This study evaluated the estimation accuracy of grip strength using surface electromyography (sEMG) recorded from the wrist, a novel electromyography (EMG) acquisition site. Previous studies on grip strength estimation has primarily been limited to EMG acquisition from the upper forearm, necessitating verification of estimation accuracy at alternative sites. In an experiment involving seven participants, root mean square was used for signal smoothing during grip force estimation, integrated electromyography (IEMG) was employed as the EMG predictor, and a linear regression model was used for prediction. Data smoothing was compared using time windows of 30 ms and 100 ms. The results showed higher prediction accuracy at 100 ms than at 30 ms. Additionally, discrepancies were observed between estimated and actual grip force values for several subjects. These findings demonstrate the feasibility of wrist-based grip force estimation under specific conditions. Future work will explore alternative EMG estimators and prediction methods, as well as strategies to reduce discrepancies between estimated and measured values.
Chiba et al. (Thu,) studied this question.