Grip strength forecast and rehabilitative guidance based on adaptive neural fuzzy inference system using sEMG

In order to resolve the problem of unstable control of force in human–computer interaction based on surface EMG signals, the adaptive neural fuzzy inference system is designed to achieve the grip strength assessment. As we know, the acquisition of surface EMG signal is non-invasive, which provides a better evaluation index for rehabilitation training in the medical process. By establishing the relationship between grip force and surface electromechanical signals, the effect of rehabilitation training can be evaluated directly while reducing the types of sensors used. Firstly, the experimental equipment are introduced, which are utilized to carry out simultaneous acquisition of surface EMG signals and forces. Then, the traditional features of sEMG and the corresponding algorithms are illustrated, based on this, supplementing the energy eigenvalue with wavelet analysis and fuzzy entropy. In which, fuzzy entropy is effective in characterizing muscle fatigue that can effectively reduce the impact of muscle fatigue on force assessment. Finally, combining fuzzy logic implication and neural network, the adaptive neural fuzzy inference system is designed, which is trained by extracted feature vectors. The experimental result shows the method used in this paper can effectively predict the grip force. Further, force prediction based on sEMG can be used to guide rehabilitation therapy in virtual space, combined with an electrical stimulator.