Method of Finger Motion Recognition Based on Polyvinylidene Fluoride Sensor Array

Finger motion recognition is one of the key technologies of human–computer interaction based on gestures. In this paper, we propose a method of recognizing finger motions by using a wearable wrist device (WWD). This method not only avoids the problem that the user's hands are limited by wearing...

Full description

Saved in:
Bibliographic Details
Published in:Sensors and materials 2019-01, Vol.31 (9), p.2931
Main Authors: Hu, Yaohui, Xie, Lingrui, Chen, Yadong, He, Ke, Fang, Yong, Kang, Wuwei, Yao, Zixian, Fang, Guoqing
Format: Article
Language:English
Subjects:
Eigenvectors
Fluorides
Fragments
Gesture recognition
Human motion
Motion perception
Optimization
Polyvinylidene fluorides
Product design
Recognition
Sensor arrays
Sensors
Wrist
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Finger motion recognition is one of the key technologies of human–computer interaction based on gestures. In this paper, we propose a method of recognizing finger motions by using a wearable wrist device (WWD). This method not only avoids the problem that the user's hands are limited by wearing motion detection sensors, but also avoids the problem that vision-sensorbased gesture recognition technology is difficult to use in a mobile environment. Moreover, this method, which uses polyvinylidene fluoride (PVDF) sensors as the detection units of WWD, has the advantages of being noninvasive, comfortable, and convenient. In this work, we first studied the distribution and optimization of the PVDF sensor array, and used this array to complete the acquisition of wrist motion signals. Then, we used short-term energy to solve the problems of the real-time detection of motion signals' endpoints and the extraction of motion signal fragments. Then, we encoded these fragments into 64-digit eigenvectors for finger motion recognition. In the experiment, we transmitted the eigenvectors as input values into a four-layer BP network to recognize three essential finger motions for mouse control. The experimental results show that the recognition effect of this method is satisfactory and the recognition accuracy is up to 96.7%.
ISSN:0914-4935
DOI:10.18494/SAM.2019.2444