Estimation and early prediction of grip force based on sEMG signals and deep recurrent neural networks

Hands are used for communicating with the surrounding environment and have a complex structure that enables them to perform various tasks their multiple degrees of freedom. Hand amputation can prevent person from performing daily activities. In event, finding suitable, fast, reliable alternative missing limb affect lives people who suffer such conditions. As most important use hands is grasp objects, purpose this study accurately predict gripping force surface electromyography (sEMG) signals during pinch-type grip. regard, sEMG derived 10 healthy subjects. Results show task, recurrent networks outperform nonrecurrent ones, as fully connected multilayer perceptron (MLP) network. Gated unit (GRU) long short-term memory (LSTM) R-squared values 0.994 0.992, respectively, prediction rate over 1300 predictions per second. The predominant advantage using frameworks be predicted straight preprocessed without any form feature extraction, not mention ability future larger horizons adequately. methods presented in myoelectric control prosthetic or robotic grippers.