つまみ及びつかみ操作における指先の動的特徴量に関する定量分析

抄録

Finger manipulation plays an important role in daily life, and the importance of its analysis is increasing in a variety of research fields. Applications in humanoid robots, artificial hands, and improvement of rehabilitation technology could be aided by a quantitative understanding of finger manipulation, which requires the measurement of natural operations. Although many previous studies examined finger manipulation, the experiments tested only a limited range of objects manipulated by the fingers or the sensation of the fingertip was inhibited by attaching sensors. The current study focused on hold and pick manipulations, exploring dynamic features of the thumb and index fingers. To measure natural operation, we used a haptic wearable sensor that measured finger pressure and 3-axis direction acceleration without inhibiting the operator's sense of touch. We conducted user experiments in seven subjects while they manipulated a cylindrical urethane object. We defined 10 features determined uniquely from the measured values, then calculated the recognition rate of pick and hold movements when various dimensions and sets of features were used. We identified common features among subjects which characterized hold and pick manipulations.