Finger Motion Measurement System for Telexistence Hand Manipulation

This research focuses on finger motion measurement system of telexistence robot to realize precise hand manipulation. Wearable sensor glove which measures operator's finger joint angles is used in most cases for reproducing his/her finger motion to robot hand. However, angle-based finger motion measurement is difficult to estimate correct position of fingertip because of hand shape difference between individuals. On the other hands, optical position measurement such as motion capture can obtain exact position, but has problem of measuring stability because of occlusion. To overcome this issue, we propose a finger motion measurement system which consists of sensor glove with motion capture. To calibrate individual differences, the kinematic parameters of operator's hand such as bone length are estimated at first, and then the hand model predicts fingertip position from sensor glove data. Once the parameters are obtained, both motion capture and hand model provide position of fingertip in parallel, and the measurement becomes more stably against occlusion. The performances of the proposed system regarding parameter estimation and precision accuracy were evaluated during finger movement in 3D space.