• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2019, Vol. 55 ›› Issue (11): 105-113.doi: 10.3901/JME.2019.11.105

• 特邀专栏:共融机器人 • 上一篇    下一篇

扫码分享

刚-柔耦合仿生手指设计及运动学特性

杨扬, 肖晓晓, 南卓江, 刘娜, 李小毛, 彭艳   

  1. 上海大学机电工程与自动化学院 上海 200444
  • 收稿日期:2018-07-02 修回日期:2019-03-20 出版日期:2019-06-05 发布日期:2019-06-05
  • 通讯作者: 李小毛(通信作者),男,1981年出生,博士,副研究员。主要研究方向为特种机器人。E-mail:lixiaomao@shu.edu.cn
  • 作者简介:杨扬,男,1986年出生,博士,讲师。主要研究方向为特种机器人、仿生机器人、执行器。E-mail:yangyang_shu@shu.edu.cn
  • 基金资助:
    国家自然科学基金(91648119,91748116,51575333)、上海市青年科技英才扬帆计划(17YF1406200)和上海高校青年东方学者(QD2016029)资助项目。

Design and Kinematic Characteristics of Rigid-flexible Coupling Bionic Finger

YANG Yang, XIAO Xiaoxiao, NAN Zhuojiang, LIU Na, LI Xiaomao, PENG Yan   

  1. School of Mechanical Engineering and Automation, Shanghai University, Shanghai 200444
  • Received:2018-07-02 Revised:2019-03-20 Online:2019-06-05 Published:2019-06-05

摘要: 仿生手能够实现各类复杂的动作,广泛应用于工业及医疗领域。现有的刚性仿生手大多具有质量重、柔顺性差等缺陷,而柔性仿生手的控制精度低、难以实现有力的抓握。针对上述不足,以人手的生理结构为设计依据,运用3D打印技术制成刚性人手骨骼模型,以日本东京工业大学铃森康一教授团队开发的1.3 mm和2 mm直径的细径McKibben型气动人工肌肉作为柔性执行器,参考人手肌肉布局分别模拟人手的内在肌与外在肌,研发一种刚-柔耦合仿生手指。利用BP人工神经网络构建手指的运动学模型,分析手指关节运动和肌肉间的耦合驱动关系,通过试验测试结果评估模型的预测精度,通过对比解剖学的研究成果发现,提出的仿生手指在驱动原理和运动学特性上与人手具有高度的相似性。

关键词: BP人工神经网络, McKibben型气动人工肌肉, 仿生手指, 运动学模型

Abstract: The bionic hand can achieve various complex movements, which is widely used in the field of industry and medical treatment. While the existing rigid bionic hand has limitations of heave weight and poor flexibility, the soft hand with low control accuracy is difficult to achieve powerful griping. Therefore, based on the structure of human hand, this study develops a rigid-flexible coupling bionic finger. While the rigid bones are fabricated by 3D printing technology, the intrinsic muscles and extrinsic muscles respectively use the thin-McKibben pneumatic artificial muscles of diameters 1.3 mm and 2 mm developed by Prof. Koichi Suzumori in Tokyo Institute of Technology. A kinematic model of bionic finger is established by using BP artificial neural network, and the coupling relationship between the joint motion and muscle is analyzed. The prediction accuracy of the model is evaluated by the experiments. Comparing with results of anatomy, it can be confirmed that the developed hand has great similar driving principle and kinematic characteristics with human hand.

Key words: bionic finger, BP artificial neural network, kinematic model, McKibben pneumatic artificial muscle

中图分类号: