Motion identification based on sEMG for flexible pneumatic hand rehabilitator

Abstract

Purpose

This paper aims to avoid the precise modeling and controlling problems of rigid structures of hand recovery device, by proposing a hand rehabilitator based on flexible pneumatic actuator with its safety and adaptability.

Design/methodology/approach

The hand rehabilitator is designed based on a flexible pneumatic bending joint. The recovery training program for an injured finger is developed via forearm sEMG (surface electromyogram) sampling, analysis, classification and motion consciousness identification. Four typical movement models of the index finger and middle finger were defined and the corresponding sEMG signals were sampled. After simulation and comparative analysis, autoregressive (AR) model back propagation (BP) network was selected for sEMG analysis and hand recovery planning because of its best recognition performance. A verification test was designed and the results showed that the soft hand rehabilitator and recovery conception are feasible.

Findings

AR model BP network can identify the index finger and middle finger movement intention via an sEMG analysis. The developed flexible pneumatic hand rehabilitator is safe and suitable for finger recovering therapy.

Research limitations/implications

Because of the limitation of experimental samples, the prototype rehabilitator of this work may lack generalizability for other situations. Therefore, for further study and application, systematic structure revising, experiments, data and training are necessary to improve the performance.

Practical implications

The paper includes implications for the development and application of a new style, safe and dexterous hand rehabilitator.

Originality/value

The paper tries a new approach to design a safe, flexible and easily controlled hand rehabilitator.

Citation

Bao, G., Li, K., Xu, S., Huang, P., Wu, L. and Yang, Q. (2015), "Motion identification based on sEMG for flexible pneumatic hand rehabilitator", Industrial Robot, Vol. 42 No. 1, pp. 25-35. https://doi.org/10.1108/IR-08-2014-0376