Real-time recognition of arm motion using artificial neural network multi-perceptron with arduino one microcontroller and EKG/EMG shield sensor

Luis A. Caro, Camilo Silva, Billy Peralta, Oriel A. Herrera, Sergio Barrientos

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Citation (Scopus)

Abstract

Currently, human-computer interfaces have a number of useful applications for people. The use of electromyographic signals (EMG) has shown to be effective for human-computer interfaces. The classification of patterns based on EMG signals has been successfully applied in various tasks such as motion detection to control of video games. An alternative to increasing access to these applications is the use of low-cost hardware to sample the EMG signals considering a real-time response. This paper presents a methodology for recognizing patterns of EMG signals given by arm movements in real time. Our proposal is based on an artificial Neural Network, Multilayer Perceptron, where the EMG signals are processed by a set of signal processing techniques. The hardware used for obtaining the signal is based on Ag/AgCl connected to the EKG/EMG-Shield plate mounted on a Arduino One R3 card which is used to control a video game. The implemented application achieves an accuracy above 90% using less than 0.2 s for recognition of actions in time of testing. Our methodology is shown to predict different movements of the human arm reliably, at a low cost and in real time.

Original languageEnglish
Title of host publicationAmbient Intelligence for Health - 1st International Conference, AmIHEALTH 2015, Proceedings
EditorsVladimir Villarreal, José Bravo, Ramón Hervás
PublisherSpringer Verlag
Pages3-14
Number of pages12
ISBN (Print)9783319265070
DOIs
Publication statusPublished - 1 Jan 2015
Event1st International Conference on Ambient Intelligence for Health, AmIHEALTH 2015 - Puerto Varas, Chile
Duration: 1 Dec 20154 Dec 2015

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9456
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference1st International Conference on Ambient Intelligence for Health, AmIHEALTH 2015
Country/TerritoryChile
CityPuerto Varas
Period1/12/154/12/15

Keywords

  • Action recognition
  • Arduino
  • Microcontrollers
  • Neural networks

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'Real-time recognition of arm motion using artificial neural network multi-perceptron with arduino one microcontroller and EKG/EMG shield sensor'. Together they form a unique fingerprint.

Cite this