SEMG Measurement Position and Feature Optimization Strategy for Gesture Recognition Based on ANOVA and Neural Networks

Changcheng Wu,Yuchao Yan,Qingqing Cao,Baoguo Xu,Aiguo Song,Fei Fei,Dehua Yang,Hong Zeng,Xiong Lu

doi:10.1109/access.2020.2982405

Abstract

Surface electromyography (sEMG) signals are widely used in the recognition of hand gestures. Nowadays, researchers usually increase the number of sEMG signal measurement positions and extract multiple features to improve the recognition accuracy. In this paper, we propose a sEMG measurement position and feature optimization strategy for gesture recognition based on Analysis of Variance (ANOVA) and neural networks. Firstly, four channels of raw sEMG signals are acquired, and four time-domain features are extracted. Then different neural networks are trained and tested by using different data sets which are obtained based on the combination of different measurement positions and features. Finally, ANOVA and Tukey HSD testing are conducted based on the gesture recognition results of different neural networks. We obtain the optimal measurement position sets for gesture recognition when different feature sets are used, and similarly, the optimal feature sets when different measurement position sets are used. Our experimental results show that the feature set of zero crossing and integrated sEMG provides the highest gesture recognition accuracy, which is 94.83%, when four channels of sEMG signals are used; the optimal measurement position set when four sEMG signal features are used for hand gesture recognition is P1+P3+P4, which provides an accuracy of 94.6%.

Highlights

In the field of human machine interaction, especially when it comes to the control of bionic prosthetics, gesture recognition based on Surface electromyography (sEMG) signals is a common method to obtain human motion intention.Recent decades have seen the emergence of lots of studies focusing on this method [1]–[7]
To realize high-accuracy recognition of multiple gestures with less data, we propose in this paper a sEMG measurement position and feature optimization strategy for gesture recognition based on Analysis of Variance (ANOVA) and a back propagation neural network (BPNN)
In this paper, we aim to address problem that the increase in the number of signal measurement positions and features adds computational burden to hand gesture recognition based on sEMG and neural networks

Summary

Introduction

In the field of human machine interaction, especially when it comes to the control of bionic prosthetics, gesture recognition based on sEMG signals is a common method to obtain human motion intention.Recent decades have seen the emergence of lots of studies focusing on this method [1]–[7]. Moin presented a gesture recognition system using a high-density sensor array and a robust classification algorithm, achieving an average classification accuracy of 96.64% for five gestures [10]. She et al proposed a novel time-frequency analysis method based on the Stockwell transform to improve hand movement recognition accuracy from forearm sEMG signals [11]. They collected two channels of sEMG signals and used neural networks to identify six gestures. In a study by Ariyanto et al, based on one

Objectives

Methods

Results

Conclusion