Reinforcement Learning for 4-Finger-Gripper Manipulation
(2018) International Conference on Robotics and Automation (ICRA) 2018- Abstract
- In the framework of robotics, Reinforcement Learning (RL) deals with the learning of a task by the robot itself. This paper presents a hierarchical planning approach in which the robot learns the optimal behavior for different levels. For high-level discrete actions, Q-learning was chosen, whereas for the low level we utilize Policy Improvement with Path Integrals (PI^2) algorithm to learn the parameters of policies, represented by rhythmic Dynamic Movement Primitives (DMPs). The paper studies the case of a 4-finger-gripper manipulator, which performs the task of continuously spinning a ball around a desired axis. The results demonstrate the efficacy of the hierarchical planning and the improvement obtained in the performance of the task... (More)
- In the framework of robotics, Reinforcement Learning (RL) deals with the learning of a task by the robot itself. This paper presents a hierarchical planning approach in which the robot learns the optimal behavior for different levels. For high-level discrete actions, Q-learning was chosen, whereas for the low level we utilize Policy Improvement with Path Integrals (PI^2) algorithm to learn the parameters of policies, represented by rhythmic Dynamic Movement Primitives (DMPs). The paper studies the case of a 4-finger-gripper manipulator, which performs the task of continuously spinning a ball around a desired axis. The results demonstrate the efficacy of the hierarchical planning and the improvement obtained in the performance of the task when PI^2 is used in conjunction with rhythmic DMPs in a real environment. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/1d8fbbc4-d89b-451b-995d-53dae6d6c3a5
- author
- Ojer de Andrés, Marco ; Ghazaei Ardakani, Mahdi LU and Robertsson, Anders LU
- organization
- publishing date
- 2018
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- host publication
- Proceedings of 2018 IEEE International Conference on Robotics and Automation
- conference name
- International Conference on Robotics and Automation (ICRA) 2018
- conference location
- Brisbane, Australia
- conference dates
- 2018-05-21 - 2018-05-25
- external identifiers
-
- scopus:85052794370
- language
- English
- LU publication?
- yes
- id
- 1d8fbbc4-d89b-451b-995d-53dae6d6c3a5
- date added to LUP
- 2018-05-17 13:42:23
- date last changed
- 2023-02-22 10:37:12
@inproceedings{1d8fbbc4-d89b-451b-995d-53dae6d6c3a5, abstract = {{In the framework of robotics, Reinforcement Learning (RL) deals with the learning of a task by the robot itself. This paper presents a hierarchical planning approach in which the robot learns the optimal behavior for different levels. For high-level discrete actions, Q-learning was chosen, whereas for the low level we utilize Policy Improvement with Path Integrals (PI^2) algorithm to learn the parameters of policies, represented by rhythmic Dynamic Movement Primitives (DMPs). The paper studies the case of a 4-finger-gripper manipulator, which performs the task of continuously spinning a ball around a desired axis. The results demonstrate the efficacy of the hierarchical planning and the improvement obtained in the performance of the task when PI^2 is used in conjunction with rhythmic DMPs in a real environment.}}, author = {{Ojer de Andrés, Marco and Ghazaei Ardakani, Mahdi and Robertsson, Anders}}, booktitle = {{Proceedings of 2018 IEEE International Conference on Robotics and Automation}}, language = {{eng}}, title = {{Reinforcement Learning for 4-Finger-Gripper Manipulation}}, year = {{2018}}, }