Real-Time Finger Force Estimation Robust to a Perturbation of Electrode Placement for Prosthetic Hand Control.

In the use of real-time myoelectric controlled prostheses, the low accuracy of the user's intention estimation for simultaneous and proportional control (SPC) and the vulnerability to electrode shifts make application to real-world scenarios difficult. To overcome this barrier, we propose a method to estimate muscle unit activation in real time through neurophysiological modeling of the forearm. We also propose a high-performance finger force intention estimation model that is robust to perturbation of electrode placement based on estimated muscle unit activation. We compared the proposed model with previous studies for quantitative validation of finger force intention estimation and electrode shift compensation performance. Compared to other regression-based models in the on/offline test, our model achieved a significantly high intention estimation performance (p < 0.001). In addition, it attained high performance in electrode shift compensation, and at this time, the amount of data required and the number of models utilized were small. In conclusion, the model proposed in this study was verified to be robust to electrode shift and has high finger force intention estimation accuracy.