The design of E glove hand function evaluation device based on fusion of vision and touch

Abstract

This paper presents an E glove hand function evaluation device based on visual and haptic fusion, and uses the Principal Component Analysis (PCA) algorithm to establish hand sensor distribution model. The PCA analysis chart shows that three sensors distributed on the thumb, forefinger, and middle finger could effectively estimate the grasp motions. Moreover, threshold values for all category models can be selected by the way of adaptive pressure threshold integrating visual aid. At last, five subjects dressed E glove judging the grasp motions under different combinations of sensors. The results show that: the classification accuracy rate depended on the pressure and visual sensor fusion method reached 94 %; the identification rate of the adaptive pressure threshold method to judge the grasp motions can be increased 1.6–1.7 times than only using single camera vision sensor or pressure sensor. Next step, the E glove hand function evaluation device will be further improved such as function of active control to the collected data will be added.