Control de movimiento robótico con detección cognitive y facial mediante Emotiv EEG

Abstract

This papers presents the development of a brain-computer interface using the Emotiv EEG device, where a facial gestures and cognitive thought are detected, which will be translated into movements by a Lego Mindstorms robot. This translation is made to verify if the Emotiv is able to perform a good recognition and it can be used to control the movement of devices. This research work was motivated by previous works, where different recognition techniques were applied, but in some cases the speed and in others the precision of detection were not the expected. Therefore, in this research work uses others methods to solve these problems. Recognition tests were performed with 6 persons, resulting on an 81% of precision with cognitive recognition for two actions learned and a 67% of precision with facial recognition for six gestures used. Lately, a test was performed in a system implemented under an event-driven architecture, using the Emotiv device and the Lego robot, obtaining a precision on detection of 70%, evaluated by comparison between the action desired by the user and the action effectively performed by the robot, which shows that the Emotiv device is recommendable for its use, given the quality/price factor, because it gives good recognition results and allows multiple actions in real time.