| Presentation | 2020-03-04 Inverted pendulum control with redundancy by freezing model using deep reinforcement learning Koki Hirakawa, Naohiro Fukumura, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | In recent years, various intelligent robots have been researched and developed, and multi-degree-of-freedom robots that can perform skillful tasks is expected. In order to learn tasks, the robot needs to operate in the real environment by itself, and it is necessary to reduce the number of learning trials. Since the living body including humans has many joints, the freezing hypothesis has been proposed that increases the number of joints used for motor task according to the learning progress during motor learning. In previous research, it was confirmed that the learning efficiency of the freezing model including multiple modules with different degrees of freedom increased. In this research, we introduce a deep reinforcement learning to the freezing model, and verify whether the model is effective in a simulation environment closer to a real robot. From the results of simulation experiment to control an inverted pendulum attached to the tip of the robot arm, we confirm that the deep reinforcement learning is effective to learn the control rule and verify the effect of the parameter for the selection probability in competitive learning. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Freezing model / Deep reinforcement learning / Inverted pendulum control / ROS |
| Paper # | NC2019-76 |
| Date of Issue | 2020-02-26 (NC) |
| Conference Information | |
| Committee | NC / MBE |
|---|---|
| Conference Date | 2020/3/4(3days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | University of Electro Communications |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Neuro Computing, Medical Engineering, etc. |
| Chair | Hayaru Shouno(UEC) / Taishin Nomura(Osaka Univ.) |
| Vice Chair | Kazuyuki Samejima(Tamagawa Univ) / Takashi Watanabe(Tohoku Univ.) |
| Secretary | Kazuyuki Samejima(NAIST) / Takashi Watanabe(NTT) |
| Assistant | Takashi Shinozaki(NICT) / Ken Takiyama(TUAT) / Yasuyuki Suzuki(Osaka Univ.) / Akihiro Karashima(Tohoku Inst. of Tech.) |
| Paper Information | |
| Registration To | Technical Committee on Neurocomputing / Technical Committee on ME and Bio Cybernetics |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Inverted pendulum control with redundancy by freezing model using deep reinforcement learning |
| Sub Title (in English) | |
| Keyword(1) | Freezing model |
| Keyword(2) | Deep reinforcement learning |
| Keyword(3) | Inverted pendulum control |
| Keyword(4) | ROS |
| 1st Author's Name | Koki Hirakawa |
| 1st Author's Affiliation | Toyohashi University of Technology(Toyohashi Univ. of Tech) |
| 2nd Author's Name | Naohiro Fukumura |
| 2nd Author's Affiliation | Toyohashi University of Technology(Toyohashi Univ. of Tech) |
| Date | 2020-03-04 |
| Paper # | NC2019-76 |
| Volume (vol) | vol.119 |
| Number (no) | NC-453 |
| Page | pp.pp.3-8(NC), |
| #Pages | 6 |
| Date of Issue | 2020-02-26 (NC) |