A Deep Learning Approach for Position Control of a Mobile Robot

Abstract

The study presents an innovative method for decoding the brain signals of subjects to control a Khepera IV mobile robot, thereby providing assistance to those with neuro-motor disorders. The afore-said objective is accomplished through a novel Convolutional Neural Network (CNN) that precisely classifies four motor imagination labels (left hand, right hand, left leg, and right leg) corresponding to four Khepera robot maneuvers (left, right, forward, and backward).