単・二関節筋同時駆動を模擬したマルチリンクロボットレッグの外力のある環境での歩行制御

カテゴリ: 研究会(論文単位)

論文No: CT11035

グループ名: 【C】電子・情報・システム部門 制御研究会

発行日: 2011/12/20

タイトル(英語): Walking Control of Multiple Link Robotic Leg under External Forces by Emulating Functionality of Combined Mono and Bi-Articular Muscles

著者名: ドゥマン エムレ(東京大学),古関 隆章(東京大学)

著者名(英語): DUMAN EMRE(The University of Tokyo),Koseki Takafumi(The University of Tokyo)

キーワード: 二関節筋|ロボット|歩行制御|ｂｉａｒｔｉｃｕｌａｒ　ｍｕｓｃｌｅ|ｒｏｂｏｔ|ｗａｌｋｉｎｇ　ｃｏｎｔｒｏｌ

要約(日本語): Conventional humanoid robots and biological subjects differ in their mechanisms and control strategy. Biological subjects have musculoskeletal systems that can drive two joints at the same time while conventional robots use one joint drive mechanisms. In this research, a dynamic model of a multiple-link robotic leg that mimics biological subjects is obtained and walking scheme for both legs under external forces are designed and analyzed. The aim of this research is to model and simulate walking behavior of a robotic leg with mono and biarticular muscles by rejecting the effect of external forces. Advantages of using muscle mechanism are that feedforward control can be implemented by using the viscoelasticity of biarticular muscles, stiffness control becomes easier and maximum output force at the end effector is larger in compared to conventional type robot manipulators; but we have actuator redundancy and coordinated activation is more difficult.

要約(英語): Conventional humanoid robots and biological subjects differ in their mechanisms and control strategy. Biological subjects have musculoskeletal systems that can drive two joints at the same time while conventional robots use one joint drive mechanisms. In this research, a dynamic model of a multiple-link robotic leg that mimics biological subjects is obtained and walking scheme for both legs under external forces are designed and analyzed. The aim of this research is to model and simulate walking behavior of a robotic leg with mono and biarticular muscles by rejecting the effect of external forces. Advantages of using muscle mechanism are that feedforward control can be implemented by using the viscoelasticity of biarticular muscles, stiffness control becomes easier and maximum output force at the end effector is larger in compared to conventional type robot manipulators; but we have actuator redundancy and coordinated activation is more difficult.

原稿種別: 日本語

PDFファイルサイズ: 718 Kバイト