An Adaptive Imitation Learning Framework for Robotic Complex Contact-Rich Insertion Tasks
Complex contact-rich insertion is a ubiquitous robotic manipulation skill and usually involves nonlinear and low-clearance insertion trajectories as well as varying force requirements. A hybrid trajectory and force learning framework can be utilized to generate high-quality trajectories by imitation...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2022-01-01
|
| Series: | Frontiers in Robotics and AI |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/frobt.2021.777363/full |
| _version_ | 1819228829362683904 |
|---|---|
| author | Yan Wang Cristian C. Beltran-Hernandez Weiwei Wan Kensuke Harada Kensuke Harada |
| author_facet | Yan Wang Cristian C. Beltran-Hernandez Weiwei Wan Kensuke Harada Kensuke Harada |
| author_sort | Yan Wang |
| collection | DOAJ |
| description | Complex contact-rich insertion is a ubiquitous robotic manipulation skill and usually involves nonlinear and low-clearance insertion trajectories as well as varying force requirements. A hybrid trajectory and force learning framework can be utilized to generate high-quality trajectories by imitation learning and find suitable force control policies efficiently by reinforcement learning. However, with the mentioned approach, many human demonstrations are necessary to learn several tasks even when those tasks require topologically similar trajectories. Therefore, to reduce human repetitive teaching efforts for new tasks, we present an adaptive imitation framework for robot manipulation. The main contribution of this work is the development of a framework that introduces dynamic movement primitives into a hybrid trajectory and force learning framework to learn a specific class of complex contact-rich insertion tasks based on the trajectory profile of a single task instance belonging to the task class. Through experimental evaluations, we validate that the proposed framework is sample efficient, safer, and generalizes better at learning complex contact-rich insertion tasks on both simulation environments and on real hardware. |
| first_indexed | 2024-12-23T11:03:30Z |
| format | Article |
| id | doaj.art-24cbd21bddc846a281a0841d8bebd2b4 |
| institution | Directory Open Access Journal |
| issn | 2296-9144 |
| language | English |
| last_indexed | 2024-12-23T11:03:30Z |
| publishDate | 2022-01-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Robotics and AI |
| spelling | doaj.art-24cbd21bddc846a281a0841d8bebd2b42022-12-21T17:49:33ZengFrontiers Media S.A.Frontiers in Robotics and AI2296-91442022-01-01810.3389/frobt.2021.777363777363An Adaptive Imitation Learning Framework for Robotic Complex Contact-Rich Insertion TasksYan Wang0Cristian C. Beltran-Hernandez1Weiwei Wan2Kensuke Harada3Kensuke Harada4Department of Systems Innovation, Graduate School of Engineering Science, Osaka University, Suita, JapanDepartment of Systems Innovation, Graduate School of Engineering Science, Osaka University, Suita, JapanDepartment of Systems Innovation, Graduate School of Engineering Science, Osaka University, Suita, JapanDepartment of Systems Innovation, Graduate School of Engineering Science, Osaka University, Suita, JapanAutomation Research Team, Artificial Intelligence Research Center, National Institute of Advanced Industrial Science and Technology, Tsukuba, JapanComplex contact-rich insertion is a ubiquitous robotic manipulation skill and usually involves nonlinear and low-clearance insertion trajectories as well as varying force requirements. A hybrid trajectory and force learning framework can be utilized to generate high-quality trajectories by imitation learning and find suitable force control policies efficiently by reinforcement learning. However, with the mentioned approach, many human demonstrations are necessary to learn several tasks even when those tasks require topologically similar trajectories. Therefore, to reduce human repetitive teaching efforts for new tasks, we present an adaptive imitation framework for robot manipulation. The main contribution of this work is the development of a framework that introduces dynamic movement primitives into a hybrid trajectory and force learning framework to learn a specific class of complex contact-rich insertion tasks based on the trajectory profile of a single task instance belonging to the task class. Through experimental evaluations, we validate that the proposed framework is sample efficient, safer, and generalizes better at learning complex contact-rich insertion tasks on both simulation environments and on real hardware.https://www.frontiersin.org/articles/10.3389/frobt.2021.777363/fullcompliance controlimitation learningreinforcement learningrobotic assemblyrobot autonomy |
| spellingShingle | Yan Wang Cristian C. Beltran-Hernandez Weiwei Wan Kensuke Harada Kensuke Harada An Adaptive Imitation Learning Framework for Robotic Complex Contact-Rich Insertion Tasks Frontiers in Robotics and AI compliance control imitation learning reinforcement learning robotic assembly robot autonomy |
| title | An Adaptive Imitation Learning Framework for Robotic Complex Contact-Rich Insertion Tasks |
| title_full | An Adaptive Imitation Learning Framework for Robotic Complex Contact-Rich Insertion Tasks |
| title_fullStr | An Adaptive Imitation Learning Framework for Robotic Complex Contact-Rich Insertion Tasks |
| title_full_unstemmed | An Adaptive Imitation Learning Framework for Robotic Complex Contact-Rich Insertion Tasks |
| title_short | An Adaptive Imitation Learning Framework for Robotic Complex Contact-Rich Insertion Tasks |
| title_sort | adaptive imitation learning framework for robotic complex contact rich insertion tasks |
| topic | compliance control imitation learning reinforcement learning robotic assembly robot autonomy |
| url | https://www.frontiersin.org/articles/10.3389/frobt.2021.777363/full |
| work_keys_str_mv | AT yanwang anadaptiveimitationlearningframeworkforroboticcomplexcontactrichinsertiontasks AT cristiancbeltranhernandez anadaptiveimitationlearningframeworkforroboticcomplexcontactrichinsertiontasks AT weiweiwan anadaptiveimitationlearningframeworkforroboticcomplexcontactrichinsertiontasks AT kensukeharada anadaptiveimitationlearningframeworkforroboticcomplexcontactrichinsertiontasks AT kensukeharada anadaptiveimitationlearningframeworkforroboticcomplexcontactrichinsertiontasks AT yanwang adaptiveimitationlearningframeworkforroboticcomplexcontactrichinsertiontasks AT cristiancbeltranhernandez adaptiveimitationlearningframeworkforroboticcomplexcontactrichinsertiontasks AT weiweiwan adaptiveimitationlearningframeworkforroboticcomplexcontactrichinsertiontasks AT kensukeharada adaptiveimitationlearningframeworkforroboticcomplexcontactrichinsertiontasks AT kensukeharada adaptiveimitationlearningframeworkforroboticcomplexcontactrichinsertiontasks |