research Achievements |
The purpose of this study is to achieve locomotion control of legged robot as a hierarchical multi-objective optimization problem. With the establishment of this technology, physical constraints and trade-offs can be explicitly considered, and animal-like natural locomotion can be expected.
Three outcomes related to this study, (1) regularization technique that can continuously accumulate learning results without forgetting, (2) policy with search ability to discover global optimal solutions, (3) structured neural networks that facilitate modularization and hierarchy of knowledge. By combining these techniques on a curriculum that sequentially learns locomotion of a quadrupedal robot from the lower hierarchical modules to the upper hierarchical modules, we succeeded in generating the locomotion on a simulation model of the developed quadrupedal robot.
|
| Academic Significance and Societal Importance of the Research Achievements |
本研究の脚ロボットの学習制御は，近年の入出力関係を直接学習してしまう大雑把なやり方では隠蔽されてしまう知識の階層関係や構成要素を．これまでの歩容に関する研究を踏まえて明示的に与えて継続的に学習を積み重ねていくことが可能な枠組みを提供しており，機械学習分野とロボティクス分野の融合領域として高い学術的意義がある．また，安定かつ高効率な歩容制御の確立はロボットの移動範囲を格段に広げて日常的にロボットが活躍するための基礎技術となり，今後のロボット共生社会の実現に繋がるものと期待できる．
|
