It is amazing for human to control highly complex hand with many degrees of freedom. To explore the mystery of hand, we use correlation analysis on human hand movement dataset, which is recorded from 33 kinds of grasping tasks in daily life, and obtain correlation relationships of all joints by hierarchical cluster analysis. The correlation relationships imply the feature of human hand movement. Thumb move relatively independently and other fingers move relatively synergistically during all grasping tasks. Moreover, DIP and PIP joints of all four fingers connect closer together than MCP joints. Before that work in this paper, we try to use dimensional reduction method, which is the main technique, to study the synergistic characteristic. It also supports the conclusion by the considerable inhomogeneity of index of RREV, which is raised to assess the error of each joint variable.
展开▼
