Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : May 10, 2017 - May 13, 2017
In this paper, identification of the grasping motion through electroencephalogram (EEG) was performed. First, for the EEG signals measured at top position of the head, independent component analysis (ICA) was executed to obtain original EEG signals. Next, short time Fourier transform was performed to the obtained EEG signals. Then, the features to identify grasping motion were determined through the results of the spectrum analysis. The neural network (NN) to distinguish existence or nonexistence of the grasping motion, and NN to distinguish right grasping motion or left grasping motion were constructed respectively based on the features. In order to verify the validity of the proposed identification method, identification experiments were carried out. In addition, verification of the effectiveness of performing ICA before spectrum analysis was conducted. The results showed that the grasping motion can be identified using the proposed method and ICA was effective for the identification.
