4. Human experiment

• We have conducted extra experiments with human hand to show how the robot hysteresis changes when the human hand is used.  

• Three experiments related to hysteresis (section III.C to section III.E) were conducted using the hand of a non-disabled person instead of a dummy hand to better represent hysteresis in real-world situations. We chose a non-disabled participant over a disabled one to help readers easily infer the amount of hysteresis, as there is greater variability in joint stiffness among disabled individuals. 

• Further, to show whether the subject exerts voluntary control of the fingers, we measured the EMG signals of the flexion and extension muscles during the experiments that used the human hand.

  • Two EMG sensors (Trigno, Delsys, USA) were used to measure the mean absolute value (MAV) of the signal. 

  • When the subject makes voluntary motions, the maximum MAV of the extensor and flexor were measured as 0.14mV and 0.11mV, respectively; the MAV of both muscles remained less than 0.02mV in the actual experiment, confirming that the subject did not exert force during the experiment. 

• All procedures are approved by the Institutional Review Board of Seoul National University (IRB No. 22014/001-004).

4.1 Positional Hysteresis

4.2 Hysteresis analysis using extension force

 Despite the change of the experimental conditions, the conclusion of the paper remains unchanged because the overall trends in the results remain consistent, as outlined below:

·   Positional Hysteresis (Section III.C): The hysteresis losses for the previous and proposed robots were 63.27% and 27.91%; when the non-disabled human hand is used, these values are 78.26% and 49.21%, respectively. 

·    Force-level Hysteresis (Section III.D): The normalized extension force indicator was previously 1.07 and 0.27 for the previous and proposed routing, respectively; when the non-disabled human hand is used, these values are 2.19 and 0.55, respectively. 

·    Positional Hysteresis of Exo-Glove II with a passive tendon (Section III.E): The hysteresis loss for the robot with a passive tendon was initially 4.65% when using a dummy hand. Using the human hand, the loss is now measured at 18.23%

Overall, all three experiments demonstrated better results when using a dummy hand compared to a human hand. This phenomenon is likely due to the torsional spring in the dummy hand having greater stiffness than anticipated, despite using a spring designed to match the stiffness of a human hand. While there are slight differences in the numerical results, the overall trends remain consistent: the robot with the proposed tendon routing consistently outperforms the one with the previous routing, and the results in Section IV.E confirm that the use of a passive tendon effectively mitigates hysteresis.