|
JAIST Repository >
b. 情報科学研究科・情報科学系 >
b11. 会議発表論文・発表資料等 >
b11-1. 会議発表論文・発表資料 >
このアイテムの引用には次の識別子を使用してください:
http://hdl.handle.net/10119/11608
|
| タイトル: | Autonomous Learning Based on Depth Perception and Behavior Generation |
| 著者: | Jeong, Sungmoon
Park, Yunjung
Lee, Minho |
| キーワード: | component
Autonomous learning
sensory invariance driven action
size invariance of object perception |
| 発行日: | 2013-08-18 |
| 出版者: | Institute of Electrical and Electronics Engineers (IEEE) |
| 誌名: | 2013 IEEE Third Joint International Conference on Development and Learning and Epigenetic Robotics (ICDL) |
| 開始ページ: | 1 |
| 終了ページ: | 2 |
| DOI: | 10.1109/DevLrn.2013.6652531 |
| 抄録: | We propose a new neuro-robotic network that can achieve a goal oriented behavior for a visually-guided object manipulation tasks based on learning by examples. The proposed model considers a brain-like interaction between behavior generation and depth perception in mammal brain to autonomously improve the robot’s action performance and perception accuracy. The brain exploits action to develop perception qualities, and this updated perceptual process helps to develop qualified-behavior. The perceptual accuracy can be enhanced by observing the effects of actions that preserve a physical and/or perceptual invariance. Also, the improved perceptual accuracy can influence the optimal selection and/or modification of actions. In order to import those action and perception abilities of a brain into a humanoid robot, we considered two key inspirations: 1) Sensory Invariant Driven Action (SIDA) and 2) Object Size Invariance (OSI) in depth perception. Considering robot manipulation of a target object with distance estimation as a perceptual process, we develop a new autonomous learning method based on the SIDA for behavior generation and OSI property for perceptual judgment. The proposed method is evaluated by using a humanoid robot (NAO) with stereo cameras, and the experimental results show that the proposed method is effective on autonomously improving the behavior generation performance as well as depth perception accuracy. |
| Rights: | This is the author's version of the work. Copyright (C) 2013 IEEE. 2013 IEEE Third Joint International Conference on Development and Learning and Epigenetic Robotics (ICDL), 2013, 1-2, DOI:10.1109/DevLrn.2013.6652531. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. |
| URI: | http://hdl.handle.net/10119/11608 |
| 資料タイプ: | author |
| 出現コレクション: | b11-1. 会議発表論文・発表資料 (Conference Papers)
|
このアイテムのファイル:
| ファイル |
記述 |
サイズ | 形式 |
|---|
| 19555.pdf | | 558Kb | Adobe PDF | 見る/開く |
|
当システムに保管されているアイテムはすべて著作権により保護されています。
|
