Motion modeling and search method based on hidden Markov model for motion database
Yoshihiro Ohnishi, Seiichiro Katsura · 2011
In order to address the problem such as aging of workers and craftsmen, `motion database' is considered to be important. Conventionally, motion capture or acceleration sensor is generally used for motion modeling. However, these methods can not get force information, which is important factor for motion modeling. In this research, motion search method based on hidden Markov model is proposed in order to make the `motion database' practical. In the motion modeling, force information is obtained using motion copying system. Because the force information is include in the model, more developed motion database is expected. In addition, motion search becomes possible by using hidden Markov model. The validity of proposed method is confirmed by experiment.