부산대학교 도서관

Sensory gloves convert hand postures and movements of fingers into electric signals. Different technologies can be adopted to achieve this conversion, and different approaches can be used to evaluate its effectiveness. In this study, we adopted two types of sensory gloves based on two types of sensors, namely, the resistive flex sensor (RFS) and the inertial measurement unit (IMU). We evaluated the conversion effectiveness in terms of repeatability, reproducibility, and reliability of quasi-static measurements. In particular, to consider the nonlinear characteristics of sensors, we propose an improvement of the usually adopted measurement test protocol used for assessing the performance of sensory gloves. According to our results, the two sensory gloves have similar reliability. However, the IMU-based glove provides better repeatability, and the RFS-based glove provides better reproducibility. Overall, the average range ± standard deviation and the intraclass correlation coefficient of the RFS-based glove were 5.66°± 2.22° and 0.73 ± 0.17, respectively, and those of the IMU-based glove were 7.80°± 2.47° and 0.76 ± 0.14, respectively. All in all, the novelty of this work concerns the comparison of two types of sensory gloves in terms of quasi-static measurement reproducibility and reliability and the improvement of an existing standard protocol for sensory glove assessment aimed at providing a more comprehensive analysis.