Real-time isometric pinch force prediction from sEMG

Abstract

This paper describes a real-time isometric pinch force prediction algorithm using surface electromyogram (sEMG). The activities of seven muscles related to the movements of the thumb and index finger joints, which are observable using surface electrodes, were recorded during pinch force experiments. For the successful implementation of the real-time prediction algorithm, an off-line analysis was performed using the recorded activities. From the seven muscles, four muscles were selected for monitoring using the Fisher linear discriminant paradigm in an off-line analysis, and the recordings from these four muscles provided the most effective information for mapping sEMG to the pinch force. An ANN structure was designed to perform efficient training and to avoid both under-fitting and over-fitting problems. Finally, the pinch force prediction algorithm was tested with five volunteers and the results were evaluated using two criteria: normalized root mean squared error (NRMSE) and correlation (CORR). The training time for the subjects was only 2 min 29 sec, but the prediction results were successful with NRMSE = 0.093 ±0.047 and CORR = 0.957 ±0.031. These results imply that the proposed algorithm is useful to measure the generated pinch force without force sensors. The possible applications of the proposed method include controlling bionic finger robot systems to overcome finger paralysis or amputation.

Neuro-Based Thumb-Tip Force and Joint Angle Modelling for Development of Prosthetic Thumb Control

“…For the developed network, the choice of the optimal number of hidden neurons used in the hidden layer is based on the evaluation of the ANN performance using Normalized Root Mean Square Error (NRMSE). The best ANN architecture is shown by the smallest value of NRMSE or test error [20]. It should be noted that an insufficient number of neurons could cause the network to be unable to model the complex data, resulting in poor fitting of the model.…”