A two-fingered underactuated anthropomorphic manipulator based on human precision manipulation motions

2016 IEEE International Conference on Robotics and Automation (ICRA)(2016)

引用 4|浏览30
暂无评分
摘要
While designing robot hands based on grasping data is more common, fewer previous works have used details of human manipulation kinematics to improve robot hand design. The current work involves an underactuated, tendon driven, anthropomorphic manipulator with two flexor tendons and an abduction-adduction tendon, and describes its design based on experimental human precision manipulation data. Link lengths, joint axis alignment, and moment arms were derived from human subject data and values in the literature. The spring ratios, determining the torque relationships between joints, were then selected to maximize the achievable manipulation workspace from the human trial without requiring large forces, which are likely to lead to instability and object ejection. This is done by minimizing the stored spring energy in the robotic fingers across the range of precision manipulation workspace positions achieved by a representative human subject. After fabricating the hand, the energy characteristics of the resulting prototype are analyzed, and the robotic workspace is compared against the original human one. Despite only having three actuators, the hand is able to manipulate the test object within a 2.7 cm 3 workspace volume, compared to an average human workspace of 5.4 cm 3 for the same object. Future work could include adding antagonist actuators to achieve a larger motion range along the palmar-dorsal axis, which is currently the most limited axis of motion in comparison to the original human workspace.
更多
查看译文
关键词
average human workspace,antagonist actuators,palmar-dorsal axis,robotic fingers,spring energy,moment arms,joint axis alignment,link lengths,flexor tendons,abduction-adduction tendon,robot hand design,human manipulation kinematics,human precision manipulation motions,two-fingered underactuated anthropomorphic manipulator
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要