2017 Fiscal Year Annual Research Report
Cutting-edge multi-contact behaviors
| Project/Area Number |
16H02886
|
|---|
| Research Institution | National Institute of Advanced Industrial Science and Technology |
|---|
Principal Investigator |
Kheddar Abder 国立研究開発法人産業技術総合研究所, 情報・人間工学領域, 客員研究員 (90572082)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
森澤 光晴 国立研究開発法人産業技術総合研究所, 情報・人間工学領域, 主任研究員 (00392671)
吉田 英一 国立研究開発法人産業技術総合研究所, 情報・人間工学領域, 研究部門付 (30358329)
金広 文男 国立研究開発法人産業技術総合研究所, 情報・人間工学領域, 研究グループ長 (70356806)
|
|---|
| Project Period (FY) |
2016-04-01 – 2019-03-31
|
| Keywords | 知能ロボティクス / ディジタルヒューマンモデル / 多点接触動作 / 機械学習 / 最適化 |
|---|
| Outline of Annual Research Achievements |
This second fiscal year of the project we achieved tasks and goals from different work-packages (WP).
In WP1 we proposed technique for unobtrusively estimating interaction forces exerted by human subjects in multi-contact. We also addressed Sense of Agency (SoA) techniques in multi-contact planning and published in the following book chapters.
In WP2 we enhanced the planner and the controller with sliding contacts, and preliminary experiments have been conducted on the HRP-4 in the Airbus mock-up. Planning closed-loop chains to support stable force tasks was achieved in the circuit-breaker use-case from Airbus, where the robot HRP-4 plan contact on the panel to support pulling and pushing, implemented together with the multi-modal robot controller. We have devised new algorithms to deal with 3D multi-contact stabilizer and stability criteria. We then developed dynamics multi-contact planning using Model-Predictive Control (MPC) based of mixed integer programming, and finally studied theoretical closed-loop stability and robustness of Quadratic Programming (QP) multi-contact controllers.
In WP3 force regulation in multi-contact motion was tackled, related to the results in WP2. The development is still in progress but force control is already integrated to the multi-contact QP controller. We also investigated integration architecture and closed-loop control and recovery implementation.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
Although the research was delayed due to the delay of delivery of proximity sensors in the collaborating laboratory at CNRS, we could almost achieve the results we have expected except the learning part in WP1.
|
| Strategy for Future Research Activity |
We will intensively work on experimental validation and the investigation on application of learning.
|
Research Products
(5 results)
