日本フルードパワーシステム学会論文集
Online ISSN : 1349-7022
Print ISSN : 1880-3121
ISSN-L : 1349-7022
最新号
選択された号の論文の2件中1~2を表示しています
研究論文
  • 見上 慧, 只野 耕太郎
    2022 年 53 巻 1 号 p. 1-9
    発行日: 2022年
    公開日: 2022/02/25
    ジャーナル フリー

    A robot hand is required to generate a large grasping force. Furthermore, it is required that one finger should possess multiple degrees of freedom and the fingertips should be positioned freely to grasp and manipulate objects of various sizes and shapes. A function to adjust the stiffness is also required to control the force and realize stable grasping. The development of a hand with these functions at a high level is a technical problem. We propose a multi-degree-of-freedom finger module that uses pneumatic cylinders and a parallel link mechanism. The proposed hand has adjustable stiffness while maintaining high grasping force-to-weight ratio. In addition, the proposed hand was able to estimate the external force acting on the fingertip from the differential pressure and a kinematics model. By using the estimated external force for control, adjustable stiffness was realized without using a force sensor. This paper also describes the design method of the finger module. We experimentally confirmed that the proposed hand was able to satisfy various target specifications. In addition, it was able to realize multiple functions of a hand using a compact mechanism.

  • 加藤 輝雄, 新井 遼, 酒井 悟
    2022 年 53 巻 1 号 p. 10-17
    発行日: 2022年
    公開日: 2022/02/25
    ジャーナル フリー

    This paper discusses an identification input for nominal hydraulic arms. First, we focus on an input-output nonlinearity which exists even in the absence of nonlinear friction and prevents us from observing and measuring the input-output behaviors. Second, we propose a new identification input based on the physical parameters of hydraulic arms via the special nondimensionalization technique. Third, we show the effectiveness of the new identification input experimentally. Finally, we show an application of the proposed new identification input to observe and measure the input-output behaviors. Especially, in the frequency domain, we identify the uncertainty that was impossible for the conventional identification input.

feedback
Top