Abstract: The tendon-driven robotic hands have the disadvantages of more motors,larger mechanism volume and more complex control system compared with underactuated hands,which have many limitations in practical application.Thus,a novel robotic hand based on muscle-mimicking and elastic mechanism was presented.A humanoid finger design was carried out based on anatomical characteristic of human finger.The robotic finger's underactuated system was achieved by double-tendon and elastic mechanism.A method of how to choose the elastic parameter was proposed which guaranted the reachable workspace of the fingertip and avoided unstable grasping situations efficiently.Then,the mapping relationship from joint space to tendon space was established by analysing the finger's kinematics.The position parameter of the finger in the palm was optimized to achieve a better palm size and range of grasping,and the evaluation index was also given.Grasping experiments on objects of different shapes and sizes were carried out,the result showed that the robotic hand had strong adaptive grasping ability and large grasping range.

Key words: underactuated robotic hand, mechanism design, parameter optimization, experiments analysis

摘要： 针对仿肌腱传动灵巧手存在的驱动单元多、机构体积过大、控制系统复杂等问题,提出一种新型的仿肌弹性欠驱动手。以人类手指解剖学特征为思路,进行了仿人手指设计,通过引入双肌腱和弹性元件来实现手指的欠驱动,并给出合理选择弹性参数的方法,保证了手指的可达空间和抓取可靠性。进行了手指运动学分析,建立了关节空间到肌腱空间的映射关系。为获得最优的手掌尺寸及抓取范围,提出一种手指位置参数优化方法,并给出了评价指标,通过对比验证了该方法的优势。最后,进行了不同形状、尺寸物体的抓取实验,结果表明该欠驱动手具有较强的自适应抓取能力、广泛的物体抓取范围。

关键词: 欠驱动手, 机构设计, 参数优化, 实验分析