Modeling of multi-degree-of-freedom motions for motion database

Recently, the aging of worker and craftsmen is becoming a great issue because of low birthrate and longevity. Further, importing the knowledge of craftsmen to future generations is considered to be important. However, it is difficult to import such knowledge to many people at once. The "motion database" that realizes recording, searching, and reproduction of human motions is considered to be effective in overcoming this problem. In order to construct the motion database, modeling of human motions is needed. In this research, human motions are modeled the basis of a hidden Markov model, and a searching method based on the model is proposed. By using the modeling, it becomes possible to search for a motion in the motion database. This motion database can be applied for skill acquisition.