This paper discusses the grasp force sensing capabilities of the fingernail imaging method integrated with a visual servoing robotic system. The effectiveness of the fingernail imaging method has been demonstrated on the previous works in the prediction of 3-D fingertip forces. In this study, the fingernail imaging method has been modified to be used in constrained grasping studies. Moreover, the technique can be extended to be applied to the unconstrained grasping study as well. Visual servoing has been utilized in this paper to solve the issue of keeping fingernail images in the field of view of the camera during unconstrained grasping motions. The experimental results show the effectiveness of applying visual servoing for use with the fingernail imaging method to be used in grasping studies. Experimental studies were performed on 2 human subjects and the mean value of RMS errors for predicted normal forces during grasping has been found as 0.57 N. (5.7% for the range of 0–10 N)

The intricate tendon system of the human muscular-skeletal system contributes to the human hand’s dexterity. A complex bond graph model of the index finger was developed to give insight into this system. Previous validation of this model by use of the ACT hand was difficult due to static joint friction. A new robotic testbed, Utah’s Anatomically-correct Robotic Testbed (UART) finger, has been developed to mitigate this friction. Static force and position experiments were conducted with the UART finger in contact with a surface and were compared to the bond graph model. The results suggest that the model is capable of simultaneously predicting static poses and fingertip force. The average predicted joint angle error was 2.9°. The average fingertip force magnitude error was 7.4%, and the average fingertip force direction error was 4.3°.