| Presentation | 2001/3/16 Acquisition of Reaching Movement Based on Reinforcement Learning Katsunari SHIBATA, Masanori SUGISAKA, Koji ITO, |
|---|---|
| PDF Download Page | PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | A learning system with a neural network whose inputs are visual sensory signals and state of a manipulator, and whose outputs are joint torques, obtains the hand reaching movement by reinforcement learning. This paper shows some simulation results in the case of (1)variable link length and in the case of (2)external force which is loaded at the hand. In the former, a hidden neuron's representation is similar to the postcentral neurons of a monkey in Iriki's experiment. The neuron activates when the food is located at the place where the hand can reach it, and the range is extended to the range where a tool can reach when the monkey uses the tool. In the latter, it is shown that the system obtains inverse dynamics of its hand and environment. When it learns in the random force at shown trial, the system becomes to use feedback control mainly to control its hand. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Reinfbrcement Learning / Neural Network / Reaching Task / Tool / Inverse Dynamics / Feedback Control |
| Paper # | NC2000-170 |
| Date of Issue |
| Conference Information | |
| Committee | NC |
|---|---|
| Conference Date | 2001/3/16(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Neurocomputing (NC) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Acquisition of Reaching Movement Based on Reinforcement Learning |
| Sub Title (in English) | |
| Keyword(1) | Reinfbrcement Learning |
| Keyword(2) | Neural Network |
| Keyword(3) | Reaching Task |
| Keyword(4) | Tool |
| Keyword(5) | Inverse Dynamics |
| Keyword(6) | Feedback Control |
| 1st Author's Name | Katsunari SHIBATA |
| 1st Author's Affiliation | Faculty of Engineering, Oita University() |
| 2nd Author's Name | Masanori SUGISAKA |
| 2nd Author's Affiliation | Faculty of Engineering, Oita University |
| 3rd Author's Name | Koji ITO |
| 3rd Author's Affiliation | Dept. of Computational Intolligence & Systems Science, Tokyo Institute of Techinology |
| Date | 2001/3/16 |
| Paper # | NC2000-170 |
| Volume (vol) | vol.100 |
| Number (no) | 688 |
| Page | pp.pp.- |
| #Pages | 8 |
| Date of Issue | |