Abstract
It is known that human gait includes information that is specific to the individual and is affected by mood. In addition, identification of emotion using electromyogram (EMG) information has been done with a variety of methods until now. The purpose of this research is to develop methods that estimate human mood from the EMG information during gait. The subjects were stimulated by doing the tasks to evoke mood. After each task, the subjects assessed their mood at that time using the psychological test, and then the subjects walked. We extracted the feature using the principal component analysis from the measured EMG, applied regression analysis with the score of the psychological test. Resulting linear model is statistically significant, and estimates 32% of the variation in the mood based on the EMG data of the upper limb. These results indicate that EMG data during gait is available for mood estimation.
