论文部分内容阅读
分析了人手抓取的各种姿态，提出了３种基本的抓取姿态结构，并举例说明了人手的姿态可分解为３种基本姿态的组合．基于人手的灵巧性，机械手的抓取可模仿人手的抓取特性，根据姿态、物体、任务之间的关系提出了神经网络规划的抓取方法，并设计了相应的神经网络
This paper analyzes the various gestures grasped by the human hand and puts forward three kinds of basic grasping gesture structures and illustrates the combination of the human hand gesture can be decomposed into three basic gestures. Based on the dexterity of the human hand, the grasping of the manipulator can imitate the grasping characteristics of the human hand and put forward the grasping method of the neural network plan according to the relationship between the attitude, the object and the task, and design a corresponding neural network