Remotely Controlled Human-Machine Interface, Managed by Means of Characterized Bioelectric Signals

Abstract

Today, a variety of human-machine interfaces are available. Many of them require electronic or mechanical elements to generate the control words, which can make them cumbersome and complicated to use. For this reason, it is necessary to have interfaces that are capable of responding in a natural and accessible way to any person, regardless of social or physical features. In this same sense, the generation of nano interfaces is sought, which may be imperceptible when used, avoiding any type of user annoyance.In this work, a more natural human-machine interaction system is proposed, which depends on the muscular electrical impulses that a person can create through arm movements. Such impulses are characterized by a back-propagation neural network. It was possible to obtain control commands by using only electrical impulses that are generated by compound movements of an arm. The interface that was developed in this project generates control commands that are sent remotely via XMPP to manipulate a vehicle that is in a geographical position that differs from the position of the user, with which we demonstrate that a bioelectric impulse can be sufficient for controlling electronic devices within the Internet of Things.The proposed interface can be used by almost anyone, including people with different abilities and even illiterate people.