Discrete Hand Motion Intention Decoding Based on Transient Myoelectric Signals

Yurong Li,Jun Chen,Nianyin Zeng,Qian Zhang,Qianhui Zhang

doi:10.1109/access.2019.2923455

Yurong Li, Jun Chen + Show 3 more

Open Access

https://doi.org/10.1109/access.2019.2923455

Copy DOI

Abstract

In the application of human-robot interaction (HRI) rehabilitation exercise controlled by Electromyogram (EMG), if the discrete motion intention decoded by EMG signals is within the range of electromechanical delay (EMD), through the rehabilitation training, it will largely enhance the user's feeling feedback and fully activate the brain plasticity. So the decoding of hand movement intention by transient EMG is investigated to reach ideal characteristics needed in the HRI. The high-density EMG signal (HDEMG) database CapgMyo was used to decode the motion intention based on the transient EMG signals within the EMD rather than the whole recorded signals. We investigate the impact of the different decoding window lengths (WL) and training sets constructing methods when using several machine learning algorithms. In addition, the transient EMG signals decoding performance of the sparse multi-electrode EMG signal database NinaPro performing the same hand movement was compared. The visual inspection of the EMG map was used to determine the onset of HDEMG. The proposed approach was tested on EMG decoding window length of 150 ms, demonstrating a mean±SD testing performance of 94.21%±4.84% after voting. However, it is worth noting that sparse EMG signal did not achieve the desired decoding accuracy. The result showed that the high-density EMG signal could be used to decode the motion intention within EMD by simple machine learning algorithms, and extending the window length of training set could improve the decoding accuracy.

Highlights

Robot-assisted training provides an effective approach to impairment of hand function, such as neural prosthesis and exoskeleton rehabilitation robot generally controlled by neurophysiological signals, such as EEG [1] and EMG [2], [3]
In view of the superior performance of machine learning in pattern recognition, this paper focuses on solving the above problems of decoding transient EMG signals with it
The aim of this paper is to study the influence of the decoding window length, training set constructing methods and the myoelectric detection type on the discrete decoding of transient EMG signals, so the selection of feature quantities and classifiers are not selected preferably under this premise

Summary

Introduction

Robot-assisted training provides an effective approach to impairment of hand function, such as neural prosthesis and exoskeleton rehabilitation robot generally controlled by neurophysiological signals, such as EEG [1] and EMG [2], [3]. The surface EMG is a non-invasive bioelectrical signal containing rich motion information that reflects users’ motion intentions. EMG-based motion intention decoding is widely used in the brain-muscle computer interface (BMI), or called muscle-computer interface (MCI). The key of human-computer interaction through EMG is to recognize the users’ motion intentions accurately, including discrete and continuous situations. As the extension of continuous motion estimation, the estimation of softness such as joint stiffness and impedance is important to improve the natural interaction ability of human-machine. In order to realize the hierarchical control as the human brain, only the effective synergy of discrete and continuous intent decoding can realize the natural control of human-machine

Objectives

Methods

Results

Conclusion