WEKO3
アイテム
模倣学習を用いた両腕ロボット着衣介助システムのデザインと開発
https://doi.org/10.18997/00007950
https://doi.org/10.18997/00007950e6037b66-7309-451a-adf2-86a755afd5e1
| 名前 / ファイル | ライセンス | アクション |
|---|---|---|
sei_k_384.pdf (13.6 MB)
|
|
| アイテムタイプ | 学位論文 = Thesis or Dissertation(1) | |||||||
|---|---|---|---|---|---|---|---|---|
| 公開日 | 2020-10-23 | |||||||
| 資源タイプ | ||||||||
| 資源タイプ識別子 | http://purl.org/coar/resource_type/c_db06 | |||||||
| 資源タイプ | doctoral thesis | |||||||
| タイトル | ||||||||
| タイトル | Design and Development of a Dual-arm Robotic Clothing Assistance System using Imitation Learning | |||||||
| 言語 | en | |||||||
| タイトル | ||||||||
| タイトル | 模倣学習を用いた両腕ロボット着衣介助システムのデザインと開発 | |||||||
| 言語 | ja | |||||||
| 言語 | ||||||||
| 言語 | eng | |||||||
| 著者 |
Prakash, Joshi Ravi
× Prakash, Joshi Ravi
|
|||||||
| 抄録 | ||||||||
| 内容記述タイプ | Abstract | |||||||
| 内容記述 | The recent demographic trend across developed nations shows a dramatic increase in the aging population and fallen fertility rates. With the aging population, the number of elderly who need support for their Activities of Daily Living （ADL） such as dressing, is growing. The use of caregivers is universal for the dressing task due to the unavailability of any effective assistive technology. Unfortunately, across the globe, many nations are suffering from a severe shortage of caregivers. Hence, the demand for service robots to assist with the dressing task is increasing rapidly. Robotic Clothing Assistance is a challenging task. The robot has to deal with the following two complex tasks simultaneously, （a） non-rigid and highly flexible cloth manipulation, and （b） safe human-robot interaction while assisting a human whose posture may vary during the task. On the other hand, humans can deal with these tasks rather easily. In this thesis, a framework for Robotic Clothing Assistance by imitation learning from a human demonstration to a compliant dual-arm robot is proposed. In this framework, the dressing task is divided into the following three phases, （a） reaching phase, （b） arm dressing phase, and （c） body dressing phase. The arm dressing phase is treated as a global trajectory modification and implemented by applying the Dynamic Movement Primitives （DMP）. The body dressing phase is represented as a local trajectory modification and executed by employing the Bayesian Gaussian Process Latent Variable Model （BGPLVM）. It is demonstrated that the proposed framework developed towards assisting the elderly is generalizable to various people and successfully performs a sleeveless T-shirt dressing task. Furthermore, in this thesis, various limitations and improvements to the framework are discussed. These improvements include the followings （a） evaluation of Robotic Clothing Assistance, （b） automated wheelchair movement, and （c） incremental learning to perform Robotic Clothing Assistance. Evaluation is necessary for our framework. To make it accessible in care facilities, systematic assessment of the performance, and the devices’ effects on the care receivers and caregivers is required. Therefore, a robotic simulator that mimicks human postures is used as a subject to evaluate the dressing task. The proposed framework involves a wheeled chair’s manually coordinated movement, which is difficult to perform for the elderly as it requires pushing the chair by himself. To this end, using an electric wheelchair, an approach for wheelchair and robot collaboration is presented. Finally, to incorporate different human body dimensions, Robotic Clothing Assistance is formulated as an incremental imitation learning problem. The proposed formulation enables learning and adjusting the behavior incrementally whenever a new demonstration is performed. When found inappropriate, the planned trajectory is modified through physical Human-Robot Interaction （HRI） during the execution. This research work is exhibited to the public at various events such as the International Robot Exhibition （iREX） 2017 at Tokyo （Japan）, the West Japan General Exhibition Center Annex 2018 at Kokura （Japan）, and iREX 2019 at Tokyo （Japan）. | |||||||
| 目次 | ||||||||
| 内容記述タイプ | TableOfContents | |||||||
| 内容記述 | 1 Introduction||2 Related Work||3 Imitation Learning||4 Experimental System||5 Proposed Framework||6 Whole-Body Robotic Simulator||7 Electric Wheelchair-Robot Collaboration||8 Incremental Imitation Learning||9 Conclusion | |||||||
| 備考 | ||||||||
| 内容記述タイプ | Other | |||||||
| 内容記述 | 九州工業大学博士学位論文 学位記番号：生工博甲第384号 学位授与年月日：令和2年9月25日 | |||||||
| キーワード | ||||||||
| 主題Scheme | Other | |||||||
| 主題 | Assistive Robotics | |||||||
| キーワード | ||||||||
| 主題Scheme | Other | |||||||
| 主題 | Robotic Clothing Assistance | |||||||
| キーワード | ||||||||
| 主題Scheme | Other | |||||||
| 主題 | Dynamic Movement Primitives | |||||||
| キーワード | ||||||||
| 主題Scheme | Other | |||||||
| 主題 | Bayesian GPLVM | |||||||
| キーワード | ||||||||
| 主題Scheme | Other | |||||||
| 主題 | Human-Robot Interaction | |||||||
| キーワード | ||||||||
| 主題Scheme | Other | |||||||
| 主題 | Learning form Demonstration | |||||||
| アドバイザー | ||||||||
| 柴田, 智広 | ||||||||
| 学位授与番号 | ||||||||
| 学位授与番号 | 甲第384号 | |||||||
| 学位名 | ||||||||
| 学位名 | 博士（情報工学） | |||||||
| 学位授与年月日 | ||||||||
| 学位授与年月日 | 2020-09-25 | |||||||
| 学位授与機関 | ||||||||
| 学位授与機関識別子Scheme | kakenhi | |||||||
| 学位授与機関識別子 | 17104 | |||||||
| 学位授与機関名 | 九州工業大学 | |||||||
| 学位授与年度 | ||||||||
| 内容記述タイプ | Other | |||||||
| 内容記述 | 令和2年度 | |||||||
| 出版タイプ | ||||||||
| 出版タイプ | VoR | |||||||
| 出版タイプResource | http://purl.org/coar/version/c_970fb48d4fbd8a85 | |||||||
| アクセス権 | ||||||||
| アクセス権 | open access | |||||||
| アクセス権URI | http://purl.org/coar/access_right/c_abf2 | |||||||
| ID登録 | ||||||||
| ID登録 | 10.18997/00007950 | |||||||
| ID登録タイプ | JaLC | |||||||
