Estimate human-force from sEMG signals for a lower-limb rehabilitation robot

Abstract

This paper presents an application of artificial neural network (ANN) to estimate human forces from Electromyogram (sEMG) signals. There are plenty of algorithms that are used to obtain the optimal ANN setting. The accuracy of ANN model is highly dependent on the network parameter settings and the accuracy of target data. However, in the majority of previous studies the force data, which are collected from the force sensors or dynamiters, used as target data in the train phase. Whereas the force sensors only measured the contact force, while the EMG signals are included of contact force and limb's dynamics. Therefore, in this paper, we present the new model to estimate the force from sEMG signals. In this method, the sum of the limb's dynamics and the contact force is used as target data in the train phase. To determine the limb's dynamics, the patient's body and rehabilitation robot are modeled in the OpenSIM. The results indicated that presented model can estimate the human force from sEMG signals precisely.